Improving Tuberculosis Control in Ethiopia
Improving Tuberculosis Control in Ethiopia
Improving Tuberculosis Control in Ethiopia
2011
Improving Tuberculosis Control in Ethiopia
cost-effectiveness
By
ISBN 978-82-308-1693-6
%HUJHQ1RUZD\
k'DQLHO*HPHFKX'DWLNR
ii
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
&
iii
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Table of contents ....................................................................................................... pages
Original papers................................................................................................................iii!
Summary........................................................................................................................... vi!
Annexes!
iv
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Acknowledgments
for his dedication, understanding and technical support to make this work a reality. I thank
the Norwegian government for financing my study and the SNNPRS Health Bureau for
providing technical and material support. My appreciation also goes to the administrative
I owe sincere gratitude to brothers and sisters in Agape Evangelical Church in Bergen for
the fellowship we had in Christ Jesus and special moments we shared together. Special
thanks to families of Yonna, Eyob and Solomon whose hospitality made me feel at home.
I am grateful to health office managers, GHWs and HEWs in Dale and Bokaso districts
for their support during the field work; special thanks to Mr. Melkamsew Aschalew, Mr.
Thank you my father, Gemechu Datiko Tatole, for sending me to school, my brother
Sisay, for supporting my family and my sisters: Yamrot, Zertihun, Wengelawit and
Gedamnesh. My late brother Dawit Gemechu, I owe sincere gratitude and have a special
place for you in my life: your prayers, blessing and wishes remain with me. This world is
a lesser place to live in without you. Late Frew Haile, thank you for your goodness to my
family.
i
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
My beloved wife Yewolega work Haile, no words can express my appreciation for your
strength, prayer and care. Our beloved children: Yohannes, Eyu and Shalom, you are at
the heart of our family. My appreciation also goes to relatives and friends who have
encouraged my family during the study period. Dr. Degu Jerene, mission accomplished.
Dr. Shelemo Shawula and your family, I have no words to express your goodness, God
My Lord and God, praise you for your great deeds, you accomplished as you have spoken.
I know that you can do all things; no plan of yours can be thwarted. Job 42:2
ii
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Original papers
This thesis is based on the following papers, which will be referred to in the text by their
Roman numerals:
PAPER I. Yassin MA, Datiko DG, Shargie EB (2006) Ten-year experiences of the
PAPER II. Datiko DG, Yassin MA, Chekol LT, Kabeto LE, Lindtjorn B: The rate
PAPER III. Datiko DG, Lindtjørn B (2009) Health Extension Workers Improve
doi:10.1371/journal.pone.0005443
PAPER IV. Datiko DG, Lindtjørn B (2010) Cost and Cost-Effectiveness of Treating
iii
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
List of abbreviations
iv
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
TB Tuberculosis
v
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Summary
World Health Organization (WHO) recommends directly observed treatment short course
(DOTS) strategy to control tuberculosis (TB). It aims to detect 70% of new smear-positive
cases and cure 85% of them. Implementing the DOTS strategy has improved the case
detection rate (CDR) and treatment success rate (TSR) in many settings.
CDR mainly because of inability to access the health service was the limit. We also
explored alternatives that could improve access to the health service, its cost-
patients under DOTS strategy and the rate of human immunodeficiency virus (HIV)
infection in TB patients.
In ten years, TB control programme of southern Ethiopia has improved the case detection
(from 22% to 45%) and treatment success (from 53% to 85%). However, the target of
CDR seemed unachievable. Some of the reasons were low health service coverage,
shortage of general health workers (GHWs), HIV epidemic and poor socioeconomic
collection and providing directly observed treatment (DOT). This improved the CDR,
vi
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
more significantly for women. This could be mainly because of the community-based
sputum collection that had increased access to the diagnostic service. Moreover,
compared with to health facility-based DOT (83%). This could be due to the improved
access to the service that was created through the provision of DOT in the community
the available resources and existing supporting evidences. This is more important in
resource-constrained settings with high disease burden. We, therefore, estimated the cost
smear-positive cases in the community reduced the total, patient and caregiver costs by
We also estimated the recurrence rate and mortality in TB patients cured under DOTS
strategy and the rate of HIV infection in TB patients and the community. The rate of
recurrence in smear-positive TB patients cured under DOTS strategy was 1 per 100 PYO
(0.01 per annum). The rate of TB-HIV co-infection varied with the prevalence of HIV in
the community. We found mortality rate of 2.5% per annum in successfully treated TB
patients. The mortality was associated with sex, age and occupation.
involving HEWs in TB control programme as we found improved the CDR and TSR.
vii
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
viii
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
1.0. Introduction
tuberculosis (PTB) patient discharging the bacilli. It mainly spreads by airborne route
when the infectious patient expels droplets containing the bacilli. It is also transmitted by
The risk of infection depends on the susceptibility of the host, the extent of the exposure
and the degree of infectiousness of the index case [3-5]. Once an individual inhales the
infectious aerosols, the bacilli lodge into the alveoli where it multiply and form a primary
lesion [6]. Under normal condition, in most of the cases, the immune system either clears
the bacilli or arrests the growth of the bacilli within the primary lesion in which case the
host is said to harbor latent TB infection (LTBI). However, in 5 - 10% of the cases, the
bacilli overwhelm the immune system resulting in a primary TB within a few months to
years. In the rest, post-primary TB occurs when reinfection occurs or the LTBI is
infection, diabetes and other medical conditions that suppress immunity) and poor
Although TB affects many parts of the body, it mainly affects the lung. Its clinical
presentation, therefore, depends on the site of infection, the organ affected and its
1
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
severity. Patients with PTB present with pulmonary symptoms (like productive cough,
haemoptysis, chest pain and shortness of breath), constitutional symptoms (like fever,
poor appetite, weight loss, night sweats and anorexia) and other symptoms depending on
the site of the infection [8-11]. Understanding of the symptoms is important to inform the
community about the symptoms to seek medical advice and to inform health workers in
order to increase the index of suspicion to easily pick suspects and detect tuberculosis
Early detection of the cases and prompt treatment are crucial for TB control. TB diagnosis
mainly depends on the clinical presentation of the disease and identification of the
offending bacilli. Many TB diagnostic tests are available although no single diagnostic
test for TB exists that can be performed rapidly, simply, inexpensively, and accurately as
feasibility and ability to identify the highly infectious forms of TB, the smear-positive
treatment. If not treated in the earliest five years, 50% of PTB cases die, 25% self cure
and 25% remain sick and infectious. Untreated smear-positive PTB patient can infect 10 -
15 people per year on average [15]. Thus, treatment of TB is not only a matter of treating
the individual patient, but also is an important public health intervention. Treatment is the
centerpiece of TB control and can reduce the risk of infection if implemented with
2
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
adequate coverage and acceptable quality [16, 17]. So far, DOTS remains a cost-effective
TB has been a scourge of humanity throughout recorded history. Even today after the
availability of effective drugs for more than half a century, it is a major cause of
infected with MTB [6, 20]. There were about 9.27 million new TB cases (including 4.1
million new smear-positive cases) and 1.3 million deaths from TB in 2008. There
were about 11.1 million prevalent TB cases and half million multidrug resistant (MDR)
and 98% of TB deaths occurred in developing countries [21, 22]. TB also is a leading
infectious cause of death among women in the reproductive age group and affects the
HIV infection and accounts for 26% of preventable deaths in the world [21, 24-26].
Age
The risk of acquiring TB infection increases with age from infancy to early adult life,
disease of adults in the age group of 15 - 49 years. In a population where the transmission
has been stable or increasing, the incidence rate is higher in children mostly because of
recent infection or reinfection. As transmission falls, the case load shifts to older adults
3
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Gender
Reports show that men account for high proportion of notified TB cases than women [28,
29]. This was explained by sex ( biological determinant - progression from TB infection
to disease is likely to be faster for women compared with men in their reproductive years)
differential access to health care (like economic problem, inability to make decisions,
poor health seeking behaviour and stigma ) that compromise the women’s ability to utilize
the available health service [11, 30-36]. In addition, higher risk of HIV infection among
women makes them susceptible to develop active TB. TB is the leading infectious cause
of death in young women in developing countries [35, 37, 38]. This could be worse in
settings with health services insensitive to gender-specific needs [39]. Studies that
consider the interplay of biological, socio-cultural and health system determinants of sex
and gender-based differences are needed to understand how and why women are affected.
Residence
More TB patients were reported from urban than rural areas because of overcrowding,
poverty and HIV infection [40]. In contrast, the presumed lower risk of TB infection in
rural settings could be misleading and should be cautiously taken in high burden
countries. In the rural settings, access to the health service is limited; health seeking
behaviour is poor and the living condition favour disease transmission. As a result,
understanding the burden of TB in rural areas will have a wider implication for TB
4
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Socio-economic conditions
overcrowding and malnutrition. The magnitude of TB is high among the poor, displaced,
homeless, drug addicts, elderly, malnourished and women [14]. The association between
TB and poverty was shown by the decline in TB burden with the improved living
condition was also found to reduce the risk of infection from 4 - 6 % per annum. In
shows its association with poor living conditions. TB was also found to disproportionately
affect the poor [43-45]. Therefore, free diagnosis and treatment was offered to TB patients
(mainly to smear-positive PTB cases) to reduce the economic burden for seeking
diagnosis and treatment and treat the highly infectious cases [46]. However, limited
access to the service because of the poor socioeconomic condition of the patients and their
households has reduced the utilization of the available service [42, 46-51]. Thus,
A complex interaction exists between TB and HIV infection. HIV increases the risk of
infection, as it reactivates LTBI and increases the progression to active disease. TB-HIV
co-infection has fatal consequences as TB becomes the leading cause of death in HIV
infected individuals and patients with acquired immunodeficiency syndrome (AIDS). HIV
lowers the host’s immune response to MTB. The lifetime risk of developing active TB in
HIV infected individuals is 10% per year compared with lifetime risk of 5 - 10% in
5
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
individuals without HIV. As a result, the TB case notification rate (CNR) has increased
cases and by compromising the treatment outcomes. It created a huge challenge to the
already overstretched and under staffed health system in high burden countries. It reduced
the proportion of smear-positive cases; and increased the rate of treatment failure,
defaulter and death, which in turn compromised the progress towards achieving the
Globally 1.37 million TB cases (14.8% of 9.27 million cases) were co-infected with HIV.
70% of TB-HIV co-infections occurred in countries with high burden of TB. Moreover,
half million deaths occurred in HIV infected people due to TB which accounted for a
quarter of deaths among HIV positive people [21, 55-58]. There is an epidemiological and
cases infected with the bacilli, from no remedy to effective treatment, from compulsory
isolation to chemical isolation (treating infectious cases with anti-TB drugs), and from
vertical to integrated approach where the service delivery was progressively decentralized
6
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Robert Koch’s identification of the bacilli and the proposal to isolate patients was
followed by compulsory isolation of the patients as the main principle of TB control. This
included social support and contact examination in TB clinics (that were accessible and
programme staffed with health workers particularly assigned to run the programme. This
reduced the annual risk of infection by 5 -13% in developed countries due to the available
resources and improved general living conditions [67]. However, similar results were not
achieved in developing countries due to the associated high cost [68]. Hence, TB control
was integrated into general health service to ensure effective and efficient use of resources
[69, 70]. However, lack of technical efficiency by the GHWs, neglect of TB control
activities, health sector reform (that resulted in collapse of TB structure because of hasty
to HIV epidemic weakened the TB control efforts [71, 72]. This was also complicated by
The affordability of rifampicin, poor treatment adherence and high TB burden paved way
to the introduction of DOTS strategy [73, 74]. The components of the strategy are
short course chemotherapy using six to eight months treatment regimens, regular and
uninterrupted supply of anti-TB drugs and standardized recording and reporting system
[64, 75].
7
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
DOTS strategy aims to detect 70% of new smear-positive cases and to cure 85% of them.
incidence, prevalence of smear-positive cases and the number of infected contacts [14,
76]. Decentralization of DOTS to peripheral health facilities has increased the number of
TB cases that were detected and treated [77] even in settings with high HIV infection and
MDR TB [18, 78]. However, its effectiveness was limited in settings with low health
service coverage.
as a public health problem (one smear-positive case per 106 population) and, ultimately, to
achieve a world free of TB. The partnership promotes TB control as an element for
TB control approach that includes providing high quality DOTS expansion and
enhancement, addressing TB-HIV, MDR TB and address the needs of the poor and
strengthening the health system, engaging all care providers, empowering TB patients and
the community with partnership, and enabling and promoting research to alleviate human
suffering [79].
Ethiopia is a located in East Africa. It covers an estimated area of 1.1 million km2. The
Government of Ethiopia has nine ethnic-based administrative regions, which are referred
8
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
to as Regional States and two Federal City Administrations. There are 611districts and
15000 kebeles (lowest administrative unit with an average population of 5000 people)
[80]. Ethiopia has a population of about 76 million people, of which 85% lives in rural
areas. It has huge topographic variations ranging from the lowest 116 meters below sea
level to 4, 620 meters above sea level. Most of the population lives in highland areas on
subsistence farming.
The National health policy of Ethiopia emphasizes the development and provision of
equitable and acceptable health service to the people. Under this provision, the
community-based health services. Primary health care unit (a health centre and five
satellite health posts) is the lowest unit in the health system (Figure 1). Health centres
provide curative and preventive health service while health posts mainly focus on disease
prevention and health promotion activities and selective curative services in the
community.
responsive health delivery system to the people living in rural areas. This was
focuses on promoting health and providing preventive and selected curative services to
9
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
ensure equitable access to the community under three major categories: disease
prevention and control, family health service, hygiene and environmental sanitation, and
The local health authorities in consultation with kebele leaders select two women from
each kebele, who have completed 10th grade. The women receive one-year training before
they are placed as HEWs in their respective kebeles. TB control is included in the training
of HEWs as one of the components under Disease Prevention and Control. HEWs are
responsible to provide health education, identify and refer TB suspects, trace defaulters
Assigned to health post, HEWs spend about three-fourth of their time on outreach
activities in the communities, the kebele in particular. Each HEW is responsible for 500 -
1000 households in each kebele. They receive salary from the government and are
accountable to health centre and the kebele administration [83, 84]. However, their
contribution to TB control has not been evaluated. The health system, administrative
10
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Federal Specialized
Ministry of referral
Health Hospital
Regional Regional
Health Referral
Bureau Hospital
Zonal Zonal
Health Hospital
Department
Woreda District
Health Hospital
Office
Health Centre
Primary
Health Unit
5 Health Posts
11
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
TB is among the leading causes of morbidity and mortality in Ethiopia [80]. The NTLCP
was started in 1992. Ethiopia adopted the WHO recommended DOTS strategy in 1995.
Since then TB control efforts have been decentralized to public health facilities (hospitals,
health centres and health stations) where GHWs are responsible for the diagnosis and
Ethiopia has one of the highest TB burden in the world. The annual incidence and
prevalence of all forms of TB was 378 and 579 per 105 populations, respectively [21].
DOTS is implemented in public health facilities in hospitals and health centres [80].
The CNR for all forms TB was 155 per 105 populations. The incidence of smear-positive
cases was 163 cases 105 populations. The CNR and CDR of new smear-positive cases was
46 per 105 populations and 28 %, respectively. The TSR, the proportion of patients who
Southern Nations, Nationalities and People’s Regional State (SNNPRS) is one of the
Federal States of Ethiopia. It has a population of about 15 million. Ninety-three per cent
of its population lives in the rural areas. The health service coverage and user rate is about
The SNNPRS Health Bureau started DOTS in three zones and four health facilities as a
vertical program in 1995. Later, the programme was integrated into the general health
service, and decentralized to zones, districts and health facilities (hospitals, health centres
12
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
and health stations). Over ten years, implementation the DOTS strategy tripled TB case
notification (45 to 143 per 105 population), doubled the case detection rate (22 to 45%),
increased the treatment success rate (from 53 to 85%), reduced the defaulter rate (from 26
TB remains the leading cause of morbidity and mortality in southern Ethiopia. It was 4th
cause of total admission, 5th cause of female admission and 5th cause of admission in
children less than five year. It was the 3rd cause of inpatient deaths, 2nd cause of deaths in
women and 6th cause of death in children less than five years [80].
Over the last two decades, the load of TB has increased in sub-Saharan Africa mainly
because of HIV infection. This has compromised the already overstretched health services
due to the associated morbidity. The implication is more in settings with low health service
DOTS strategy advocates passive case finding and provision of DOT under the direct
observation of GHWs or treatment supervisors. Passive case finding mainly serves those
who have better socioeconomic status (better knowledge and health seeking behaviour) and
geographic access to health facilities. This affects the poor and patients living in rural and
remote areas leading to delay in presentation and disease transmission [42, 44]. Moreover,
seeking diagnosis and treatment in health facilities is costly and difficult for TB patients and
their families [47, 87-89]. In TB patients who have accessed the service, adherence and
13
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Therefore, alternative approaches that improve access to diagnosis and treatment are needed
[90-95].
Ethiopia has one of the highest TB burden in the globe. TB is among the leading causes of
morbidity and mortality. Over the last two decades implementing DOTS strategy has
increased the number of TB patients diagnosed and treated. However, the CDR for smear-
positive patients remained far below the target despite increasing number of health workers
and health facilities providing DOT. This was mainly because of the limited access to the
health service. However, decentralization and performance of DOTS strategy in the era of
provide health service in each kebele by a new cadre of health workers. However, the role of
HEWs in TB control has not been explored. Therefore, we aimed to measure the
performance DOTS, identify the role of HEWs in improving the performance of the TB
14
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
2.1. Aim
The aim of this research was to improve the performance of TB Control in Ethiopia
2.2. Objectives
2. To estimate the rate of HIV infection in TB patients and its association with the
3. To find out if involving HEWs in TB control improves the CDR and TSR of smear-
positive patients
15
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
3.0. Methods
Ethiopia is the third largest and populous country in Africa (Figure 2). It covers an area of
1.1 million km2 and has a population of 76 million people. 85% of the population lives in
rural areas. It has nine regional states and two city administration. It is undergoing
enhanced health facility construction and training of health workers to improve the delivery
Ethiopia is a poor country with one of the worst health indicators in the world [96].
Communicable diseases and nutritional deficiencies are the main causes of morbidity and
service delivery and extending affordable primary health care to the community [80, 97].
16
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Indicators Rates
2008 2009
SNNPRS is the third largest and populous region in Ethiopia. It is located in the south-west
part of the country. It shares international borders with Sudan and Kenya. It covers 118, 000
square km (10% of national area). It has a population of about 15 million (20% of the
national population). It has 13 zones, one city administration and eight special districts. 93%
of the population lives in rural areas. The health service coverage and user rate was 73.5 %
and 32%, respectively. TB treatment is provided in all hospitals and health centres (DOTS
coverage was 100% in hospitals and health centres). However, only 16% (354/2230) of
DOT has not been implemented in health posts and HEWs do not provide DOT to TB
patients.
17
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Sidama zone is located in SNNPRS. It has 19 districts and two towns. It has about 3.2
million people in an area of 6,981 square km. It is one of the most densely populated
zones of the region with a population density of 463 people per square km. Fifty-five per
cent of the population lives at a two-hour walking distance from the health facilities. Dale
and Wonsho are rural districts in Sidama zone located about 50 kms from Hawassa, the
Ethiopia! Sidama!zone
Dale!and!Wonsho!districts
SNNPR!State
18
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
The SNNPRS Health Bureau started implementing DOTS strategy in 1995. TB control
programme was decentralized to hospitals, health centres and health stations. TB case
finding and treatment outcome reports were complied on quarterly basis. TB patients
reported from 1995 - 2004 were enrolled to measure the performance of TB control
In 2005, after implementation of DOTS strategy for ten years, the prevalence of HIV and
its association with the rate of TB-HIV co-infection was estimated by enrolling pregnant
women and TB patients in the southern Ethiopia. This was performed as part of regular
HIV surveillance conducted by the SNNPRS Health Bureau. Paper II was based on the
data obtained from TB patients and pregnant women attending health facilities for
antenatal care (ANC) for the first time during the study period.
completed) under DOTS strategy in two rural districts of Sidama zone (Dale and Wonsho)
were retrospectively followed to the first cohort that received DOT (Paper V & VI).
community randomized trial conducted in the two rural districts of Sidama zone. TB
patients identified from intervention and control kebeles received DOT. These patients
were followed until they completed treatment for TB (Paper III). Prospectively cost data
was collected for these patients, caregivers and health workers to find out the cost and
19
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
cases were excluded from the study. However, they received the treatment available in
The main study design was community randomized trail (Paper III). In this study, 51
kebeles from two rural districts were randomly allocated to intervention and control
groups. Kebele was the unit of randomization. We used table of random numbers for
allocation. TB patients diagnosed from the intervention kebeles were started on DOT under
the direct observation of HEWs while patients from the control kebeles received health
facility-based DOT under the direct observation of GHWs. Cost data was prospectively
collected for these patients along the main study to estimate the cost and cost-effectiveness
In Paper II, TB patients and pregnant women were enrolled in a cross-sectional study to
find out the rate of TB-HIV co-infection and the prevalence of HIV in southern Ethiopia.
TB patients and pregnant women presenting to the health facilities were consecutively
enrolled after obtaining informed consent. HIV testing was done (from the remaining
serum after routine blood test for pregnant women and from sample collected for
surveillance of HIV in TB patients) at the Centre for Health Research Laboratory (CHRL)
Paper I reports a cross-sectional study conducted to assess the ten-year performance of the
TB control programme based on the reports compiled at the SNNPRS Health Bureau. We
also conducted a retrospective cohort study to find out the recurrence and mortality rate in
20
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
TB patients after successfully receiving treatment under DOTS strategy. TB patients who
were declared treatment completed or cured were traced to their home (Paper V & VI).
21
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Table - 2 Summary of the studies conducted: the design, population and the study period
II. The rate of TB HIV co- Cross sectional 1308 TB patients 2005 - 2006
III. HEWs improved the case Community 318 TB patients 2006 - 2008
22
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
To achieve the targets of DOTS strategy, understanding the transmission of the disease,
the progress from infection to disease and effective control measures that early identify
the cases, cure the disease and maintain disease free survival of the cases and the
prevention of the transmission, recognizing the symptoms and seeking diagnosis should
be complemented by the ability of health workers to identify the cases, provide prompt
treatment and encourage the patients to adhere to the treatment. Studies and intervention
The aim of our study as described earlier was to improve the performance of TB control
relation to the patient, community and health system. To put our study in context we have
depicted how our studies fitted into the public health model of TB control adopted from
23
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Figure 3 - The schematic presentation of the studies in the public health model of TB
control
24
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
NTLCP uses standard recording and reporting formats to monitor TB control efforts. Case
finding and treatment outcome data were complied and reported on quarterly basis from
lower to higher administrative levels (from District Health Offices ! Zonal Health
departments ! Regional Health Bureaus ! Federal Ministry of Health) and copies of the
reports were kept at all levels for official documentation ( Figure - 1 ). Ten-year TB
programme review (Paper I) was based on the copies of reports remaining at SNNPRS
Health Bureau.
In Paper II, TB patients and pregnant women were enrolled from health facilities to find
out the rate of HIV infection. Trained laboratory technicians and GHWs collected the data
using a pretested questionnaire prepared for HIV surveillance among pregnant women
and TB patients.
In paper III, smear-positive patients from intervention and control kebeles were registered
in the health facilities and were treated under the two treatment options: under the direct
list of patients from unit TB registers in the health facilities to district TB registers. They
also cross-checked the sputum results of smear-positive cases in the TB unit register
against the smear results recorded in the laboratory register. TB case finding and
treatment outcome data were quarterly reported from district TB register using standard
reporting formats.
25
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
In Paper IV, smear-positive patients from intervention and control kebeles were
prospectively enrolled and cost data was obtained using a structured questionnaire.
Trained GHWs and HEWs interviewed the patients and caregivers about the travel time,
transport cost and costs related to visit to health facilities or health posts for treatment.
Similarly, the cost data for HEWs and GHWs was collected by trained data collectors
from the two districts. The salary of established positions, budget expenditures, medical
equipments, vehicles and buildings were obtained from health facilities, district health and
finance offices. Joint costs were shared based on the proportion of time used for TB
control. Capital items were annualized using 30 years for buildings, 10 years for
equipments and 5 years for motorbikes as an expected useful life. The base year for
valuing cost was 2007, and the applicable exchange rate was 8.6 Ethiopian birr for 1
USD.
In Paper V & VI, TB patients who were treated since the start of DOTS in the study area
were retrospectively followed in the two districts. The lists of TB patients who were
declared treatment completed or cured were obtained from unit TB registers in the health
facilities and district TB registers. HEWs collected the data about the current status of the
patients if they were alive or dead. This was done by taking registered history of TB
patients from unit registers for prior successful treatment in health facilities and recent
history of TB was obtained by making house-to-house visits. The data about recurrence or
26
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
The study team consists of HEWs, GHWs, district HEP experts and TB programme
experts at Districts, Zone and Region including the principal investigator and the
supervisor.
The HEWs were responsible for case finding and provision of DOT in an intervention
kebeles. They also conducted house-to-house visit to collect data about post-treatment
follow-up of TB patients and cost data for patients receiving treatment in their kebeles.
The GHWs were responsible for supervising the TB control activities conducted by
HEWs, registration of TB patients in the unit TB register, provision of drugs and supplies
to the health post and providing DOT for patients receiving treatment in health centres.
The district HEP experts were responsible for coordinating the activities of HEWs in TB
control; and District TB programme experts were responsible for providing drugs and
supplies, coordinating the activities, ensuring the completeness and consistency of the
recording and reporting in the health facilities and report the quarterly performance of
district TB control programmes. TB and Leprosy Control Programme experts from zone
and region also conducted supportive supervision to the districts, health facilities and
kebeles.
The principal investigator was responsible for organizing and supervising the overall
conduct of the studies and reports the activities to the supervisor who also facilitated the
administrative and technical issues to accomplish the work and conducted the field visits.
27
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Effective delivery of an intervention requires both that the providers adhere to the
intervention procedure and that the participants cooperate appropriately [102]. To ensure
this, HEWs, GHWs and TB programme experts were trained about community-based TB
care and its implementation. The trainees received copies of training documents and field
conduct of case finding and treatment, and checked the completeness and accuracy of the
cost and post-treatment follow-up data. They also crosschecked patient data by making
kebeles. The district programme experts checked the completeness and accuracy of the
data, the recording and reporting of patients, the patient follow-up and the availability of
resources. The experts also cross-checked the data from unit TB register with the
laboratory register in the same health facility, the data from health facilities against the
data from the kebeles and the patients. They also collected slides from diagnostic units for
blind rechecking as part of external quality assurance and it was done at CHRL as per the
recommendation of the NTLCP. The district health office reviewed the community-based
TB care and other health activities on quarterly basis. Regular supervision was also
conducted by the investigator and TB programme experts from the SNNPRS Health
Bureau. The six months performance of the community-based TB care was conducted in
Yirgalem.
28
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Definition of terms
TB Diagnosis, classification, case definitions and treatment outcomes were dealt with in
Failure: refers to a patient who remains or becomes smear positive at 5th month or later.
treatment and at least on the previous occasion (at 2nd or 5th month).
treatment but does not have smear result at 7th month of treatment or does not fulfil the
Case detection rate: is the number of smear-positive patients detected of the estimated
Treatment success rate: is the number of TB patients cured or treatment completed of the
completed in the past with or without smear positive result. It also included relapse cases.
The main study outcomes were CDR, TSR, proportion of HIV infected, mortality,
TB patients were classified into smear-positive, smear-negative and EPTB cases based on
the smear result and the site involved. The treatment outcomes were cured, treatment
29
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
completed, default, failure, transfer out and died (appendix I & II). In Paper I, CDR and
by WHO [77].
In Paper V & VI, recurrence and mortality were used to measure the post-treatment
condition of the TB patients. In Paper II, HIV test results were done using ELISA test. A
societal perspective of cost estimation was used to find out the average cost per patient
For the community randomized trial, we used the CDR of 41% (an average of previous
three years from 2003 - 2005 for the study area) [additional results -1, table - 1]. We
estimated the number of clusters needed on the basis that community-based case finding
to increase the CDR by 30% using power of 80%, 95% confidence interval and
accounting for 30% loss to follow-up. Based on the principle of allocating unequal
groups. TB patients who received treatment in the study area were included in the cost-
effectiveness study (Paper III & IV). As indicated in the individual papers, we enrolled
and analyzed the available data as a whole and the number of patients evaluated was large
Microsoft Excel and SPSS for Windows 14 were used for data entry and analyses.
Independent t-test and one way analysis of variance were used to compare the mean CDR
and TSR for cluster level values and to determine the intraclass correlation coefficient
30
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
(Paper III), respectively. Average cost of treating smear-positive patients was divided by
the number of patients treated successfully. One way sensitivity analysis was done to
Logistic regression analysis was used to estimate the effect of predictor variables on the
rate of HIV infection among TB patients and pregnant women. Linear regression analysis
was used to estimate the amount of variation explained by predictor variables (Paper II).
Kaplan-Meier and Cox Regression method were used to evaluate event free (death or
recurrence) survival and the relative effects of selected variables, respectively. Log rank
test and hazard ratios were used for statistical significance. We calculated SMR using
The Ethical Review Committee of the SNNPRS Health Bureau approved the studies. In
consultation with the NTLCP, discussion was held with TB programme experts at zone,
discussion was held with kebele leaders and we obtained community consent. Enrolment
of the study participants was done after obtaining informed consent from individual study
participants. The participants were also informed about the right to withdraw from the
study without compromising their future care. PTB suspects that were smear-negative for
acid fast bacilli were given free antibiotic treatment as part of diagnostic work up
recommended by the NTLCP. In Paper II, study participants who wanted to know their
31
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
HIV status were advised to visit the voluntary counselling and testing unit in the same
32
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Region of Ethiopia
districts and health facilities. Monitoring and evaluation is one of the components of the
DOTS strategy to understand its performance. We aimed to find out the effectiveness
The result of the study was based on the official reports of TB control programme over
ten years. The diagnosis and treatment, case notification and treatment outcome reports
In 2004, 94% of the health facilities (hospitals, health centers and health stations)
implemented the DOTS strategy. 136 572 cases were registered in ten years; of these,
47% were smear-positive, 25% were smear-negative and 28% had EPTB. The smear-
positive case notification rate increased from 45 to 143 per 105 population. Similarly, the
CDR increased from 22% to 45%, and the TSR from 53% to 85%. The defaulter rate
Decentralization of DOTS strategy improved the case detection and treatment success of
TB patients. TB control programme achieved 85% treatment success; however, with the
current low CDR (45%), the 70% WHO target seems unachievable in the absence of
33
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
4.2. Paper II: The rate TB-HIV Co-infection Depends on the Prevalence of HIV
Infection in a Community
Limited knowledge about the rate of HIV infection in TB patients and the general
population compromises the planning, resource allocation and prevention and control
activities. We aimed to determine the rate of HIV infection in TB patients and its
correlation with the rate HIV infection in pregnant women attending ANC.
TB patients and pregnant women attending health facilities were enrolled in 2004 - 2005.
TB diagnosis, treatment and HIV testing were done as per the National guideline. Logistic
regression and linear regression analysis were used to determine the risk factors and the
Of the 1308 TB patients enrolled, 226 (18%, 95%CI: 15.8 - 20.0) were HIV positive. The
rate of HIV infection was higher in TB patients from urban (25%) than rural areas (16%)
[AOR = 1.78, 95%CI: 1.27- 2.48]. Of the 4199 pregnant women, 155 (3.8%, 95%CI: 3.2 -
4.4) were HIV positive. The rate of HIV infection was higher in pregnant women from
urban (7.5%) than rural areas (2.5%) [OR = 3.19, 95% CI: 2.31- 4.41]. In the study
participants attending the same health facilities, the rate of HIV infection in pregnant
women correlated with the rate of HIV infection in TB patients (R2 = 0.732).
The rate of HIV infection in TB patients and pregnant women was higher in urban areas.
The rate of HIV infection in TB patients was associated with the prevalence of HIV
34
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
4.3. Paper III: Health Extension Workers Improve Tuberculosis Case Detection and
Early case finding and prompt treatment of smear-positive cases is at the centre of DOTS
strategy. Unfortunately, the CDRs remain low in many countries. We aimed to find out if
Ethiopia.
southern Ethiopia. HEWs from the intervention kebeles were trained on how to identify
230 smear-positive patients were identified from the intervention and 88 smear-positive
patients from control kebeles. The mean CDR was higher in the intervention than in the
control kebeles (122·2% vs. 69·4%, p < 0·001). More females were identified in the
intervention kebeles (149·0% vs. 91·6%, p < 0·001). The mean TSR was higher in the
intervention than control kebeles (89·3% vs. 83·1%, p = 0·012) and for females (89·8%
Involving HEWs in sputum collection and treatment improved smear-positive CDR and
TSR possibly because of an improved access to the service that reduced socioeconomic
burden on TB patients. This could be applied in settings with low health service coverage
35
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Trial
conditions have affected the already overstretched health system. Therefore, alternative
strategies that increase the effectiveness of identifying and treating TB patients at lower
cost are required. We present the cost and cost effectiveness of involving HEWs in TB
treatment.
Comparison of two treatment options, DOT by HEWs and GHWs was done along a
community randomized trial. Costs were analyzed from societal perspective in 2007 in
US $ using standard methods. Cost-effectiveness was calculated as the cost per smear-
The total cost per successfully treated smear-positive patient was higher in health facilities
($161.9) compared to the treatment in the community ($60.7). Community DOT (CDOT)
reduced the total, patient and caregiver cost by 62.6%, 63.9% and 88.2%, respectively.
The cost of involving HEWs ($8.8) was 14.3% of total cost per patient for CDOT.
patients and caregivers. There is an economic and public health reason to involve HEWs
in TB treatment in Ethiopia. However, due attention should be paid to ensuring initial start
36
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Decentralization of DOTS has increased the number of cured smear-positive patients after
completing treatment. However, the rate of recurrence has increased mainly due to HIV
We retrospectively enrolled smear-positive patients who were reported cured from 1998
of observation (PYO) were calculated per 100 PYO from the date of cure to date of
regression methods were used to determine the survival and the hazard ratio (HR).
368 cured smear-positive patients cured under DOTS were followed for 1463 person-
years. Of these 15 smear-positive patients had recurrence. The rate of recurrence was 1
per 100 PYO (0.01 per annum). Recurrence was not associated with age, sex, occupation,
High recurrence occurred among smear-positive patients cured under DOTS strategy.
Further studies are required to identify factors contributing to high recurrence rates to
37
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Tuberculosis control programme aims at identifying the highly infectious TB cases and
patients.
We retrospectively enrolled TB patients who were treated and reported cured or treatment
completed from 1998 to 2006. Mortality was used as an outcome measure. Person-years
of observation (PYO) were calculated per 100 PYO from the date of completing treatment
to date of interview if the patient was alive or to date of death. Kaplan-Meier and Cox-
regression methods were used to determine the survival and the hazard ratio (HR).
Indirect method of standardization was used to calculate the standard mortality ratio
(SMR).
725 TB patients were followed for 2602 person-years. 91.1% (659 of 723 patients) were
alive while 8.9% (64 of 723 patients) had died. The mortality was 2.5% per annum. Sex,
age and occupation were associated with high mortality. More deaths occurred in non-
The morality was high in TB patients compared with the general population. More deaths
occurred in non-farmers, men and elderly. Further studies are required to identify the
38
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Additional results
To enrich the understanding of the context and the significance of the studies, some
important data that were not part of the individual studies were added here. These are the
baseline data, the duration of presentation for seeking diagnosis in the community, cases
identified, cost of seeking diagnosis in public health facilities and treatment outcome of
Table 1. CDR and TSR of smear-positive cases in the study area, 2003 - 2005
Control kebeles 35% 53% 40% 43% 86% 91% 82% 86%
Intervention kebeles 23% 50% 42% 38% 80% 87% 84% 84%
*
3 years average TSR for smear-negative and EPTB cases was 77% and 79%, respectively.
2. Duration of cough on presentation: in the intervention kebeles, of the total 723 PTB
suspects who produced sputum for examination, more females (65%) were enrolled than
males (35%). Most of the suspects (75%) visited health post within 2 - 4 weeks of the
onset of cough. The duration of cough was in the range of two weeks to six months in the
first month of the intervention. It decreased as the sputum collection continued from the
39
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Table 2. Smear-positive case notification rates in the study area, 2006 - 2007
diagnosis in public health facilities for smear-positive patients from control kebeles
accounts for the loss of about 10 working days or US$ 12.9 [10% of the GDP in
2007($130)], at its least estimate. This did not include the waiting time in the health
EPTB cases, 171 cases were from the intervention while 94 cases were from control
kebeles. Of these, 75% (128/171) from intervention and 42% (39/94) from control kebeles
were treated successfully. The TSR was higher in the intervention than control kebeles for
smear-negative cases (p-value = 0.01). However, the TSR for EPTB cases was higher in
40
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
the control than intervention kebeles (p-value = 0.004). This was due to six deaths in
Table 3. Treatment outcome of smear-negative and EPTB cases in the study area,
2006 - 2008
* **
TB Groups Number of cases TSR ICC
*
TSR - treatment success rate **ICC - intraclass correlation coefficient
NB: the TSR was adjusted for clustering under the two treatment options.
41
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Table 4. Comparison of TB cases enrolled and lost to follow-up in the study area,
1998 - 2006
*
df = 92.3 mean difference = 0.28, † SD = standard deviation, ‡others = relapse and transfer-in
‡‡
More cases were lost to follow-up among treatment completed (missed sputum examination at 7th month)
42
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
5.0. Discussion
The main study design was community-randomized trial in which intact social units or
group of individuals are assigned to intervention or control groups. This study design is
when an intervention involves groups is applied at the level of health organizational units
administer and when reduction of contamination is required [103-106]. This design may
be more important in developing countries, particularly in rural areas where the sense of
community is strong, and community consent and cooperation are essential [107]. We
conducted our study in two rural districts (Dale and Wonsho) at programme (health
organizational unit) level using health workers in the districts (Paper III).
Cluster randomization trials lack the independence of observations and violate the main
assumption of statistics. This influences the design and analysis; the standard approaches
to sample size estimation and analysis no longer apply. Moreover, during recruitment,
clusters may withdraw (in our case HEWs in a cluster may not recruit participants) which
may lead to empty clusters. However, the shortcomings may be overcome by randomized
allocation, increasing number of clusters (small sized clusters are more efficient), follow-
up during the intervention and adjusting for clustering during analysis. Randomization
ensures similar distribution of known or unknown sources of bias except for a chance or a
43
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
In Paper III, we completely randomized the clusters using table of random numbers,
adjusted for clustering and used intention-to-treat analysis that is a pragmatic approach.
We did not have cluster level withdrawal but had three empty clusters one from
intervention and two from control groups which reflects what happens in practice under
programme conditions. However, this is less likely to affect the results of our study due to
fact that we added thirty percent of calculated clusters as a contingency to compensate for
loss to follow-up. The conduct of the study requires two level ethical consideration due to
consent at the two levels, from the community leaders and the patients.
Cohort studies can be thought as natural experiments in which outcomes are measured in
realistic setting [111]. Cohort studies follow two or more groups from exposure to
outcome. It can be done ahead in time from present (prospective), in the opposite
records of past exposure for a group of people who can be traced to find out their current
status [112]. In Paper V & VI, TB patients who were successfully treated under DOTS
strategy were retrospectively followed-up. The list of patients was obtained from unit TB
registers in health facilities and this was crosschecked with the list of patients in the
district TB register and the quarterly reports. Then the patients were traced to their place
of residence.
Cohort studies are useful to find out incidence and natural history of a disease. They allow
estimation of incidence rates, relative risks and other outcome measures using survival
methods. It also reduces the risk of survivor’s bias in conditions that are rapidly fatal [43,
44
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
112, 113]. The design is appropriate for studying rare exposures and helps to measure
multiple outcomes that might follow an exposure. Selection bias and loss to follow-up can
be a problem. In Paper V & VI, patients were followed to the place of their residence by
house-to-house visit and the degree of loss to follow-up was minimal (less than seven
percent). Cohort studies by nature are time consuming and expensive. However, the
benefit: cost ratio of efficient cohort study is high and retrospective cohort studies
generally reduce cost [43, 112]. In Paper V & VI, we believe that the cost of conducting
the studies was low for we involved HEWs, used unit TB registers at health facilities and
casual risk factors (analytic) and captures its prevalence. It is useful to generate
hypothesis, to examine unchanging exposures (sex, blood group) that occurred many
years back. However, cross-sectional studies have in built ‘chicken or egg’ dilemma and
selection bias [114]. We enrolled all TB patients and pregnant women available during the
study period. Therefore, selection bias is less likely to affect the results.
The results of any research are only as good as the data upon which they are based.
However, data may be affected by the study participants, instruments used, people’s
perfect except to minimize errors as far as possible, and then assess the practical effects of
any unavoidable error [115]. The validity of epidemiological study, therefore, depends on
the study design, the study conduct and data analysis [112]. In relation to the population
45
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
to which the conclusion is drawn, the term validity refers to two population groups, the
Internal validity
Internal validity refers to the accuracy of measuring what the study is designed to measure
in the study participants or refers to the extent to which the results of the study reflect the
true situation in the study sample. It is natural to take delight in interesting findings.
However, the findings of a study could be explained by other facts than the study itself. It
could be due to chance, bias and confounding that should be ruled out through closer
evaluation of the study design, the selection of the study participants and the data analysis.
The possible alternative explanations for study results are briefly discussed below.
It refers to random error or the probability that variability in sampling explains the
observed result. The role of chance is measured by conducting the test of statistical
adequate sample size [115, 116]. In Paper I, II, V & VI, we consecutively enrolled TB
that randomization could take care of the chance occurrences of outcomes under study
[102, 104]. In Paper III & IV, due to the random allocation of adequate number of clusters
46
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
This is a systematic distortion that results from procedures used to select subjects and
from factors that influence study participation. This occurs when study participants are
selected inappropriately, using different criteria, or upon prior knowledge about their
exposure or outcome. Selection bias could also arise due to self-selection of volunteers,
response rate or loss to follow-up and enrolling healthy workers. It should be considered
and reduced in the design and conduct of a study by using a clearly defined eligibility
In Paper I, TB patients reported over ten years were included in the study. It is possible
that poor outcomes might not be reported. However, the treatment outcome data was
crosschecked against the previously reported case finding which was used to calculate the
treatment outcomes. In addition, only about 4% of all TB cases were not evaluated for
which incomplete recording could be one of the reasons [118]; in one of the zones in the
region, as high 17.5% of TB cases were not evaluated and treatment was not registered for
them[119]. TB patients were prospectively (Paper III & IV) and retrospectively (Paper V
& VI) followed to their place of residence. In Paper V & VI, the list of patients was
obtained from the TB unit registers available in the health facilities where the patients
received treatment; in this case the registration was less likely to be affected by the
outcomes of the cases. In Paper II, we consecutively enrolled all TB patients and
pregnant women available during the study period. Therefore, it is less likely that the
results are affected by selection bias. However, ecological fallacy could be a problem for
we used aggregate data of few study sites to estimate the correlation (Paper II).
47
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
In intervention studies, participants lost to follow-up and protocol deviation can occur at
the level of the cluster (cluster withdrawal or lost to follow-up or inactive cluster) or the
analysis [110]. In Paper III and IV, there was no cluster level withdrawal and individual
level loss to follow was minimal to affect the comparability of the groups. Moreover,
analysis using cluster as unit of analysis and cluster level values maintains comparability
of the groups.
In Paper V, we found that no baseline difference between the cases enrolled in the study
compared to those who were lost to follow-up and moved to other places except in the
difference in treatment outcome (additional results - 6, table - 4). However, there was no
difference in mortality and recurrence among those who were reported cured or treatment
It is a systematic error that results from systematic differences in the way exposure or
outcome data are obtained. This can be reduced by using eligibility criteria, defined
exposure or outcome, objective and structured questionnaire and maintain blinding of the
using written protocol, administering the data collection under uniform conditions,
conducting regular supervision and quality check could reduce information bias [117].
48
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
In our studies, standard case and treatment outcome definitions, laboratory examination
method, registers and reporting formats were used. GHWs and laboratory technicians
were trained. CHRL prepared reagents for sputum examination and distributed to the
districts. Slides were collected for quality control to assure the quality of the tests as per
the recommendation of NTLCP (CHRL reported 98% concordance with the slides from
the peripheral laboratory in the study area). In addition, TB control programme experts
and the investigator supervised the health facilities to follow the conduct of the studies.
The data about rediagnosis was confirmed by cross-checking the report against unit TB
registers in health facilities (Paper V). The diagnosis of smear-negative and EPTB had a
component of subjective decision by the clinicians. However, the use of uniform training
materials and case definitions based on the NTLCP guideline and working under similar
setting (GHWs and the health facilities providing diagnostic service) could reduce the
during the study period. The travel time and related costs were cross-checked against the
travel distance and existing market price by the HEWs and GHWs who know the
estimates of travel time and related costs. The fact that we did not give incentives or
refund the expenses makes the estimate reasonable for the travelled distance and related
expenses. Thus, we believe that the measurements were less prone to bias to significantly
49
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Confounding
It refers to a mixing of effects that can occur when the results of a study are confused by
effect of a third factor that is associated with both the exposure and the outcome but not
distributed unevenly across study groups and can lead to either overestimation or
underestimation of the effect size, completely hide and in extreme cases reverse the
trials if they are small. In such cases, increasing the size of the study does not make any
In our studies, randomization (Paper III & IV), multiple logistic regression (Paper II) and
Cox proportional hazard regression (Paper V & VI) were done to control for confounders.
Randomization deals not only with confounders that are known and can be measured but
also with other unrecognizable or unmmeasurable confounders in the study group that
makes the results less likely to be affected by confounding except for the play of chance.
In addition, in the analysis, we stratified the data by age, sex, residence, TB classification
and category (Paper II) and by sex, level of education, marital status, occupation and TB
classification (Paper V) to control for confounders in the analysis. However, in Paper II,
using aggregate data made it difficult to control for confounding factors. Generally, if
extrapolated to the study and general population [121] while controlling for cofounding
ensures the internal validity of the study, a prerequisite for external validity.
50
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
It refers to the applicability of the results to the people outside the study population. The
applicability of the study results outside the study setting depends on the feasibility of
conducting it under routine care and its acceptability [122]. In addition, using broader
eligibility criteria increases the applicability of the findings to a wider population [123].
The results of Paper III could be applied in settings with low health service coverage (low
DOTS coverage and limited number of TB laboratories), where HEWs have the first
contact with the people to provide health education, and collect and transport sputum
specimens and provide DOT. It is considered that, this makes the service patient-centred
and improve the case finding and treatment adherence [124]. The study area is a densely
populated agrarian community, typical of the rural population on the Ethiopian highlands.
It could also be applied in areas with a shortage of health workers, especially laboratory
In the studies, we used the existing health service (health centres and health posts) and
health workers (GHWs and HEWs) within the policy provision of community-based
initiative. The study employed the routine TB care and decentralized service to the
community without requiring the patients to make extra visits for the purpose of the study.
This reduced the distance travelled by the patients and caregivers, travel time and related
expenses. This might have increased the time for productivity by caregivers and possibly
by the patients.
51
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
We believe that it was acceptable by the community as the patients received treatment
under direct observation of HEWs (a member of their kebele) at health post in the kebele
(the closest health facility to their living place) and possibly culturally acceptable as the
patients did not travel out of their kebeles and were treated by HEWs who understand the
culture and speak their language. Moreover, the improved TSR for smear-negative and
EPTB add to the applicability of the intervention to all forms of TB (additional results - 5,
table - 3). However, TB patients should be carefully evaluated for their general condition
compared to the benefits of improving case finding, reducing disease transmission and
improving the treatment outcome; and cost implication to the patients, households and the
health service, the researchers strongly believe that it is valid to apply to broader
population. The results of the study could be applied in settings with limited health
service coverage and shortage of health workers including laboratory technicians, which
learn from the experience and improve its performance by dealing with emerging
challenges.
to the service and increased the number of cases identified and treated. In our study area,
52
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
involving HEWs in TB control improved the case detection and treatment success rates. It
was also economically attractive for the patients, the household and the health service.
This is one of the ways to overcome the challenges of the TB control programme in
settings with low health service coverage and shortage of health workers in the face of
HIV epidemic.
The introduction of DOTS strategy, in response to the increasing global burden of TB, is
an important landmark in the history TB control. DOTS strategy aims to detect 70% of
incident smear-positive cases and cure 85% of them through early case finding and
decentralization to peripheral health facilities and later involved all care providers and the
improved the TSR of smear-positive cases. However, its effectiveness in CDR was
limited mainly due to low health service coverage, shortage of health workers, HIV
epidemic, MDR TB and inadequate involvement of the available care providers [19, 124-
TSR from 53% to 85%. However, its effectiveness in improving the CDR was limited; it
increased from 22% to 45%, far below the target to effectively reduce the incidence of
smear-positive cases [14]. Similarly low CDR was reported in a study conducted in rural
districts of southern Ethiopia. However, the possible explanations given were treatment of
53
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
the incidence and prevalence of TB or low disease burden in the study area [119].
However, our findings suggested that backlog of cases were not adequately reached by the
DOTS strategy to reduce the incidence and prevalence of TB due to the prevailing low
Moreover, TB diagnosis and treatment was limited to health centres and hospitals. In such
cases, TB suspects and patients travel long distance; have out of pocket expense and loss
awareness about TB in the community. However, this can also have impact on case
finding if optimally functioning diagnostic and treatment facilities are available within the
alternatives that reduce the cost of the health service and the community to access the
service. The contributions of the community health workers specifically HEWs in our
element of public health interventions and primary health care [128]. After long silence,
the following reasons called for renewed interest to look for cost-effective alternatives to
did not increase the access to the services as expected mainly because of inadequate
services and shortage of health workers and resources. This was worsened by the
54
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
increased burden of TB that overstretched the existing health services in many countries
[69, 129].
In Ethiopia the main challenge for NTLCP continues to be low CDR. Attempts to increase
CDR through deploying HEWs and training GHWs to identify and refer suspects has not
yet demonstrated an increase in CDR. This suggests the existence of problems in the
health system that needs alternative approaches to improve the programme performance
[130].
TB diagnosis depends on examination of patients that visit health facilities. However, the
awareness about TB, socioeconomic status, culture, access to the health service, the
quality of the service and the interplay among these factors affect the health seeking
diagnosis was primarily considered to be the role of GHWs. However, evidences show
geographic expansion DOTS strategy to have less effect on improving case finding [125].
Similarly, in our setting (Paper I), DOTS expansion (treatment and diagnostic units)
improved the TSR while the CDR remained far below the target.
To improve the CDR in our setting, the role of community in TB case finding that
involves activities such as raising awareness of TB, advocacy for adequate resources and
proper care, identification and referral of TB suspects and supporting patients during
treatment was revisited [14]. In Paper III, we involved HEWs in identifying PTB suspects,
sputum collection and transportation to diagnostic units and administration of DOT. This
55
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
has successfully increased the CDR in the intervention group (122%) compared to the
control group (69%). The main reasons were increased access to diagnostic facilities that
was created (by sputum collection in the kebeles and its transportation to diagnostic units
by HEWs). The intervention improved geographic and economic access to the rural
community. In addition, more women suspects were examined (additional results - 2) and
more smear-positive cases were detected among them. This could be one of the
explanations why less women TB cases are reported in DOTS strategy. We believe that
our intervention addressed socio-cultural and economic factors that affected the heath
seeking behaviour of women. In addition, for every smear-positive case detected in health
attractive the case fining is the adherence of the patients to treatment and the cost
Community-based DOT
DOT is only one range of measures aimed at promoting treatment adherence. Thus, the
fight against TB, a scourge to humanity, requires a concerted effort of the health system,
care providers and the community. This presupposes a patient-centred service that is
easily accessible, convenient and acceptable without compromising the treatment success.
To this end a range of community members, TB patients, families, community leaders and
CHWs were involved in providing DOT. Community DOT was found to be as effective
as the health facility based DOT [126, 131-133]. The study (Paper III) has shown
community DOT - treatment by HEWs to be more effective than the health facility DOT
in our setting. The TSR was 89% in the community and 83% under health facility DOT.
In addition, the rate of defaulter, failure and transfer out was low in the community-based
56
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
approach. This could be mainly due to the improved geographic and socioeconomic
Cognizant of the argument that treatment success should be improved prior to increasing
the case detection, except for improving the TSR the effectiveness of CHWs in improving
the CDR was limited [133, 134]. This could be due to the fact that the role of CHWs was
limited to improving treatment adherence and providing health education. Is this a failure
of DOTS strategy? [127] In our study, we have explored option of involving HEWs in
sputum collection and improved the CDR. However, the cost-effectiveness of involving
use in situations where a decision maker, operating with a given budget, is considering a
Studies show that involving CHWs in providing DOT is a cost-effective alternative for
economic reasons for the patients, household and health service. Compared to health
facility DOT, community DOT is effective without compromising the TSR [135-138].
But, its cost-effectiveness varied with the type and the role of the CHWs involved in
community-based approaches, the health service coverage and other related factors [139-
144]. In Paper IV, the cost per successfully treated smear-positive patient in the
intervention area ($61) was lower compared to health facility DOT ($162). It reduced the
cost per successfully treated smear-positive case by 63%. For the same amount of cost of
57
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
treating a patient in health facilities more than two patients could be treated in their
communities. This was due to the decentralization of the service and reduced the travelled
In addition, the high cost incurred for seeking diagnosis in public health facilities ($13,
about 10% of GDP loss or loss of 10 working days) could also be one of the reasons for
low CDR at its least estimate excluding the cost related to visits made to other health
facilities and traditional healers ( additional results 4). Other studies reported that the cost
of seeking diagnosis to be prohibitively high for the patients and their household [47-51].
In this study, stronger belief is held that community-based sputum collection has reduced
the cost of seeking diagnosis in public health facilities due to the reduced travel distance,
limited due to high turnover and/or exhaustion of the CHWs, additional cost incurred to
integrated into the general health service in such a way that it complements the existing
In Paper III & IV, involving HEWs improved the CDR and TSR and was cost-effective
compared to the health facility DOT. The health workers involved in our study operated
under the provision of the health policy of the country, employed and received salary and
supervised by public health professional in health centre. In addition, two HEWs were
58
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
assigned in each kebele and training schools deploy new graduates to address the attrition
Effective disease control strategy should be efficient in identifying the cases early,
providing prompt treatment and ensuring long term disease free survival of the patients.
Therefore, understanding the long term efficacy of the DOTS strategy in reducing the rate
MDR TB and its effectiveness in the face of HIV epidemic and socioeconomic
In Paper V & VI, we found high recurrence rate and mortality among successfully treated
TB patients. The plausible explanations were HIV infection, MDR TB and possibly
continued transmission due to high backlog of untreated cases in the community. In Paper
II, the prevalence of HIV infection in the community was 3.8% and the rate of TB-HIV
co-infection was 17.5%. This is high enough to affect the survival of TB patients after
successful treatment.
treatment. The mortality was higher in non-farmers (SMR=9.95, 95%CI: 7.17 - 12.73).
This could also be explained by HIV infection and to some extent by prevalence of MDR
TB. Thus, TB control programmes should strengthen patient follow-up, provide HIV
counselling and testing for TB patients during follow-up period and after successful
59
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
counselling and testing for TB patients that is performing in limited health institutions.
Less than five percent of TB patients were tested for HIV during the study period which
has now increased by three fold [21]. Further study is needed to estimate the prevalence of
Strategies that enhance TB-HIV collaboration activities are needed in such settings. All
TB patients should be encouraged to be tested for HIV and enrolled for antiretroviral
should also be advised to reduce their risk and be encouraged to live HIV free life. The
health facilities should organize their facilities in such a way that it encourages simple
In addition, the role of the community and HEWs in prevention and control of the two
treatment. This will increase the number of patients accessing and receiving the service to
benefit from the available treatment that will contribute to the long term efficacy of DOTS
strategy.
the existing health system in Ethiopia. The recommendations of our study were taken up
by the Federal Ministry of Health of Ethiopia. The SNNPRS Health Bureau required
60
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
decision making for policy change to make sure that the benefits of the study were shared
We therefore recommend further larger scale intervention to strengthen the health system
approaches based on evidences from the field. In such cases, health system strengthening
should also be part of the intervention where the GHWs and HEWs will receive adequate
training and necessary resources made available. The service should then be progressively
scaled up and based on the lessons learned should be used to improve the model of the
impact and quality of the community-based approach to improve the role of HEWs and
The NTLCP has recognized the role of HEWs in TB control and decided to start case
finding and treatment by HEWs. The WHO in collaboration with the NTLCP has started a
and decentralization of treatment to health posts under the direct observation of HEWs in
four big regions of the country. Currently the Federal Ministry of Health of Ethiopia has
at larger-scale.
61
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
6.1. Conclusions
2. The rate of TB-HIV co-infection is high and is associated with the prevalence of
3. Beyond achieving the treatment success rate of 85%, the recurrence and mortality
rate was high in successfully treated TB patients. Therefore, identifying the causes
positive TB cases improved the CDR and TSR. More women smear-positive cases
treated TB cases for the same amount of cost under health facility DOT. This is
economically attractive option for the patients, households and the health service.
6.2. Recommendations
1. GHWs should be trained about identifying TB cases that require closer follow-up
in health centres and hospital before referring them for treatment in the community
2. HEWs should be trained about identifying seriously ill TB patients and patients
with side effects that need referral and follow-up in health centres and hospitals.
62
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
3. GHWs and HEWs should inform TB patients to be tested for HIV during
treatment and even after completing treatment to benefit from the available care.
2. Routine TB data reported from districts to higher levels should be analysed and
districts.
service to the community. It should be clearly outlined and practiced at all levels.
burden and the impact of the interventions on the incidence and prevalence of
smear-positive TB
63
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
6.2.4. Policy
1. The role of HEWs in TB case finding and treatment should be clearly stated and
2. In collaboration with the HEP, the NTLCP should incorporate the role of HEWs in
diseases burden and save lives, worth adopting in resource constrained settings
64
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
References
1. Bates JH: Transmission and pathogenesis of tuberculosis. Clin Chest Med 1980,
1(2):167-174.
3. Harries AD, Dye C: Tuberculosis. Ann Trop Med Parasitol 2006, 100(5-6):415-
431.
5. Loudon RG, Spohn SK: Cough frequency and infectivity in patients with
7. Harries AD: Robert Koch and the discovery of the tubercle bacillus: the challenge
of HIV and tuberculosis 125 years later. Int J Tuberc Lung Dis 2008, 12(3):241-
249.
8. El-Sony AI, Mustafa SA, Khamis AH, Sobhi S, Enarson DA, Baraka OZ, Bjune
McAdam KPWJ: Clinical and radiological presentation of 340 adults with smear-
positive tuberculosis in The Gambia. INT J TUBERC LUNG DIS 2003, 7(10):942–
947.
65
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
12. Crawford JT, Eisenach KD, Bates JH: Diagnosis of tuberculosis: present and
13. Perkins MD, Kritski AL: Diagnostic testing in the control of tuberculosis. Bull
15. Frieden TR, Sterling TR, Munsiff SS, Watt CJ, Dye C: Tuberculosis. Lancet 2003,
362(9387):887-899.
18. Frieden TR: Can tuberculosis be controlled? Int J Epidemiol 2002, 31(5):894-899.
19. Dye C, Hosseini M, Watt C: Did we reach the 2005 targets for tuberculosis
20. Corbett EL, Watt CJ, Walker N, Maher D, Williams BG, Raviglione MC, Dye C:
The growing burden of tuberculosis: global trends and interactions with the HIV
66
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
2010.
183.
24. Murray JF: A century of tuberculosis. Am J Respir Crit Care Med 2004,
169(11):1181-1186.
25. Murray JE: The white plague in Utopia: tuberculosis in nineteenth-century Shaker
26. World Health Organization, The global burden of disease 2004 update. Geneva,
Switzerland; 2008.
27. Sutherland I, Fayers PM: The association of the risk of tuberculous infection with
28. Ottmani SE, Uplekar MW: Gender and TB: pointers from routine records and
29. Borgdorff MW, Nagelkerke NJ, Dye C, Nunn P: Gender and tuberculosis: a
30. Ahsan G, Ahmed, J., Singhasivanon, P., Kaewkungwal, J., Okanurak, K.,
Suwannapong, N., Akarasewi, P., Majid, M. A., Begum, V., Belayetali, K.:
67
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
35(1):126-135.
31. Karim F, Islam MA, Chowdhury AM, Johansson E, Diwan VK: Gender
32. Boeree MJ, Harries AD, Godschalk P, Demast Q, Upindi B, Mwale A, Nyirenda
33. Hamid Salim MA, Declercq E, Van Deun A, Saki KA: Gender differences in
2004, 8(8):952-957.
34. Agboatwalla M, Kazi, G. N., Shah, S. K., Tariq, M.: Gender perspectives on
35. Uplekar MW, Rangan S, Weiss MG, Ogden J, Borgdorff MW, Hudelson P:
Attention to gender issues in tuberculosis control. Int J Tuberc Lung Dis 2001,
5(3):220-224.
tuberculosis treatment in the Gambia, west Africa. Int J Tuberc Lung Dis 2004,
8(1):70 - 75.
37. Holmes CB, Hausler H, Nunn P: A review of sex differences in the epidemiology
68
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
38. Connolly M, Nunn P: Women and tuberculosis. World Health Stat Q 1996,
49(2):115-119.
39. Allotey P, Gyapong M: Gender in tuberculosis research. Int J Tuberc Lung Dis
2008, 12(7):831-836.
40. The social roots of urban tuberculosis. Indian Med Trib 1994, 2(11):3.
involving rural and urban population in Tamil Nadu, South India. Trop Med Int
42. Cambanis A, Yassin MA, Ramsay A, Bertel Squire S, Arbide I, Cuevas LE: Rural
43. Grimes DA, Schulz KF: Cohort studies: marching towards outcomes. Lancet
2002, 359(9303):341-345.
Organization; 2005.
45. Caminero JA: A Tuberculosis Guide for Specialist Physicians. Paris: IUATLD;
2004.
69
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
tuberculosis at the household level. Int J Tuberc Lung Dis 1999, 3(7):596-602.
48. Croft RA, Croft RP: Expenditure and loss of income incurred by tuberculosis
patients before reaching effective treatment in Bangladesh. Int J Tuberc Lung Dis
1998, 2(3):252-254.
49. Liu X, Thomson R, Gong Y, Zhao F, Squire SB, Tolhurst R, Zhao X, Yan F, Tang
analysis from community and tuberculosis patient perspectives. Trop Med Int
50. Kemp JR, Mann G, Simwaka BN, Salaniponi FM, Squire SB: Can Malawi's poor
afford free tuberculosis services? Patient and household costs associated with a
585.
51. Lonnroth K, Tran TU, Thuong LM, Quy HT, Diwan V: Can I afford free
with tuberculosis in Ho Chi Minh City, Vietnam. Soc Sci Med 2001, 52(6):935-
948.
40(2):116-122.
70
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
interactions, disease progression, and survival. Mem Inst Oswaldo Cruz 1996,
91(3):385-387.
54. McShane H: Co-infection with HIV and TB: double trouble. Int J STD AIDS 2005,
rate in tuberculosis patients in Malawi. Int J Tuberc Lung Dis 2001, 55(7):1000 -
1005.
56. Harries T: TB control programmes: the challenges for Africa. Afr Health 1996,
19(1):19-20.
60. Corbett EL, Marston B, Churchyard GJ, De Cock KM: Tuberculosis in sub-
doing with current control efforts? Int J Tuberc Lung Dis 2005, 9(1):17-24.
Africa: impact on patients and programmes; implications for policies. Trop Med
71
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
63. Frieden TR, Lerner BH, Rutherford BR: Lessons from the 1800s: tuberculosis
64. Raviglione MC, Pio A: Evolution of WHO policies for tuberculosis control, 1948-
65. Murray JF: [Historical development of tuberculosis since Robert Koch's discovery
66. Bates JH, Stead WW: The history of tuberculosis as a global epidemic. Med Clin
68. Mahler H: The tuberculosis programme in the developing countries. 1966. Bull
71. Miller B: Health sector reform: scourge or salvation for TB control in developing
72. Bosman MC: Health sector reform and tuberculosis control: the case of Zambia.
73. TB. A global emergency, WHO, July 1994. Lepr Rev 1995, 66(3):270-271.
Tuberculosis. The Hague, March 15, 1991. Bull Int Union Tuberc Lung Dis 1991,
66(4):139-214.
72
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
75. Tuberculosis and Leprosy Prevention and Control Manual, Ministry of Health,
76. Dye C, Garnett GP, Sleeman K, Williams BG: Prospects for worldwide
tuberculosis control under the WHO DOTS strategy. Directly observed short-
79. World Health Organization. The Stop TB Strategy: building on and enhancing
81. Ministry of health of Ethiopia: Guidelines for health Extension programme. Addis
Ababa; 2007.
83. Federal Ministry of Health of Ethiopia. Essential health Service Package for
2007, 5:21.
73
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
86. Yassin MA, Datiko DG, Shargie EB: Ten-year experiences of the tuberculosis
control programme in the southern region of Ethiopia. Int J Tuberc Lung Dis
2006, 10(10):1166-1171.
88. Wyss K, Kilima P, Lorenz N: Costs of tuberculosis for households and health care
providers in Dar es Salaam, Tanzania. Trop Med Int Health 2001, 6(1):60-68.
urban Zambia: the economic impact and burden on patients prior to diagnosis. Int
90. Sherman LF, Fujiwara PI, Cook SV, Bazerman LB, Frieden TR: Patient and
health care system delays in the diagnosis and treatment of tuberculosis. Int J
91. Kiwuwa MS, Charles K, Harriet MK: Patient and health service delay in
93. Tekle B, Mariam DH, Ali A: Defaulting from DOTS and its determinants in three
districts of Arsi Zone in Ethiopia. Int J Tuberc Lung Dis 2002, 6(7):573-579.
74
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
4(2):e37.
96. Wakabi W: Extension workers drive Ethiopia's primary health care. Lancet 2008,
372(9642):880.
97. Crisp N, Gawanas B, Sharp I: Training the health workforce: scaling up, saving
98. Epidemiological fact sheet on HIV AIDS core data on epidemiology and response
99. Federal Ministry of Health of Ethiopia. HIV/AIDS Prevention and Control Office.
100. Poit MA: A simulation model of case-finding and treatment in tuberculosis control
101. Piot MA: World Health Organization A simulation model of case finding and
Geneva; 1967.
102. Peter GS, Richard HM: Field trials of health interventions in developing
103. Puffer S, Torgerson DJ, Watson J: Cluster randomized controlled trials. J Eval
104. Donner A, Klar N: Design and analysis of cluster randomization trials in health
75
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
105. Green SB, Corle DK, Gail MH, Mark SD, Pee D, Freedman LS, Graubard BI,
Lynn WR: Interplay between design and analysis for behavioral intervention trials
593.
106. Chuang JH, Hripcsak G, Heitjan DF: Design and analysis of controlled trials in
107. Cluster randomised trials: Methodological and ethical considerations. In: 2002.
108. Hayes RJ, Moulton LH: Cluster randomised Trials. USA; 2009.
109. Altman DG, Bland JM: Statistics notes. Treatment allocation in controlled trials:
110. Bruno Giraudeau, Ravaud. P: Preventing Bias in Cluster Randomised Trials. PLoS
111. Rochon PA, Gurwitz JH, Sykora K, Mamdani M, Streiner DL, Garfinkel S,
Normand SL, Anderson GM: Reader's guide to critical appraisal of cohort studies:
112. Kenneth JR, Sanders G, Timothy LL: Modern Epidemiology. 3rd edition. USA;
2008.
114. Grimes DA, Schulz KF: Descriptive studies: what they can and cannot do. Lancet
2002, 359(9301):145-149.
76
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
116. Boccia S, La Torre G, Persiani R, D'Ugo D, van Duijn CM, Ricciardi G: A critical
117. Charles HH, Julie EB: Epidemiology in Medicine. 1st edition. USA: Lippincott
119. Shargie EB, Lindtjorn B: DOTS improves treatment outcomes and service
120. Normand SL, Sykora K, Li P, Mamdani M, Rochon PA, Anderson GM: Readers
121. Grimes DA, Schulz KF: Bias and causal associations in observational research.
validity of cluster randomised trials: systematic review of recent trials. Bmj 2008,
336(7649):876-880.
123. Green SB: Design of randomized trials. Epidemiol Rev 2002, 24(1):4-11.
124. World Health Organization: Community TB care: Practice and Policy. Volume
WHO/CDS/TB/2003.312; 2003.
77
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
125. Dye C, Watt, C. J., Bleed, D. M., Williams, B. G. : What is the limit to case
detection under the DOTS strategy for tuberculosis control? Tuberculosis 2003,
83(1-3):35-43.
126. Maher D, van Gorkom JL, Gondrie PC, Raviglione M: Community contribution to
tuberculosis care in countries with high tuberculosis prevalence: past, present and
127. Obermeyer Z, Abbott-Klafter J, Murray CJ: Has the DOTS strategy improved case
3(3):e1721.
129. Raviglione MC, Harries AD, Msiska R, Wilkinson D, Nunn P: Tuberculosis and
131. Omaswa F: The Community TB Care in Africa. Int J Tuberc Lung Dis 2003,
7(Suppl 1):S1-S102.
133. Community tuberculosis care policy recommendations. Int J Tuberc Lung Dis
134. Ninth Report of the WHO Expert Committee on Tuberculosis In WHO Technical
78
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
136. Singh AA, Parasher D, Shekhavat GS, Sahu S, Wares DF, Granich R:
tuberculosis patients: a field report from Haryana, North India. Int J Tuberc Lung
138. Murray CJ, Kreuser, J., Whang, W.: Cost-effectiveness analysis and policy
choices: investing in health systems. Bull World Health Organ 1994, 72(4):663-
674.
Machakos District, Kenya. Int J Tuberc Lung Dis 2003, 7(9 Suppl 1):S14-20.
in Lilongwe District, Malawi. Int J Tuberc Lung Dis 2003, 7(9 Suppl 1):S29-37.
community-based care for tuberculosis patients in rural Uganda. Int J Tuberc Lung
79
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
142. Sinanovic E, Floyd K, Dudley L, Azevedo V, Grant R, Maher D: Cost and cost-
based and health facility based directly observed treatment of tuberculosis in Dar
144. Ferroussier O, Kumar MK, Dewan PK, Nair PK, Sahu S, Wares DF, Laserson K,
mix project in Kannur District, Kerala, India, 2001-2002. Int J Tuberc Lung Dis
2007, 11(7):755-761.
2007, 21(3):232-239.
146. Stead WW, Bates JH: Recurrent tuberculosis due to exogenous reinfection. N Engl
148. Lambert ML, Hasker E, Van Deun A, Roberfroid D, Boelaert M, Van der Stuyft
3(5):282-287.
149. von Reyn CF, Horsburgh CR: Reinfection with Mycobacterium tuberculosis. Am J
150. Bates JH: Reinfection tuberculosis: how important is it? Am J Respir Crit Care
80
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
151. Cox HS, Morrow M, Deutschmann PW: Long term efficacy of DOTS regimens
152. Nolan CM: Reinfection with multidrug-resistant tuberculosis. N Engl J Med 1993,
153. Lin BT: Reinfection with multidrug-resistant tuberculosis. N Engl J Med 1993,
329(11):811-812.
154. Garcia Ordonez MA, Martinez Gonzalez J, Orihuela Canadas F, Jimenez Onate F,
155. Golub JE, Durovni B, King BS, Cavalacante SC, Pacheco AG, Moulton LH,
an endemic area may be due to relapse or reinfection. Tuber Lung Dis 1994,
75(3):199-202.
81
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Annexes
LCC LCC
83
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Total number of new patients registered during the quarter being reported:
Total number of cases registered for re-treatment during the reported quarter:
SCC
New cases
PTB-NEG.
LCC
SCC
New cases
EPTB
LCC
Re-treatment
84
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Date ___________
Age __________
Sex __________
Note: From ANC surveillance format we included study site, date, client code, age, sex
85
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
1. Socio-demographic variables
1.6. Educational status No schooling ____ Grade ____ other (specify) ______________
1.7. Occupation of suspect Farmer ____ student ____ merchant ____________ ______
2.3. Did you have fever and night sweats? Yes __________ No ________
2.8. Did you have history of tuberculosis treatment? Yes __________ No ________
2.9. Did you have closer contact with known TB patient? Yes __________ No ________
86
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
V. LABORATORY REGISTER
Lab. Woreda
Patient Health Name and address Name and address New Follow- Results
Date Serial Age Sex T B/Lep Sign Remarks
number Unit of patient of contact person patient up
number number
1 2 3
87
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
88
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
89
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Cost of companion
Visit to GHWs or HEWs
accomodation (if stayed
means of travel
travel type of type of overnight)
date transportati time in Cost Cost
cost meal drinks amount
on minutes place reason
paid
90
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Cost of companion
Visit to GHWs or HEWs
accomodation (if stayed
means of travel
travel type of type of overnight)
date transportati time in Cost Cost
cost meal drinks amount
on minutes place reason
paid
Note: Use extra sheet for more than one care takers
91
Daniel Gemechu Datiko
Improving Tuberculosis Control in Ethiopia
Date______/_______/________
8. Treatment category New ____ Relapse _____ Failure ____ Defaulter ____ others ___
3. TB category New ____ Relapse _____ Failure ____ Defaulter ____ others ________
6. Treatment outcome Cured ___Completed ___ Died ___ Failure __ Default __ TO ___
Note: 1. Collect sputum specimen for patients with productive cough of two weeks or more
92
Daniel Gemechu Datiko
I
INT J TUBERC LUNG DIS 10(10):1166–1171
© 2006 The Union
SUMMARY
SETTINGS: The tuberculosis control programme, south- positive case notification rate increased from 45 to 143
ern region of Ethiopia. per 100 000 population, the case detection rate from
O B J E C T I V E : To assess the impact of the expansion of the 22% to 45%, and the treatment success rate from 53%
DOTS strategy on tuberculosis (TB) case finding and to 85%. The default and failure rates decreased from
treatment outcome. 26% to 6% and from 7% to 1%, respectively.
D E S I G N : Reports of TB patients treated since the intro- D I S C U S S I O N : There was a steady increase in the treat-
duction of DOTS in the region were reviewed. Patients ment success rate with the decentralisation of DOTS. Al-
were diagnosed and treated according to World Health though 94% coverage was achieved after 10 years, the
Organization (WHO) recommendations. Case notification stepwise scale-up was important in securing resources
and treatment outcome reports were compiled quarterly and dealing with challenges. The programme achieved
at district level and submitted to the regional programme. 85% treatment success; however, with the current low
R E S U L T S : Of 136 572 cases registered between 1995 case detection rate (45%), the 70% WHO target seems
and 2004, 47% were smear-positive, 25% were smear- unachievable in the absence of alternative case-finding
negative and 28% had extra-pulmonary tuberculosis mechanisms.
(EPTB). In 2004, 94% of the health institutions were K E Y W O R D S : tuberculosis; TB control; DOTS; case find-
covered by DOTS. Between 1995 and 2004, the smear- ing; Ethiopia
TUBERCULOSIS (TB) is a major infectious cause of In Africa, the case detection rate is below 50% and
death among adults in sub-Saharan Africa. The situ- the treatment success rate could not exceed 73%.3
ation is compounded by low socio-economic status, With estimated new cases of TB at 356 per 100 000
displacement due to famine, drought and war and, population, Ethiopia is the seventh among the 22 high
in the last two decades, due to the human immuno- TB burden countries in the world.3 The National Tuber-
deficiency virus (HIV)/acquired immune-deficiency syn- culosis and Leprosy Control Programme (NTLCP) was
drome (AIDS). The World Health Organization (WHO) established in 1994, the DOTS strategy was adopted,
declared TB to be a global emergency in 1993.1 The and more than 90% geographic coverage had been
DOTS strategy is believed to be the most valuable achieved by 2004 (NTLCP report, 2004). Although
strategy for TB control;2 at the end of 2003, 182 DOTS has been implemented in the Southern Na-
countries in the world had adopted and implemented tions, Nationalities and People’s Region (SNNPR) of
DOTS.3 Studies from resource-poor settings demon- Ethiopia since 1995, its performance, including trends
strated that DOTS is effective for TB control,4–6 and in TB cases detected and their treatment outcomes,
the World Bank report stated that TB chemotherapy has not been evaluated. A recent retrospective trend
is ‘one of the most cost-effective of all health interven- analysis of TB patients in Hadiya Zones in SNNPR
tions’.7 WHO set a target to detect 70% of smear- demonstrated that the implementation of DOTS re-
positive cases and to treat 85% of them successfully sulted in improved treatment success and a decrease
by 2005.8 Because of low access to diagnostic labora- in defaulter rates.11 The present study aims to assess
tories, shortage of resources and trained personnel, the impact of the implementation and expansion of
coupled with low sensitivity of smear microscopy, the DOTS on the trends of TB cases and their treatment
global target in case detection has been challenged.9,10 outcome in the SNNPR.
Correspondence to: Dr Mohammed A Yassin, Liverpool School of Tropical Medicine, Pembroke Place, L3 5QA, Liverpool,
UK. Tel: (!44) 15 170 53219. Fax: (!44) 15 170 53329. e-mail: [email protected]
Article submitted 24 April 2006. Final version accepted 23 June 2006.
Experiences of the TB control programme in Ethiopia 1167
STUDY POPULATION AND METHODS tion form contains information on reporting district,
quarter of the year, and number of patients registered
SNNPR is located in the south-western part of Ethio- by sex and disease status (new smear-positive, return
pia and is one of the largest regions, with about 14 after default, failure, relapse, smear-negative, EPTB).
million inhabitants. Ninety-three per cent of the pop- Information on follow-up smear examination is also
ulation live in the rural part of the region and acces- included. The treatment outcome form contains in-
sibility of the health services is limited, as only half of formation on the quarter the patients were registered
the population resides within 2 hours walking dis- (15 months earlier), the total number registered, the
tance from a public health facility. A pilot DOTS number of patients who were cured, completed treat-
programme, supported financially and technically by ment, defaulted, failed, died or transferred out and
ALERT, the German Leprosy/TB Relief Association the total number evaluated. The regional coordinator
(GLRA) and the NTLCP, was introduced in a few checks the completeness, quality and accuracy of the
health facilities in the region in 1995–1996, gradually reports, analyses and interprets the data and sends a
expanded to other sites, and was integrated into the compiled report to the NTLCP. All coordinators keep
general health services in 1999. copies of the reports for their documentation.
The NTLCP adopted the WHO-recommended re- For the purpose of this study, the data retained
cording and reporting guidelines and forms for mon- at the regional level were entered into Epi Info 2000
itoring and evaluation of programme activities. Briefly, (Centers for Disease Control and Prevention, Atlanta,
patients with signs and symptoms suggestive of TB12 GA, USA) for analysis. The case detection rate (CDR)
are screened for confirmation of the diagnosis and ini- was calculated by dividing the number of smear-
tiation of treatment. Patients with symptoms compat- positive cases by the WHO-estimated number of
ible with pulmonary tuberculosis (PTB) submit three smear-positive cases per 100 000 population in the same
sputum samples (spot-morning-spot). Smears are graded year for the country. The CDR and the proportion
according to the guidelines of the International Union of smear-positive cases treated by SCC who success-
Against Tuberculosis and Lung Disease (The Union)13 fully completed treatment were considered as the
and to the recommendations of the National Labora- main outcome variables. Ethical approval was not re-
tory Manual.14 Patients with at least two positive quired as the survey was based on retrospective data.
smears are considered smear-positive and those with
three negative smears are requested to undergo chest
RESULTS
X-ray (CXR) or are treated with antibiotics and then
re-evaluated. The diagnosis of extra-pulmonary TB Following the integration of the TB control programme
(EPTB) is usually made clinically and by the decision activities into the general health services, the pro-
of the clinician. gramme was expanded to most of the districts and
Patients who are diagnosed with TB are referred to health facilities in the region (Table). All 13 zones
TB clinics where they receive health education and are have been covered by the programme since 2001, and
registered. Smear-positive patients registered in the 100% coverage of the districts was achieved in early
DOTS clinics receive 8-month short-course chemo- 2004. In mid 2004, 445 (94%) of the 475 health facil-
therapy (SCC) including daily supervised streptomycin ities in the region—14 (100%) hospitals, 114 (100%)
(S), rifampicin (R, RMP), isoniazid (H) and pyrazina- health centres and 317 (91%) health stations—were
mide (Z) for 2 months followed by self-administered implementing DOTS. Over the 10-year period from
ethambutol (E) and H for 6 months for adults (2SRHZ/ 1995 to 2004, 136 572 patients with all forms of TB
6EH) and RH for 4 months for children (2SRHZ/4RH). were registered for treatment. Of these, 47% were
Smear-positive patients are monitored by smear exam- smear-positive, 25% smear-negative and 28% EPTB,
ination at the end of months 2, 5 and 7 of treatment. with no marked differences throughout the years. The
Smear-negative and EPTB cases receive RHZ in the case notification rate of all forms of TB increased
first 2 months followed by EH for 6 months (2RHZ/ from 45/100 000 in 1995 to 143/100 000 in 2004, and
6EH) and are monitored by regularity of attendance the smear-positive CDR doubled from 22% in 1995
and clinical improvement. to 45% in 2004 (Figure). The increase in the number
The Woreda (district) communicable diseases co- of registered TB cases, from 4648 in 1995 to 20 196
ordinator (WCDC) compiles the information about all in 2004, correlated with the decentralisation and ex-
TB patients entered into the unit registers of all health pansion of the DOTS strategy to more health facili-
facilities in the district and assigns a unique district num- ties, which increased from four in 1995 to 445 in 2004
ber to each patient in the district TB register. The (R2 " 0.65).
WCDC completes the forms for case notification and The follow-up smear results at month 2 were not
treatment outcome quarterly and submits the report to available for 51% of the patients in 1996; this pro-
the zonal coordinator, who is responsible for compil- portion had decreased to 22% in 2004. Among patients
ing the zonal summary, and submits them in turn to the who underwent follow-up smear testing, the negative
regional programme coordinator. The case notifica- conversion rate at month 2 increased from 85% in
1168
Table Performance of the DOTS strategy in the southern region of Ethiopia, 1995 to 2004
Year 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
Population of the region (million) 10.4 10.5 10.8 11.1 12.1 12.5 12.9 13.3 13.7 14.1
Programme expansion, n (%)
Zones implementing DOTS 3 (33) 4 (44) 5 (56) 6 (67) 6 (67) 8 (89) 13 (100) 13 (100) 13 (100) 13 (100)
Woredas with DOTS 4 (5) 22 (29) 24 (31) 26 (34) 31 (40) 43 (46) 80 (77) 92 (89) 100 (96) 104 (100)
Health facilities with DOTS 4 (1) 22 (6) 32 (7) 52 (11) 61 (13) 95 (20) 180 (38) 236 (49) 350 (73) 445 (94)
All forms of TB, n 4648 6197 9725 10 004 15 167 17 067 17 858 17 246 18 464 20 196
New smear-positive PTB, % 48 43 40 41 38 44 46 49 48 49
New smear-negative PTB, % 17 24 22 23 33 27 26 22 22 23
New EPTB, % 33 32 37 35 28 27 26 27 26 26
Other,* % 1 2 1 1 2 2 3 3 4 2
Treatment and follow-up
Percentage of PTB! on SCC 20 27 33 46 56 73 91 100 100 100
Sputum tested at month 2, n (%) 1751 2674 2 731 4 922 6 783 7 695 8 880 8 238
Positive 141 (8) 204 (7) 58 (2) 172 (3.5) 222 (3) 213 (3) 222 (3) 295 (4)
Negative 767 (41) 1227 (46) 1 533 (56) 3 074 (62) 5 027 (74) 6 074 (79) 7 015 (79) 6 052 (74)
The International Journal of Tuberculosis and Lung Disease
Not done 843 (51) 1243 (49) 1 140 (43) 1 676 (35) 1 534 (23) 1 408 (19) 1 891 (18) 1 891 (22)
Treatment outcome, n (%)
PTB! treated with SCC a year before 450 723 1 272 1 850 3 189 5 452 7 459 8 386 8 880
New PTB! evaluated NA 560 736 826 2 158 2 981 5 516 7 488 8 501 6 245
New PTB! not evaluated NA #110 (24)† #13 (2)† 446 (35) #308 (17)† 208 (7) #64 (1)† #29 (0.4)† #115 (1)† 2 635 (30)‡
Cured NA 140 307 376 926 1 430 2 708 4 386 4 827 3 933
Treatment completed NA 154 252 260 591 855 1 673 1 771 2 160 1 349
Success rate NA 294 (53) 559 (76) 636 (77) 1 517 (70) 2 285 (77) 4 381 (79) 6 157 (82) 6 987 (82) 5 282 (85)
Default NA 89 (16) 88 (12) 100 (12) 311 (14) 368 (12) 538 (10) 507 (7) 534 (6) 378 (6)
Failure NA 39 (7) 24 (3) 31 (4) 14 (1) 38 (1) 63 (1) 78 (1) 79 (1) 38 (1)
Death NA 27 (5) 39 (5) 44 (5) 252 (11) 167 (6) 274 (5) 384 (5) 446 (5) 270 (4)
Transferred out NA 111 (20) 26 (4) 15 (2) 64 (3) 123 (4) 260 (5) 362 (5) 455 (5) 277 (4)
Treated by LCC, not evaluated 1687 (75) 1942 (73) 2 590 (67) 2 200 (54) 2 506 (44) 1 940 (26) 704 (9) 0 0
PTB " pulmonary tuberculosis; SCC " short-course chemotherapy; LCC " long-course chemotherapy; NA" not applicable as at least 15 months are required to compile the treatment outcome of these patients.
Experiences of the TB control programme in Ethiopia 1169
The treatment success rate showed a significant in- treatment centres after 2000 did not yield a propor-
crease, reaching the 85% target in 2004. This was ac- tional increase in the number of cases detected. Al-
companied by a corresponding decrease in defaulter though this may partly be explained by a declining
and failure rates, corroborating findings in other coun- trend in the incidence of active TB, one cannot expect
tries.10 However, the treatment outcomes of 13 569 a dramatic decline in such a short period of time, and
smear-positive patients treated with LCC were not in- other factors merit consideration. Increasing the num-
cluded in this report due to incomplete information. ber of diagnostic and treatment centres improves phys-
Although these patients were registered and notified, ical access to health care, but cannot address other
their treatment outcomes were not routinely reported, barriers to access, such as socio-economic obstacles.19,20
as they did not receive SCC, which is one of the com- Such barriers have to be addressed to help TB suspects
ponents of the DOTS strategy. The treatment out- access diagnostic and treatment services as early as
comes of the patients treated with LCC are no doubt possible.
less favourable than for those treated with SCC. Fur- In conclusion, 10 years after the introduction of
thermore, we were unable to ascertain the treatment DOTS in our region, it was possible to triple TB case
outcomes of 2650 patients treated with SCC because notification and nearly double treatment success rates,
the treatment outcomes of the cohort registered in the while significantly reducing the defaulter and failure
last quarter of 2003, which accounted for the majority rates. There is a need to increase the CDR to achieve
of the missing cases, were not yet available. the 70% WHO target, and this warrants investigation
As our review was based on the quarterly reports of alternative intensified case-finding mechanisms.
retained at the region, we cannot exclude the possibil-
ity of poor recording and reporting systems resulting Acknowledgements
in the discrepancies between the registered and evalu- We are grateful to the Southern Region Health Bureau and all
ated cases. There was, however, a reduction in the the staff at the different levels of the TB Control Programme in the
proportion of cases with missing treatment outcome Region. We thank Mr M Aschalew for his assistance during data
information over the years, indicating an improvement collection.
in recording and reporting.
Evaluating the performance of TB control pro-
References
grammes may need to take into account to what ex-
1 World Health Organization. TB—a global emergency. WHO
tent the five elements of the DOTS strategy are in
Report on TB epidemic. WHO/TB/94.177. Geneva, Switzerland:
place. The DOTS-based programme in the southern WHO, 1994.
region of Ethiopia has tried to effectively address this 2 Khatri G R, Frieden T R. Rapid DOTS expansion in India. Bull
issue. There was clear political commitment, demon- World Health Organ 2002; 80: 457–463.
strated by the establishment of the control programme 3 World Health Organization. WHO Report 2005. Global tuber-
and budget allocation (although a larger proportion culosis control: surveillance, planning and financing. Geneva,
Switzerland: WHO, 2005.
of financial support comes from external assistance); 4 Xianyi C, Fengzeng Z, Hongjin D, et al. The DOTS strategy in
the introduction and expansion of the programme China: results and lessons after 10 years. Bull World Health
was accompanied by the establishment of a labora- Organ 2002; 80: 430–436.
tory network for sputum smear microscopy and qual- 5 Zhang L X, Tu D H, Enarson D A. The impact of directly-
ity control procedures;18 efforts have been made to observed treatment on the epidemiology of tuberculosis in Beijing.
Int J Tuberc Lung Dis 2000; 4: 904–910.
ensure an uninterrupted supply of anti-tuberculosis 6 Norval P Y, San K K, Bakhim T, Rith D N, Ahn D I, Blanc L.
drugs; SCC was provided under supervision during DOTS in Cambodia. Directly observed treatment with short-
the intensive phase of treatment; and well-organised course chemotherapy. Int J Tuberc Lung Dis 1998; 2: 44–51.
patient registration, follow-up and reporting mecha- 7 World Bank. World Development Report. Investing in health.
nisms were in place and the quality of record keep- New York, NY: Oxford University Press, 1993.
8 World Health Organization. Fifty-third World Health Assem-
ing has improved. The availability of unit registers at
bly. Stop TB initiative, report by the Director-General. A53/5,
treatment centres and district registers has facilitated 5 May 2000; Resolution WHA53.1. Geneva, Switzerland: WHO
the assessment of performance at various levels of the 2000.
programme, although due attention should be given 9 Veron L J, Blanc L J, Suchi M, Raviglione M C. DOTS expan-
to reporting the outcome of all patients registered for sion: will we reach the 2005 targets? Int J Tuberc Lung Dis
2004; 8: 139–146.
treatment.
10 Dye C, Watt C J, Bleed D. Low access to a highly effective ther-
These findings indicate that with the current global apy: a challenge for international tuberculosis control. Bull
TB control strategy, it is possible to achieve 85% World Health Organ 2002; 80: 437–444.
treatment success even in resource-poor settings with 11 Shargie E B, Lindtjorn B. DOTS improves treatment outcomes
low health service coverage.10 However, raising the and service coverage for tuberculosis in South Ethiopia: a ret-
CDR to 70% may continue to be an unachievable rospective trend analysis. BMC Public Health 2005; 5: 62.
12 Crofton J, Horne N, Miller F. Clinical Tuberculosis. 2nd ed.
target in the absence of alternative mechanisms for London, UK: Macmillan Education Ltd, 1999.
improved and intensified case finding.9 As discussed 13 Enarson D A, Reider H L, Arnadottir T, Trébucq A. Manage-
earlier, an increase in the number of diagnostic and ment of tuberculosis: a guide for low income countries. 5th ed.
Experiences of the TB control programme in Ethiopia 1171
Paris, France: International Union Against Tuberculosis and 17 World Health Organization. WHO Report 2006. Global tuber-
Lung Disease, 2000. culosis control: surveillance, planning, financing. WHO/HTM/
14 Ministry of Health. Laboratory manual for the National Tuber- TB/2006.362. Geneva, Switzerland: WHO, 2006.
culosis and Leprosy Control Programme in Ethiopia. 2nd ed. 18 Shargie E B, Yassin M A, Lindtjorn B. Quality control of spu-
Addis Ababa, Ethiopia: Ministry of Health, 2002. tum microscopic examinations for acid fast bacilli in southern
15 Yassin M A, Takele L, Gebresenbet S, et al. HIV and tubercu- Ethiopia. Ethiop J Health Dev 2005; 19: 104–108.
losis coinfection in the southern region of Ethiopia: a prospec- 19 Cambanis A, Yassin M A, Ramsay A, Squire S B, Arbide I, Cue-
tive epidemiological study. Scand J Infect Dis 2004; 36: 670– vas L E. Rural poverty and delayed presentation to tuberculosis
673. services in Ethiopia. Trop Med Int Health 2005; 10: 330–335.
16 Rieder H L. Sputum smear conversion during directly observed 20 Demissie M, Lindtjorn B, Berhane Y. Patient and health service
treatment for tuberculosis. Tubercle Lung Dis 1996; 77: 124– delay in the diagnosis of pulmonary tuberculosis in Ethiopia.
129. BMC Public Health 2002; 2: 23.
RÉSUMÉ
C O N T E X T E : Programme de lutte antituberculeuse, Région été couvertes par le DOTS. Entre 1995 et 2004, le taux
du Sud, Ethiopie. de déclaration des cas à bacilloscopie positive est passé de
O B J E C T I F : Evaluer l’impact de l’expansion de la stratégie 45 à 143 par 100 000 habitants, le taux de détection
DOTS sur le dépistage des cas de tuberculose (TB) et sur de 22% à 45%, le taux du succès du traitement de 53%
les résultats du traitement. à 85%, alors que les taux d’abandon et d’échec diminu-
S C H É M A : On a revu les rapports sur les patients TB aient respectivement de 26% à 6% et de 7% à 1%.
traités depuis l’introduction du DOTS dans la région. Le D I S C U S S I O N : On a noté une augmentation régulière du
diagnostic et le traitement des patients ont été menés taux de succès du traitement grâce à la décentralisation
conformément aux recommandations de l’Organisation du DOTS. Bien qu’après 10 ans on ait atteint une cou-
Mondiale de la Santé (OMS). Les rapports de déclara- verture de 94%, l’escalade par étapes a été importante
tion des cas et de résultats du traitement ont été revus pour l’obtention de ressources et pour faire face aux défis.
tous les trimestres au niveau du district et soumis au Pro- Le programme a atteint le taux de 85% de succès du
gramme Régional. traitement bien qu’actuellement, avec un faible taux de dé-
R É S U L T A T S : Sur 136 572 cas enregistrés entre 1995 et tection des cas de 45%, l’accès à la cible de 70% de l’OMS
2004, 47% avaient une bacilloscopie positive, 25% une semble irréalisable en l’absence de mécanismes alternatifs
bacilloscopie négative et 28% étaient des TB extra- de dépistage des cas.
pulmonaires. En 2004, 94% des institutions de soins ont
RESUMEN
MARCO DE REFERENCIA : El programa de lucha contra del programa DOTS. Entre 1995 y 2004, la tasa de noti-
la tuberculosis (TB) en la región meridional de Etiopía. ficación de casos con baciloscopia positiva aumentó de
O B J E T I V O : Evaluar la repercusión de la expansión de la 45 a 143 por 100 000 habitantes, la detección de casos
estrategia DOTS sobre la búsqueda de casos y el desen- del 22% al 45%, la tasa de tratamiento exitoso del 53%
lace terapéutico. al 85% y disminuyó la tasa de abandono del 26% al 6%
M É T O D O : Se analizaron los informes de tratamiento de y la de fracasos del 7% al 1%.
pacientes con TB desde la introducción de DOTS en la D I S C U S I Ó N : Con la descentralización de DOTS se ob-
región. El diagnóstico y el tratamiento de los pacientes servó un aumento progresivo de la tasa de tratamiento
cumplieron con las recomendaciones de la Organización exitoso. Si bien la cobertura del 94% se alcanzó después
Mundial de la Salud (OMS). A escala del distrito, se de 10 años, esta expansión gradual fue importante para
compilaron trimestralmente las notificaciones de casos y garantizar los recursos y responder a las dificultades. El
los informes sobre el desenlace terapéutico y se remi- programa obtuvo un tratamiento exitoso del 85% de los
tieron al programa regional. casos, pero dada la baja tasa actual detección (45%),
R E S U L T A D O S : De los 136 572 casos registrados entre pareciera inalcanzable la meta de la OMS, a menos que
1995 y 2004, el 47% tuvo baciloscopia positiva, 25% se implementen mecanismos diferentes de búsqueda de
baciloscopia negativa y 28% TB extrapulmonar. In 2004, casos.
el 94% de los establecimientos sanitarios formaba parte
II
BMC Public Health BioMed Central
Address: 1Southern Nations, Nationalities and Peoples' Regional Health Bureau, P.O. Box 149, Awassa, Ethiopia, 2Liverpool School of Tropical
Medicine, Pembroke place, L3 5QA, Liverpool, UK and 3Centre for International Health, University of Bergen, Armauer Hansen Building, N-5012,
Bergen, Norway
Email: Daniel G Datiko* - [email protected]; Mohammed A Yassin - [email protected]; Luelseged T Chekol - [email protected];
Lopisso E Kabeto - [email protected]; Bernt Lindtjørn - [email protected]
* Corresponding author †Equal contributors
Abstract
Background: A complex interaction exists between tuberculosis (TB) and human
immunodeficiency virus (HIV) infection at an individual and community level. Limited knowledge
about the rate of HIV infection in TB patients and the general population compromises the planning,
resource allocation and prevention and control activities. The aim of this study was to determine
the rate of HIV infection in TB patients and its correlation with the rate HIV infection in pregnant
women attending antenatal care (ANC) in Southern Ethiopia.
Methods: All TB patients and pregnant women attending health institutions for TB diagnosis and
treatment and ANC were consecutively enrolled in 2004 – 2005. TB diagnosis, treatment and HIV
testing were done according to the national guidelines. Blood samples were collected for
anonymous HIV testing. We used univariate and multivariate logistic regression analysis to
determine the risk factors for HIV infection and linear regression analysis to determine the
correlation between HIV infection in TB patients and pregnant women.
Results: Of the 1308 TB patients enrolled, 226 (18%) (95%CI: 15.8 – 20.0) were HIV positive. The
rate of HIV infection was higher in TB patients from urban 25% (73/298) than rural areas 16% (149/
945) [AOR = 1.78, 95%CI: 1.27–2.48]. Of the 4199 pregnant women attending ANC, 155 (3.8%)
[95%CI: 3.2–4.4] were HIV positive. The rate of HIV infection was higher in pregnant women from
urban (7.5%) (80/1066) than rural areas (2.5%) (75/3025) [OR = 3.19, 95% CI: 2.31–4.41]. In the
study participants attending the same health institutions, the rate of HIV infection in pregnant
women correlated with the rate of HIV infection in TB patients (R2 = 0.732).
Conclusion: The rate of HIV infection in TB patients and pregnant women was higher in study
participants from urban areas. The rate of HIV infection in TB patients was associated with the
prevalence of HIV infection in pregnant women attending ANC.
Page 1 of 8
(page number not for citation purposes)
BMC Public Health 2008, 8:266 http://www.biomedcentral.com/1471-2458/8/266
Page 2 of 8
(page number not for citation purposes)
BMC Public Health 2008, 8:266 http://www.biomedcentral.com/1471-2458/8/266
Figure
Map 1 Southern Nations, Nationalities and Peoples' Region of Ethiopia showing the survey sites, 2004 – 2005
of the
Map of the Southern Nations, Nationalities and Peoples' Region of Ethiopia showing the survey sites, 2004 –
2005. _____ Zonal boundary. Regional boundary. International boundary. Regional capital,
Awassa. Lake. TB-HIV survey sites. ANC-based sentinel survey sites. Overlapping sites.
course of broad-spectrum antibiotics and decided by a cli- Health Research Laboratory (CHRL) and stored at – 20°C
nician to be treated with anti-TB chemotherapy were clas- until analysis. The serum samples were anonymously
sified as EPTB cases. In children, TB was diagnosed if there tested for HIV using ELISA test (Vironostica ® Uniform II
were symptoms and signs suggestive of TB, contact history Ag/Ab BIOMÉRIEUX). All the samples were sent to the
with a known TB patient and x-ray findings consistent Ethiopian Health and Nutrition Research Institute
with active TB. (EHNRI) to repeat ELISA test using Enzygnost Anti-HIV1/
2 Plus (Dade Behring, Germany) and quality control.
Data and specimen collection ELISA reactive specimens at CHRL and EHNRI were con-
Trained laboratory technicians and health workers from sidered positive and discordant specimens were retested
TB and ANC functions collected the data using pretested using similar tests [11,12].
questionnaires. The main variables were age, sex, resi-
dence and survey site for all participants, and disease clas- Data analysis
sification and category for TB patients. 5 ml of blood We used SPSS 14.0 (SPSS Inc, Chicago, IL, USA) for data
samples were collected from TB patients and pregnant entry and analysis. We determined the rate of HIV infec-
women. Routine blood tests except for HIV were done tion in TB patients and pregnant women. Univariate and
locally and reported to the attending health workers. The multivariate logistic regression analysis were used to
remaining serum samples were stripped off individual determine the risk factors for HIV infection in TB patients
identifying markers and were assigned unique codes. They and pregnant women. Socio-demographic variables that
were kept at 4°C, transported to the regional Centre for were significant by univariate analysis were included in
Page 3 of 8
(page number not for citation purposes)
BMC Public Health 2008, 8:266 http://www.biomedcentral.com/1471-2458/8/266
the model to calculate adjusted odds ratio and 95% con- There was no difference in the rate of HIV infection by TB
fidence interval by HIV status in TB patients. We also did disease classification: the rate of HIV infection among
linear regression analysis to determine the variation of smear-positive PTB cases 17.5% (92/526) was similar to
HIV infection in TB patients explained by the prevalence smear-negative PTB 18.1% (78/432) [OR = 1.048, 95%CI:
of HIV infection among pregnant women from all study 0.723–1.519] and EPTB cases 18.2% (54/297) [OR =
sites and then for the study participants from the same 1.009, 95%CI: 0.687–1.480]. The rate of HIV infection
health institutions. P-value < 0.05 was considered as sta- was higher in TB patients from urban (24.5%, 73/298)
tistically significant. than rural areas (15.8%, 149/945) [AOR = 1.78, 95%CI:
1.27–2.48] as shown in Table 1 &3.
Ethical clearance
Ethical Review Committee of the Regional Health Bureau Of the 4199 pregnant women attending ANC, 3097
approved the study. Oral informed consent was obtained (74%) came from rural and 1096 (26%) from urban
for all study participants. The study participants who areas. Their mean age was 25.7 years. The prevalence of
wanted to know their HIV status were advised to go to vol- HIV infection among the pregnant women was 3.8%
untary counselling and testing service located within the (155/4091) [95%CI: 3.2–4.4] ranging from 1.5% (in
health institutions or nearby. Gamo Goffa zone) to 10.5% (in Wolaita zone). The rate
of HIV infection was higher among women from urban
Results 7.5% (80/1066) than rural 2.5% (75/3025) areas [OR =
1308 TB patients and 4199 pregnant women were 3.19, 95% CI: 2.31–4.41] (Table 2 &3).
included in the study. Of the TB patients, 729 (56%) were
men and 569 (44%) were women. 309 (24%) patients In all survey sites, where both surveys were conducted in
came from urban and 978 (76%) patients from rural the same as well as in different health institutions, we
areas. Their mean age was 28.4 years. 544 (42%) patients found no correlation between the rate of HIV infection
had smear-positive PTB, 449 (34%) smear-negative PTB among pregnant women and TB patients (R2 = 0.034).
and 308 (24%) EPTB. The rate of HIV infection in TB Briefly, South Omo zone with the highest TB-HIV co-
patients was 18% (226/1261) [95%CI: 15.8–20.0] rang- infection rate did not have higher rate of HIV infection
ing from 8.3% (in Silte zone) to 35.3% (in South Omo among pregnant women whereas Silte zone that had the
zone). The rate of HIV infection in TB patients was similar lowest rate of TB-HIV co-infection did not have the lowest
for men and women (OR = 1.00, 95%CI: 0.75 – 1.34). rate of HIV infection among pregnant women (Table 3).
Table 1: Socio-demographic characteristics and HIV status of TB patients, southern Ethiopia, 2004 – 2005
Variables TB Patients without TB patients with HIV OR (95%CI) P-value AOR (95%CI) P-value
HIV (N = 226), n (%)
(N = 1035), n (%)
TB = Tuberculosis, HIV = Human immunodeficiency virus, OR = odds ratio, CI = confidence interval, AOR = adjusted OR for age and residence
SD = standard deviation, PTB +ve = smear positive pulmonary TB, PTB -ve = smear negative pulmonary TB, EPTB = extrapulmonary TB, R =
relapse,
F = failure, D = return after default, O = others. Missing variables: age – 15 (1.2%), sex – 12(0.9%), address – 21(1.6%), disease classification –
7(0.5%), disease category – 58(4.4%), HIV result – 47(3.6%), sex & HIV result – 58(4.4%), age group and HIV – 61(4.7%), address and HIV –
21(1.6%), disease classification and HIV – 53(4.1%) and age category and HIV- 61(4.7%).
Page 4 of 8
(page number not for citation purposes)
BMC Public Health 2008, 8:266 http://www.biomedcentral.com/1471-2458/8/266
Table 2: Socio-demographic characteristics and HIV status of pregnant women attending ANC, Southern Ethiopia, 2004 – 2005
Variables ANC attendants without HIV ANC attendants with HIV OR (95%CI) P – value
(N = 3936), n (%) (N = 155), n (%)
ANC = antenatal care, HIV = human immunodeficiency virus, OR = odds ratio, CI = confidence interval, SD = standard deviation
Missing variables: age – 6(0.1%), address – 6(0.1%), HIV result – 108(2.6%), age group & HIV – 108(2.6%), address & HIV – 108(2.6%)
In contrast, in the six study sites where the two surveys The information about the rate of HIV infection among
were conducted in the same health institutions, there was different groups of a community is important to under-
a strong correlation between the rate of HIV infection stand the extent of the problem and to implement appro-
among pregnant women and TB patients (R2 = 0.732). priate prevention and control measures.
Upon further analysis by residence, the magnitude of cor-
relation was stronger for study participants from urban In a large representative survey of TB patients in southern
(R2 = 0.998) than rural areas (R2 = 0.546) as shown in Ethiopia, less than a fifth of them were HIV infected sim-
Table 4 and Figure 2. From a linear regression analysis, we ilar to other reports from the region [9,16]. Higher TB-HIV
found the equation, prevalence of HIV among pregnant co-infection rates, as high as 47% was reported from Ethi-
women = -6.22 + 0.89* the rate of HIV infection in TB opia [17,18]. These studies however were hospital-based
patients. Each per cent increase of HIV seroprevalence in and were conducted in few major towns where the preva-
TB patients corresponded to an increase in seroprevalence lence of HIV infection in the general population was
of 0.89% among pregnant women. much higher.
Table 3: The rate of HIV infection among TB patients and pregnant women attending antenatal care in southern region of Ethiopia
2004 – 2005
Survey sites by zones* ANC attendants with HIV % (N) TB patients with HIV % (N) R2† Adjusted R2 P-value‡
*The survey sites were areas where we conducted the two surveys in the same and different health institutions.
†R2-coefficient of determination weighed for the number of study participants
‡P-value for adjusted R2 HIV = Human immunodeficiency virus, ANC = Antenatal care
Page 5 of 8
(page number not for citation purposes)
BMC Public Health 2008, 8:266 http://www.biomedcentral.com/1471-2458/8/266
Table 4: The rate of HIV infection among TB patients and pregnant women attending antenatal care in the same health institutions of
southern region of Ethiopia 2004 – 2005
Survey sites by zones* ANC attendants with HIV % (N) TB patients with HIV % (N) R2† Adjusted R2 P – value
*The survey sites were areas where we conducted the two surveys in the same health institutions in a district.
†R2-coefficient of determination weighed for the number of study participants
different TB classifications. This could be due to the rela- has been used as a proxy for HIV prevalence in the general
tively low prevalence of HIV infection in the region [13]. population [9]. In our study, the prevalence of HIV infec-
Another possible explanation could be under diagnosis or tion among pregnant women attending ANC was 3.8%.
referral of some smear-negative and EPTB suspects with a This was similar to the previous reports from the region
potentially higher risk of HIV infection due to limited [10] but lower than the reports of sentinel surveillance
diagnostic facilities. from other parts of the country [7] and sub-Saharan Afri-
can countries [19,20]. As expected, the prevalence of HIV
Although the ANC-based HIV sentinel surveillance has among pregnant women was higher in urban areas than
weaknesses as the results may be affected by low attend- rural areas; this could be due to the difference in the risk
ance of ANC, exclusion of private clinics, the rate of con- and rate of HIV infection in urban and rural communities
traceptive use and provides no information about men, it [21,22].
Page 6 of 8
(page number not for citation purposes)
BMC Public Health 2008, 8:266 http://www.biomedcentral.com/1471-2458/8/266
In our study, the correlation between the seroprevalence 6(8):483-495. Nunn P, Williams B, Floyd K, Dye C, Elzinga G, Ravigli-
one M
in pregnant women and TB patients coincided with the 5. Harries A, Maher D, Graham S, TB/HIV: A Clinical Manual. 2nd
spread and stage of HIV epidemic in a community. This ed. WHO/HTM/TB/2004.329. Geneva. Switzerland. WHO
was reflected by the higher rate of HIV infection among TB 2004:1-210.
6. Ministry of Health of Ethiopia: Tuberculosis and Leprosy Preven-
patients and pregnant women in urban areas. This could tion and Control Programme Manual. 2nd ed. Addis Ababa.
be because of matured HIV epidemic in urban areas that Ethiopia: MOH 2002.
led to an increased number of TB cases and number of 7. Federal Ministry of Health of Ethiopia HIV/AIDS Prevention and Con-
trol Office: AIDS in Ethiopia. Sixth report. Addis Ababa. Ethi-
HIV infected TB patients [24,25]. In rural areas, we found opia: MOH 2006:1-52.
lower correlation possibly due to the low HIV prevalence 8. World Health Organization: Guidelines for HIV Surveillancea-
mong Tuberculosis patients. 2nd ed. WHO/HTM/TB/
in the rural communities [13] and a lag period between 2004.339. Geneva. Switzerland. WHO 2004:1-32.
the spread of HIV infection and maturity of the epidemic. 9. UNAIDS/WHO Working Group on Global HIV/AIDS and STI Sur-
In Zimbabwe, an increase in TB incidence occurred four to veillance: Guidelines for Conducting HIV Sentinel Serosur-
veys among Pregnant Women and Other Groups: UNAIDS/
five years after the spread of HIV infection in the commu- 03.49E. Geneva, Switzerland: UNAIDS; 2003:1-66.
nity [4] and a lag period of seven years was reported from 10. Yassin MA, Takele L, Gebresenbet S, Girma E, Lera M, Lendebo E:
HIV and tuberculosis coinfection in the southern region of
Kenya [26]. Generally, HIV prevalence surveys in Africa, Ethiopia: a prospective epidemiological study. Scand J Infect Dis
Asia and Pacific showed HIV prevalence in TB patients to 2004, 36(9):670-673.
be many times higher than that was seen in the general 11. Ministry of Health of Ethiopia: Revised National Guideline for
ANC – Based HIV Surveillance. Addis Ababa. Ethiopia. MOH
population [27-29]. Similar to the report from Cameroon, 2004.
surveillance of HIV infection in TB patients could be used 12. World Health Organization: Guidelines for HIV surveillance
as an estimate of the rate of HIV infection in the general among tuberculosis patients. 2nd edition. Geneva, Switzerland;
2004.
population [30]. 13. Central Statistical Agency of Ethiopia: Demographic and
Health Survey 2005. Addis Ababa, Ethiopia 2006:1-410.
14. Liberato IR, de Albuquerque Mde F, Campelo AR, de Melo HR: Char-
Conclusion acteristics of pulmonary tuberculosis in HIV seropositive and
The rate of HIV infection in TB patients was associated seronegative patients in a Northeastern region of Brazil. Rev
with the prevalence of HIV infection among pregnant Soc Bras Med Trop 2004, 37(1):46-50.
15. Maher D, Harries A, Getahun H: Tuberculosis and HIV interac-
women in the general population. The seroprevalence tion in sub-Saharan Africa: impact on patients and pro-
information for TB patients and pregnant women could grammes; implications for policies. Tropical Medicine and
be valuable for planning, monitoring and evaluation of International Health 2005, 10:734-742.
16. Madebo T, Nysaeter G, Lindtjorn B: HIV infection and malnutri-
joint prevention and control activities. The trend and level tion change the clinical and radiological features of pulmo-
of interaction of HIV infection in TB patients and preg- nary tuberculosis. Scand J Infect Dis 1997, 29(4):355-359.
17. Converse PJ, Dual Infection: The Challenge of HIV/AIDS and
nant women need further study. Tuberculosis in Ethiopia. Northeast African Studies 2000,
7(1):147-166.
Competing interests 18. Demissie M, Lindtjørn B, Tegbaru B: Human immunodeficiency
virus (HIV) infection in tuberculosis patients in Addis Ababa.
The authors declare that they have no competing interests. Ethiop J Health Dev 2000, 14(3):277-282.
19. World Health Organization: Global Tuberculosis Control, sur-
veillance, planning and financing. WHO Report 2005. WHO/
Authors' contributions HTM/TB/2005.349. Geneva, Switzerland: WHO; 2005:1-247.
DGD, LTC and LEK supervised data collection and labora- 20. Yahya-Malima KI, Olsen BE, Matee MI, Fylkesnes KM: The silent
tory testing. DGD, MAY and BL analysed, interpreted the HIV epidemic among pregnant women within rural North-
ern Tanzania. BMC Public Health 2006, 6(109):.
findings and prepared the drafts. All authors contributed 21. Denise V, Sarbani C, Taha T: Evaluation of the World Bank's
to the final manuscript. Assistance in Responding to the AIDS Epidemic: Ethiopia
Case Study. 2005.
22. Shabbir I, Larson CP: Urban to rural routes of HIV infection
Acknowledgements spread in Ethiopia. J Trop Med Hyg 1995, 98(5):338-342.
We would like to thank the SNNPR Health Bureau for providing financial, 23. HIV testing policies and HIV surveillance among tuberculo-
technical and material support for the study. We are also grateful to the sis patients in Europe. Andrea Infuso (EuroTB) and Françoise F. Ham-
ers (EuroHIV), Dept. of Infectious Diseases, Institut de veille sanitaire, Saint-
staff working in the ANC, TB units and laboratories of the participating
Maurice, France .
health institutions. Our special thanks go to TB patients and pregnant 24. Kebede D, Aklilu M, Sanders E: The HIV epidemic and the state
women who consented to take part in the survey. of its surveillance in Ethiopia. Ethiop Med J 2000, 38(4):283-302.
25. Buve A, Bishikwabo-Nsarhaza K, Mutangadura G: The spread and
effect of HIV-1 infection in sub-Saharan Africa. Lancet 2002,
References 359(9322):2011-2017.
1. MacDougall DS: TB & HIV: the deadly intersection. J Int Assoc 26. Nunn P, Williams B, Floyd K, Dye C, Elzinga G, Raviglione M: Tuber-
Physicians AIDS Care 1999, 5(5):20-27. culosis control in the era of HIV. Nat Rev Immunol 2005,
2. Servilio J: HIV/TB dual infection cause for concern. Posit Aware 5(10):819-826.
1995:8. 27. Lawn SD, Bekker LG, Middelkoop K, Myer L, Wood R: Impact of
3. Harries A, Dye C: Tuberculosis. Annals of Tropical Medicine and Par- HIV Infection on the Epidemiology of Tuberculosis in a Peri-
asitology 2006, 100(5):415-431. Urban Community in South Africa: The Need for Age-Spe-
4. Reid A, Scano F, Getahun H, Wiliams B, Dye C, Nunn P, Cock K, cific Interventions. Clin Infect Dis 2006, 42(7):1040-7.
Hankins C, Miller B, Castro K, Raviglion M: Towards universal
access to HIV prevention, treatment, care, and support: the
role of tuberculosis/HIV collaboration. Lancet Infect Dis 2006,
Page 7 of 8
(page number not for citation purposes)
BMC Public Health 2008, 8:266 http://www.biomedcentral.com/1471-2458/8/266
Pre-publication history
The pre-publication history for this paper can be accessed
here:
http://www.biomedcentral.com/1471-2458/8/266/pre
pub
Page 8 of 8
(page number not for citation purposes)
III
Health Extension Workers Improve Tuberculosis Case
Detection and Treatment Success in Southern Ethiopia: A
Community Randomized Trial
Daniel G. Datiko1,2*, Bernt Lindtjørn1
1 Centre for International Health, University of Bergen, Bergen, Norway, 2 Southern Nations, Nationalities, and Peoples’ Regional Health Bureau, Awassa, Ethiopia
Abstract
Background: One of the main strategies to control tuberculosis (TB) is to find and treat people with active disease.
Unfortunately, the case detection rates remain low in many countries. Thus, we need interventions to find and treat
sufficient number of patients to control TB. We investigated whether involving health extension workers (HEWs: trained
community health workers) in TB control improved smear-positive case detection and treatment success rates in southern
Ethiopia.
Methodology/Principal Finding: We carried out a community-randomized trial in southern Ethiopia from September 2006
to April 2008. Fifty-one kebeles (with a total population of 296, 811) were randomly allocated to intervention and control
groups. We trained HEWs in the intervention kebeles on how to identify suspects, collect sputum, and provide directly
observed treatment. The HEWs in the intervention kebeles advised people with productive cough of 2 weeks or more
duration to attend the health posts. Two hundred and thirty smear-positive patients were identified from the intervention
and 88 patients from the control kebeles. The mean case detection rate was higher in the intervention than in the control
kebeles (122.2% vs 69.4%, p,0.001). In addition, more females patients were identified in the intervention kebeles (149.0 vs
91.6, p,0.001). The mean treatment success rate was higher in the intervention than in the control kebeles (89.3% vs 83.1%,
p = 0.012) and more for females patients (89.8% vs 81.3%, p = 0.05).
Conclusions/Significance: The involvement of HEWs in sputum collection and treatment improved smear-positive case
detection and treatment success rate, possibly because of an improved service access. This could be applied in settings with
low health service coverage and a shortage of health workers.
Citation: Datiko DG, Lindtjørn B (2009) Health Extension Workers Improve Tuberculosis Case Detection and Treatment Success in Southern Ethiopia: A
Community Randomized Trial. PLoS ONE 4(5): e5443. doi:10.1371/journal.pone.0005443
Editor: Delia Goletti, National Institute for Infectious Diseases (INMI) L. Spallanzani, Italy
Received November 25, 2008; Accepted March 21, 2009; Published May 8, 2009
Copyright: ! 2009 Datiko, Lindtjørn. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: The University of Bergen funded this study. The university had no role in the design, data collection, analysis and interpretation or writing of the report.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: [email protected]
(HEP) to ensure equitable access to health services. The aim of result was reported as treatment completed. If a patient remained
the HEP is to prevent major communicable diseases and pro- or became smear-positive at the end of 5 months or later, he/she
mote health in the community. A new cadre of community level was reported as treatment failure. A patient who missed treatment
health workers, health extension workers (HEWs), was trained for for eight consecutive weeks after receiving treatment for at least 4
1 year at an undergraduate level. With the aim of preventing weeks was reported as a defaulter. A patient who was transferred
major communicable diseases, HEWs are trained on how to to another district after receiving treatment for at least 4 weeks and
identify and refer TB suspects, trace defaulters, and provide whose treatment outcome was not reported to the referring district
treatment and health education [6,7]. However, their role in TB was reported as transferred out. A patient who died while on
control has not been evaluated. The aim of the present study was treatment was reported as dead irrespective of the cause of death
to establish whether involving HEWs in TB control improved [8].
smear-positive case detection and treatment success rates in Ethics. We obtained ethical clearance from the Ethical
southern Ethiopia. Review Committee of the Regional Health Bureau in southern
Ethiopia. We obtained permission from TB programme managers
Methods and kebele leaders after discussing with them community-based
TB care. TB patients were enrolled after giving informed
The protocol for this trial and supporting CONSORT checklist consent after explaining the aim of the study and the right
are available as supporting information; see Checklist S1 and to refuse or to withdraw from the study. HIV testing was not
Protocol S1. offered to TB patients because of the unavailability of HIV testing
and treatment in the study area at the time the study was
Study area and population conducted.
This study was conducted in Dale and Wonsho, rural districts of
Sidama zone in southern Ethiopia from September 2006 to April
The intervention
2008. There were 51 kebeles (lowest administrative units) in the
Training on how to identify TB suspects and administer
two districts. Fifty-five per cent of the population live within two- DOT. We trained health workers, laboratory technicians and
hour walking distance of health facilities. There were 21 health
HEWs for 2 days. The training focused on symptoms and
posts (operational unit for HEWs), two health stations, two nucleus
transmission of TB, how to identify TB suspects, how to collect,
health centres (health stations upgrading to health centres) and one
label, store and transport sputum specimens, administer DOT,
health centre. Three health facilities (one health centre and two
and follow patients during treatment. The messages and the
health stations) conducted sputum microscopy, and DOT was
content of our training were similar to the curriculum of training
provided in the health centre, nucleus health centres and health
HEWs. HEWs, in the in the intervention kebeles, received on job
stations. None of the health posts provide DOT.
training about how to collect sputum samples and support patients
to adhere to treatment. HEWs collected sputum specimens once a
Health service and HEP month. An ice box was used to keep the sputum specimens in the
The Government of Ethiopia has a four-tier health service, and health post and during their transportation on foot to diagnostic
the lowest level is a primary health care unit (a health centre and units. The intervention included sputum collection and providing
five satellite health posts). On average, a health post serves a kebele DOT.
with 5000 people. The health policy focuses on provision of During health education sessions at health posts, HEWs
preventive and promotive health care to the population under the informed people living in the kebele about TB and advised
HEP, which involves prevention and control of diseases, including them to come to a health post if they had productive cough of
TB. The local health authorities in consultation with kebele
2 weeks or more duration. TB suspects who came to the health
leaders select two female residents, who have completed tenth
posts were told about community-based TB care. HEWs
grade, from each kebele. The women receive training for 1 year
collected spot-morning-spot sputum specimens, and labelled and
and are placed as HEWs in their respective kebele. They receive a
transported them to the diagnostic units every month for
salary from the government and they are accountable to the health
examination for acid-fast bacilli by direct microscopy. Smear-
centre [7].
positive patients in the intervention kebeles received standard
DOTS under the direct observation of HEWs. TB patients visited
Participants health posts daily during the intensive phase and once a month in
the continuation phase.
TB case finding and treatment outcome
Case finding. TB suspects, who had cough for two weeks or
more, were referred for further investigations. A smear-positive Control kebeles
pulmonary TB case was defined by two positive sputum smears or Identifying TB suspects and DOT administration. HEWs
one positive smear and x-ray findings consistent with active TB. in the in the control kebeles did not received on job training about
Treatment regimen and duration. The treatment regimen how to collect sputum samples and how to support patients to
for new smear-positive cases consisted of two months intensive adhere to treatment. However, they provided health services,
phase treatment with ethambutol, rifampicin, isoniazid and including health education about TB, to the people living in their
pyrazinamide followed by continuation phase treatment for 6 kebeles. TB suspects presented themselves to diagnostic units.
months with ethambutol and isoniazid. For children, in the However, the health workers from health facilities were trained as
continuation phase, ethambutol/isoniazid was replaced by they provided the service to intervention and control kebeles.
rifampcin/isoniazid for 4 months. Follow-up sputum smear Smear-positive patients in the control kebeles received standard
examination was done at the end of 2, 5 and 7 months treatment. DOTS were treated under the direct observation of general health
Treatment outcome. A patient with at least two negative workers at health centres. TB patients visited health centres and
smears including that at 7 months was reported as cured. A patient health stations daily during the intensive phase and once a month
who finished the treatment but did not have the 7-month smear in the continuation phase.
Objective. the objective of the study was to investigate intervention and control groups using a table of random numbers
whether involving HEWs in TB control improves the case (Figure 1).
detection and treatment success rate in southern Ethiopia
Blinding
Outcome variables Neither the general health workers nor TB programme
Case detection rate. the number of new smear-positive cases managers were blinded to the allocation. Although we did not
detected divided by the estimated number of incident smear- blind the laboratory technicians, they were not informed whether
positive cases, expressed as a percentage. the sputum specimens were from the intervention or control
Treatment success rate. cure or treatment completion rate kebeles.
was calculated as the number of patients cured or treatment
completed divided by the total number of patients reported Data collection
expressed as a percentage. Treatment success rate (TSR) was the TB case finding and treatment outcome data were collected
sum of cure and treatment completion rate. from TB registers at health facilities and districts. The information
collected included date, age, sex, address, TB classification, smear
Sample size calculation results and treatment outcome using the official reporting system
The sample size was calculated based on a difference in effect of the NTLCP.
size of 30%, power of 80%, 95% significance level, and coefficient
of variation of 0.25. Based on the average annual smear-positive Statistical analysis
case detection rate (CDR) of 41% (unpublished review of three We used Microsoft excel and SPSS for Windows 14 (SPSS Inc,
years of DOTS in the study area; the national CDR was 29%), we Chicago, USA) for data entry and analysis. We analysed the data
calculated the number of clusters required per group with 30% on the basis that all TB patients in the intervention kebeles
contingency. Based on the principle of allocating an unequal intended to use community-based case finding and treatment. We
number of clusters for randomization [9], we allocated 30 kebeles described the patients by age, sex, season and treatment outcome.
to the intervention and 21 kebeles to the control group. We calculated summary values of case detection and treatment
success rates for each kebeles. We used independent sample t test,
Randomization: generation and implementation weighted by cluster size, to compare the mean CDR and TSR
Before starting the intervention, we explained the aim of the using kebele as a unit of analysis. This is robust for cluster level
study to the programme coordinators of the districts and health analysis of binary outcomes [9]. The intra-cluster correlation
facilities. After we obtained their consent, we used the list of coefficient was calculated using one-way analysis of variance
kebeles in the two districts and randomly allocated them to [10,11].
Figure 1. Map of the study area in Sidama zone in south Ethiopia. Shaded area - Intervention kebeles. White area with black box - Control
kebeles. Red box - Health centers and health stations.
doi:10.1371/journal.pone.0005443.g001
Results (89.3% vs 83.1%, p = 0.012). Similarly, the mean TSR for females
was higher in the intervention than control kebeles (89.8 vs 81.3%,
Participants flow, recruitment and number analysed p = 0.05) as shown in Table 3.
In a year, the number of pulmonary TB suspects examined was
723 from intervention and 328 from control kebeles. Among these, Discussion
230 and 88 smear-positive patients were identified from the
intervention and control kebeles, respectively. All the smear- Interpretation and overall evidences
positive patients were analyzed (figure 2). We showed that involving HEWs in TB control improved the
smear-positive CDR and TSR in the intervention kebeles. Both
Baseline data the CDR and TSR were higher for female patients in the
Of the 51 kebeles included in the study, 30 were intervention intervention kebeles.
kebeles with a population of 178,138 and mean kebele population DOTS uses passive case finding to detect TB cases, through
of 5938 people, while 21 were control kebeles with a population of health education and tracing contacts of index cases [6]. However,
decades after implementing the strategy, smear-positive CDR has
118,673 and mean kebele population of 5651 people. 53.4% (123/
remained far below the target. In particular, the trend in CDR was
230) of patients from intervention and 42% (37/88) from control
consistently low for women, to the extent that passive case finding
kebeles were female (Table 1).
seems to favour men [12,13,14,15]. The reasons are low health
service coverage, shortage of trained health workers and poor
Outcomes and estimation health seeking behaviour [3,16,17]. Alternatively the advantage of
Patients from control kebeles were younger than those from active case finding in improving case detection is limited due to the
intervention kebeles (26 vs 29 years, p = 0.011). The mean CDR associated high cost in resource-constrained settings [5,18].
was higher in intervention kebeles (122.2% vs 69.4%, p,0.001) Moreover, neither rapid community surveys [19,20] nor commu-
and for female patients (149.0% vs 91.6%, p,0.001) (Table 2). nity DOT [21,22] seems to improve CDR.
Among the 230 patients from the intervention kebeles, 172 In our study, community-based case finding significantly
(74.8%) were cured, 33 (14.3%) completed treatment, eight (3.5%) improved the CDR for all age groups more for women than for
died, two (0.9%) had treatment failure, 15 (6.5%) defaulted and no men. The increase in CDR was lower for children compared to
patient was transferred out. Of the 88 patients in the control those aged 15 years and above. This could be explained by an
kebeles, 60 (68.2%) were cured, 14 (15.9%) completed treatment, inability to produce sputum specimens, low disease burden, or the
two (2.3%) died, nine (10.2%) defaulted, three (3.4%) were low number of children enrolled in the study [23]. Patients from
transferred out, and none had treatment failure (Figure 2). The the intervention kebeles were older than those from control kebeles
mean TSR was higher in the intervention than control kebeles for both sexes. This may have been caused by poor access,
Figure 2. Trial profile for smear-positive TB case finding and treatment outcome.
doi:10.1371/journal.pone.0005443.g002
Table 1. The baseline characteristics of the study area and poverty, and low health seeking behaviour that might have
smear-positive tuberculosis cases of southern Ethiopia 2006/ hindered them from coming to the health facilities.
07. Routine surveillance reports have repeatedly shown higher
CDRs for men than women [24]. However, in our study, the CDR
was higher for females in the intervention group. This could be
Variable Intervention Control explained by the improved geographic and socioeconomic access
to the service as sputum collection was done in the intervention
Communities
kebeles. As expected, the number of TB cases detected was greater
Number of clusters 30 21 than that estimated. This may have resulted from underestimation
Study population 178, 138 118, 673 of TB incidence as reported from Myanmar [8], the backlog of TB
Male 91,206 63,464 cases that were not reached by the health service [19,24], or
Female 86,932 55,209
underestimation of the population in the study area. Further study
is required to determine the magnitude of TB in the community.
Mean kebele population 5938 5651
Studies have shown that using different treatment supervisors
Male 3040 3022 for DOT has improved the TSR for passively detected TB cases
Female 2898 2629 [25,26,27]. However, poor treatment adherence remains a
Smear-positive TB patients challenge for patients identified by active and enhanced case
Mean age (SD) 29 (13) 26 (11)
finding [28]. In our study, decentralisation of the treatment to the
kebele improved the TSR for TB patients detected by enhanced
Male 29 (13) 26 (13)
case finding. Similar to CDR, the TSR was higher for women
Female 29 (13) 24 (8) aged above 14 years because of improved access created by DOT
Number (%) of TB cases by sex provision in the kebele.
Male 107 (46.6) 51 (58) Our findings suggest seasonal variation in CDR and TSR. In
Female 123 (53.4) 37 (42) the intervention kebeles, the rates peaked in spring (September–
November) and winter (December–February) possibly as a result
Number (%) of TB cases by age(in years)
of the economic gain from the harvest in spring. However, in the
#14 23 (10.0) 9 (10.3)
intervention and control kebeles, the rates were low in autumn
15–24 63 (27.4) 34 (39.1) (March–May) when farmers prepare for the farming season, and
25–34 72 (31.3) 28 (32.2) this was followed by another peak in early summer (June–August).
35–44 58 (25.2) 13 (14.9) Previous studies have suggested that overcrowding and staying
45–54 14 (6.1) 3 (3.4)
indoors during the rainy season favour transmission of TB, which
results in greater seasonal variation in children [29,30,31]. In our
Number (%) of TB cases by season
setting, further study is required to establish more about the
Spring 55 (23.9) 29 (33.0) seasonal variation and its associated factors.
Winter 69 (30.0) 18 (20.4) Although cluster randomized controlled trials are considered
Autumn 45 (19.6) 22 (25.0) valid studies, their methodological limitations should be addressed.
Summer 61 (26.5) 19 (21.6) The baseline demographic and clinical characteristics were similar
in the two groups. We kept potential for bias to a minimum by
doi:10.1371/journal.pone.0005443.t001 comparing and analysing information from complementary
Table 2. Case detection rates of smear-positive tuberculosis cases in southern Ethiopia, 2006/07.
*
ICC - intraclass correlation coefficient.
{
CDR - case detection rate.
doi:10.1371/journal.pone.0005443.t002
Table 3. Treatment success rates of smear-positive tuberculosis cases in southern Ethiopia, 2006/07.
*
ICC - intraclass correlation coefficient.
{
treatment success rate.
doi:10.1371/journal.pone.0005443.t003
sources. Although we did not blind the sputum samples sent to the especially laboratory technicians, with or without adequate health
diagnostic units, the laboratory technicians received the standard service coverage. The findings of the study were disseminated to
information for sputum analysis, and were not informed if the managers of TB programmes in the southern region and at
sputum specimens were from an intervention or control kebele. In national level. We believe our findings are relevant for policy
addition, external quality control was performed for the slides formulation on community TB care in Ethiopia. With limited
examined in health facilities at the Centre for Health and health care coverage and shortage of health workers, similar to
Research Laboratory in the southern Ethiopia. In our sample size that in many developing countries, we believe that our findings are
calculations, the intraclass correlation was small. The CDR and applicable to similar settings.
TSR in the control group was 45% and 83%, respectively, which In conclusion, involving HEWs in TB control improved the
was similar to the CDR of 40% in the intervention and 42% in the CDR and TSR for smear-positive patients and females in
control kebeles, and TSR of 78% in the intervention and 74% in particular. It could be used as an option to improve the trend in
the control kebeles (unpublished review of 3 years of DOTS in the low CDR and provide patient-centred services in high-burden
study area), which suggests the completeness of our data collection. countries. However, the cost-effectiveness of enhanced case finding
However, as control and intervention kebeles were neighbouring and treatment outcome needs further study.
each other and health facilities delivered the service to both
groups, we cannot rule out the effect of the intervention in the
control kebeles. This might have reduced the effect size in the Supporting Information
intervention kebeles. Checklist S1 CONSORT Checklist
Our intervention used enhanced case finding, a variant of active Found at: doi:10.1371/journal.pone.0005443.s001 (0.06 MB
case finding, in which HEWs encouraged TB suspects to visit DOC)
health posts for sputum collection, and provided DOT. The
strength of the study was that it included a sufficient number of Protocol S1 Trial Protocol
clusters to address an important challenge of DOTS strategy, Found at: doi:10.1371/journal.pone.0005443.s002 (0.27 MB
namely low CDR, and improved treatment adherence of patients DOC)
identified by enhanced case finding by providing DOT. It also
explored a practical way of involving HEWs in TB control under Acknowledgments
the community-based initiative of HEP in Ethiopia.
We are grateful to Regional Health Bureau, Sidama Zone Health
Department, Dale and Wonsho Woreda Health Office and TB programme
Generalizability coordinators for their technical and material support. We are grateful to
The results of our study could be applied in settings with low health workers and laboratory technicians in the health facilities and
health service coverage (low DOTS coverage and limited number HEWs in the intervention kebeles. We are also thankful to TB patients who
of TB laboratories), where HEWs have the first contact with the voluntarily participated in the study. Special thanks go to Dr. Estifanos
people to provide health education, and collect and transport Biru for his technical support and provision of resources for data analysis.
sputum specimens to diagnostic units. This makes the service
patient-centred, to improve case finding and treatment adherence Author Contributions
[22]. Our study area is a densely populated agrarian community, Conceived and designed the experiments: DGD BL. Performed the
typical of the rural population on the Ethiopian highlands. It could experiments: DGD. Analyzed the data: DGD BL. Wrote the paper: DGD
also be applied in areas with a shortage of health workers, BL. Supervised the conduct of the experiment: BL.
References
1. Harries A, DermotMaher, StephanGraham (2004) TB/HIV: A Clinical 17. Wahyuni C, Budiono X, Rahariyani LD, Sulistyowati M, Rachmawati T, et al.
Manual. Geneva, Switzerland: WHO. 210 p. (2007) Obstacles for optimal tuberculosis case detection in primary health
2. Subramani R, Radhakrishna S, Frieden TR, Kolappan C, Gopi PG, et al. (2008) centers (PHC) in Sidoarjo district, East Java, Indonesia. BMC Health Serv Res
Rapid decline in prevalence of pulmonary tuberculosis after DOTS implemen- 7: 135.
tation in a rural area of South. India Int J Tuberc Lung Dis 12: 916–920(915). 18. Enarson DA (2000) World tuberculosis control: how far have we to go?
3. Dye C, Watt CJ, Bleed DM, Williams BG (2003) What is the limit to case Int J Tuberc Lung Dis 4: 219–223.
detection under the DOTS strategy for tuberculosis control? Tuberculosis 83: 19. Shargie EB, Morkve O, Lindtjorn B (2006) Tuberculosis case-finding through a
35–43. village outreach programme in a rural setting in southern Ethiopia: community
4. Murray CJ, Salomon JA (1998) Expanding the WHO tuberculosis control randomized trial. Bull World Health Organ 84: 112–119.
strategy: rethinking the role of active case finding. Int J Tuberc Lung Dis 2: 20. Santha T, Renu G, Frieden TR, Subramani R, Gopi PG, et al. (2003) Are
S9–S15. community surveys to detect tuberculosis in high prevalence areas useful?
5. Golub JE, Mohan CI, Comstock GW, Chaisson RE (2005) Active case finding of Results of a comparative study from Tiruvallur District, South India. Int J Tuberc
tuberculosis: historical perspective and future prospects. Int J Tuberc Lung Dis Lung Dis 7: 258–265.
9: 1183–1203. 21. Omaswa F (2003) The Community TB Care in Africa. Int J Tuberc Lung Dis 7:
6. Federal Ministry of Health of Ethiopia (2008) Tuberculosis, Leprosy and TB/ S1–S102.
HIV Prevention and Control Manual. Addis Ababa, Ethiopia: Ministry of 22. World Health Organization (2003) Community TB care: Practice and Policy.
Health. 23. Shargie EB, Yassin MA, Lindtjorn B (2006) Prevalence of smear-positive
7. Federal Ministry of Health of Ethiopia (2005) Essential health Service Package pulmonary tuberculosis in a rural district of Ethiopia. Int J Tuberc Lung Dis 10:
for Ethiopia, Health FMo, editor. Addis Ababa: Artistic printing Enterprise. 87–92.
8. World Health Organization (2008) Global Tuberculosis Control, surveillance,
24. World Health Organization (2007) Global Tuberculosis Control, surveillance,
planning and financing. Geneva, Switzerland: World Health Organization. pp
planning and financing. Geneva, Switzerland. pp 1–270.
1–304.
25. Wandwalo E, Kapalata N, Egwaga S, Morkve O (2004) Effectiveness of
9. Donner A, Klar N (2000) Design and analysis of cluster randomization trials in
community-based directly observed treatment for tuberculosis in an urban
health research. London: Arnold.
10. Donner A, N. K (1994) Methods for comparing event rates in intervention setting in Tanzania: a randomised controlled trial. Int J Tuberc Lung Dis 8:
studies when the unit of allocation is a cluster. Am J Epidemiology 140: 279–289. 1248–1254.
11. Ukoumunne OC, Gulliford MC, Chinn S, Sterne JA, Burney PG (1999) 26. Singh AA, Parasher D, Shekhavat GS, Sahu S, Wares DF, et al. (2004)
Methods for evaluating area-wide and organization-based interventions in health Effectiveness of urban community volunteers in directly observed treatment of
and health care: a systematic review. Health Technol Assess 3: 92. tuberculosis patients: a field report from Haryana, North India. Int J Tuberc
12. Kivihya-Ndugga LE, van Cleeff MR, Ng’ang’a LW, Meme H, Odhiambo JA, et Lung Dis 8: 800–802.
al. (2005) Sex-specific performance of routine TB diagnostic tests. Int J Tuberc 27. Kangangi JK, Kibuga D, Muli J, Maher D, Billo N, et al. (2003)
Lung Dis 9: 294–300. Decentralisation of tuberculosis treatment from the main hospitals to the
13. Agboatwalla M, Kazi GN, Shah SK, Tariq M (2003) Gender perspectives on peripheral health units and in the community within Machakos district, Kenya.
knowledge and practices regarding tuberculosis in urban and rural areas in Int J Tuberc Lung Dis 7: S5–13.
Pakistan. East Mediterr Health J 9: 732–740. 28. Golub JE, Mohan CI, Comstock GW, Chaisson RE (2005) Active case finding of
14. Ahsan G, Ahmed J, Singhasivanon P, Kaewkungwal J, Okanurak K, et al. (2004) tuberculosis: historical perspective and future prospects. Int J Tuberc Lung Dis
Gender difference in treatment seeking behaviors of tuberculosis cases in rural 9: 1183–1203.
communities of Bangladesh. Southeast Asian J Trop Med Public Health 35: 29. Leung CC, Yew WW, Chan TY, Tam CM, Chan CY, et al. (2005) Seasonal
126–135. pattern of tuberculosis in Hong Kong. Int J Epidemiol 34: 924–930.
15. Hamid Salim MA, Declercq E, Van Deun A, Saki KA (2004) Gender differences 30. Thorpe LE, Frieden TR, Laserson KF, Wells C, Khatri GR (2004) Seasonality
in tuberculosis: a prevalence survey done in Bangladesh. Int J Tuberc Lung Dis of tuberculosis in India: is it real and what does it tell us? Lancet 364:
8: 952–957. 1613–1614.
16. Dye C, HosseiniMehran, WattCatherine (2007) Did we reach the 2005 targets 31. Luquero FJ, Sanchez-Padilla E, Simon-Soria F, Eiros JM, Golub JE (2008)
for tuberculosis control? Bulletin of the World Health Organization 85: Trend and seasonality of tuberculosis in Spain, 1996–2004. Int J Tuberc Lung
364–369. Dis 12: 221–224.
Abstract
Background: Evidence for policy- and decision-making related to the cost of delivering tuberculosis (TB) control is lacking in
Ethiopia. We aimed to determine the cost and cost-effectiveness of involving health extension workers (HEWs) in TB
treatment under a community-based initiative in Ethiopia. This paper presents an ancillary cost-effectiveness analysis of
data from a RCT, from which the main outcomes have already been published.
Methodology/Principal Findings: Options of treating TB patients in the community by HEWs in the health posts and
general health workers at health facility were compared in a community-randomized trial. Costs were analysed from a
societal perspective in 2007 in US dollars using standard methods. We prospectively enrolled smear-positive patients, and
calculated the cost-effectiveness in terms of the cost per patient successfully treated. The total cost for each successfully
treated smear-positive patient was higher in health facilities (US$161.9) compared with the community-based approach
(US$60.7). The total, patient and care giver costs of community-based treatment were lower than health facility DOT by
62.6%, 63.9% and 88.2%, respectively. Involving HEWs added a total cost of US$8.80 to the health service per patient treated
in the health posts in the community.
Conclusions/Significance: Community-based treatment by HEWs costs only 37% of what treatment by general health
workers costs for similar outcomes. Involving HEWs in TB treatment is a cost-effective treatment alternative to the health
service and to the patients and their caregivers. Therefore, there is both an economic and public health reason to consider
involving HEWs in TB treatment in Ethiopia. However, community-based treatment would require initial investment for
implementation, training and supervision.
Citation: Datiko DG, Lindtjørn B (2010) Cost and Cost-Effectiveness of Treating Smear-Positive Tuberculosis by Health Extension Workers in Ethiopia: An Ancillary
Cost-Effectiveness Analysis of Community Randomized Trial. PLoS ONE 5(2): e9158. doi:10.1371/journal.pone.0009158
Editor: J. Jaime Miranda, London School of Hygiene and Tropical Medicine, Peru
Received May 23, 2009; Accepted January 13, 2010; Published February 17, 2010
Copyright: ! 2010 Datiko, Lindtjørn. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: University of Bergen, Centre for International Health funded the study. The funders had no role in study design, data collection and analysis, decision to
publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: [email protected]
Introduction Under the NTLCP, there are three levels of function for
coordinating TB control in hospitals and health centres: regions,
Ethiopia has one of the highest tuberculosis (TB) burdens in zones and districts. TB diagnosis and treatment is provided by
the world [1]. Directly observed treatment short course (DOTS), general health workers (GHWs) in health facilities [5]. The
the World Health Organization (WHO) recommended TB treatment success rate (TSR) of smear-positive cases in the study
control strategy, was started in 1995 in Ethiopia by the National area was 76% (unpublished three year review of TB programme
TB and Leprosy Control Programme (NTLCP), being decen- performance in the study area), while 84% at the National level.
tralized to hospitals and health centres [2]. However, less than The case detection rate (CDR) of smear-positive cases was 41% in
half of the population has access to the health service [3]. Thus, the study area and 27% at the National level, far below the target
many TB patients remain undiagnosed, untreated and continue of 70%. However, the cost implication for the health service and
transmitting the infection. The interaction between TB and the community has not been estimated.
human immunodeficiency virus (HIV) infection has fuelled the In 2004, the Government of Ethiopia launched a community-
TB burden and affected the already overstretched health service, based initiative focused on disease prevention and health
which needs alternative ways of making the service accessible promotion to ensure equitable access to health service. To this
[4]. end, health extension workers (HEWs) were trained and deployed
to each kebele (the lowest administrative unit in Ethiopia) to their caregivers or freed by the programme. Weighted mean cost
provide health service. The HEWs function from an operational was calculated to costs related to patients and caregivers. In this
unit - a health post in each kebele, receiving training on TB as study, hereafter, cost values refer to mean cost values per
part of communicable disease prevention and control [6]. successfully treated smear-positive TB patient.
However, community DOTS was not yet implemented and Programme costs. Programme costs are the health service
HEWs were not providing directly observed treatment (DOT) to costs including the expenses required to establish the health
TB patients [5]. service, and run the TB programme in the districts and health
Studies show that involving community health workers in TB facilities including the health posts in the kebeles. The average cost
control is cost-effective in improving the treatment success for each component of treatment (drugs, sputum examination,
compared with health facility-based DOTS [7,8,9,10,11,12, treatment and other medical expenses) was calculated from the
13,14,15]. Community DOTS requires supervision, some initial quantity and unit prices of resources. Time costs were estimated
investment and a well-coordinated TB control programme. Its from the health facility providing DOT to the patient’s place of
effectiveness therefore depends on how well the health system residence. Joint costs (cost items shared by two or more services)
functions in coordination with the community [16]. were allocated to TB patients based on the proportion of the total
We conducted a community randomized trial (CRT) in health facility visits which they accounted for and the associated
Southern Ethiopia to determine whether involving HEWs in TB health workers time. Annuitization was done on the basis of the
control would improve smear-positive CDR and TSR compared expected useful life of 30 years for buildings, 10 years for cars and
with health facility-based TB treatment. We found both improved equipment and 5 years for motorcycles [19]. The base year for
CDR (122% vs. 69%, p,0?001) and TSR (89% vs 83%, p = 0?012) valuing costs was 2007, and the exchange rate was 8.6 Ethiopian
in the community-based DOT (CDOT) compared with the health Birr to US $1.
facility-based DOT (HFDOT) [17]. Therefore, determining The cost of HEWs, part of the health service, included the time
whether involving a community-based approach is also more spent for treatment supervision in the kebele, travel time and
cost-effective would seem a relevant issue for policy- and decision expenditures associated with visits to the health facilities to collect
making. To our knowledge, there have been no studies of cost and drugs. The time costs were converted to a monetary value based
cost-effectiveness of alternative ways of treating TB in Ethiopia. In on the monthly income of HEWs in US dollars. The cost of
this study, we aimed to determine the cost and cost-effectiveness of training and supervision was also included.
involving HEWs in TB treatment in Southern Ethiopia. This Patient costs. Patient costs include the costs related to the
paper presents an ancillary cost-effectiveness analysis of data from TB patient and their caregiver. The costs were estimated for the
a RCT, from which the main outcomes have already been smear-positive patients and their caregivers using a structured
published. questionnaire. TB patients and the caregivers were asked about
the travel time and expenses associated with visits to HEWs in the
Methods health post to take drugs. This included transport, food and other
costs. The cost data was collected for all caregivers who
The protocol for this trial and supporting CONSORT checklists accompanied the patients to health centres and health posts.
are available as supporting information; see Checklist S1, Travel time was estimated from the patient’s home to the health
Flowchart S1 and Protocol S1. Full description of trial method- post in the kebele. The time costs were converted to a monetary
ology is given in the paper reporting main trial findings [17]. value based on unskilled wage rates [18] which was US$1.39 per
Briefly, two treatment options of treating smear-positive patients day (US$0.17 per hour) in the study area.
were compared: health facility and community DOT (CDOT - the The cost data was case specific for all study participants and was
intervention). standard in each arm of the intervention. At least ten visits in the
intensive phase and six visits in the continuation phase were used
Health Facility-Based DOT (HFDOT) as a standard for smear-positive patients and care givers for both
TB patients receive treatment under the direct observation of the CDOT and HFDOT.
GHWs in hospitals and health centres. They visit health facilities For each treatment option, average costs were multiplied by the
daily for two months during the intensive phase. During the number of times each cost was incurred to calculate the cost per
continuation phase, patients visit health facilities once a month to patient successfully treated. For each kebele, we calculated
collect drugs but take the drugs unsupervised. summary values of costs and then used an independent sample t
test, weighted by cluster size, to compare the mean costs using
Community DOT (CDOT): The Intervention kebele as a unit of analysis.
TB patients visit the health post daily for two months during the The data sources were budget and expense files of the districts
intensive phase to receive treatment under the direct observation finance and health offices, health facilities, health workers’
of HEWs in their kebele. During the continuation phase, patients payroll, drug and supply prices, funds used from research
collect drugs from the HEWs on a monthly basis. projects (training, supervision and review of activities), TB
Trained HEWs and GHWs prospectively collected the cost data control programme, bank reports and interview of the study
by using a structured questionnaire. GHWs also used a checklist to participants.
observe the conduct of DOT in the health facilities and the
kebeles. Effectiveness
The measure of effectiveness was based on sputum smear
Costing results at the end of the 2nd, 5th, and 7th months of treatment.
Costs were assessed from a societal perspective in 2007 in US Patients with at least two negative smears including the smear at
dollars, using recommended standard methods [18]. We classified the 7th month were reported as cured. Patients who finished the
costs in to programme and patient costs. Direct cost refers to treatment but did not have the 7th month smear result were
patient’s out-of-pocket expenses for seeking treatment, while reported as treatment completed. We used TSR as a measure of
indirect cost refers to the cost of the time spent by the patient or effectiveness, which is a standard indicator used by WHO to
measure programme success and which has been adopted by the Table 1. Baseline characteristics of smear-positive
NTLCP in Ethiopia. TSR was calculated as the sum of the tuberculosis patients in Southern Ethiopia.
number of TB patients who were cured and the number of TB
patients for which treatment was completed divided by the total
number of smear-positive cases reported, expressed as a Variable Community DOT Health facility DOT
percentage [1,5]. The effectiveness data was obtained from
Mean age (SD) 26.8 (13.7) 25.2 (11.8)
CRT. Briefly, we calculated the summary values of TSRs for
each kebele. Then we used an independent sample t test, Gender
weighted by cluster size, to compare the mean TSR using the Men 62 (38.5%) 43 (63.2%)
kebele as a unit of analysis. This is robust for cluster level analysis Women 99 (61.5%) 25 (36.8%)
of binary outcomes [20]. Education
In a no intervention scenario (‘‘do-nothing alternative’’), a self- Illiterate 55 (37.2%) 27 (45.0%)
cure rate of 20% was used but 0% for HIV infected TB patients.
Literate 93 (62.8%) 33 (55.0%)
The reported rate of TB-HIV co-infection in southern Ethiopia
was 17.5% [21]. The self-cure rate was calculated using the Missing 13 8
following formula: [(estimated percentage of HIV+ patients x 0) + Occupation
(estimated percentage of TB patients who are not HIV infected x Student 38 (32.2%) 16 (26.7%)
20)]/100 [22]. Farmer 34 (28.8%) 24 (40.0%)
Housewife 38 (32.2%) 18 (30.0%)
Cost-Effectiveness Others 8 (6.8%) 2 (3.3%)
Cost-effectiveness was calculated as the average cost per patient
Missing 42 8
treated successfully. This was done by dividing the total cost by the
number of TB patients successfully treated for each of the two Marital status
treatment options, the CDOT and HFDOT. Single 51 (34.5%) 19 (28.8%)
Married 93 (62.8%) 47 (71.2%)
Sensitivity Analysis Widowed/divorced 4 (2.7%) 0 (0.0%)
Sensitivity analysis determines the level of uncertainty in the Treatment outcome*
components of the evaluation by repeating the comparison
Cured 132 (82.0%) 56 (82.4%)
between cost items and consequences while varying the assump-
Treatment completed 29 (18.0%) 12 (17.6%)
tions underlying the estimates. A one-way sensitivity analysis varies
one cost item at a time while others are held at base value to *A smear-positive tuberculosis patient with at least two negative smears
measure its impact on the results of the evaluation [18,23]. We including that at 7th month was reported as cured, while a patient who finished
performed one-way sensitivity analysis to assess the robustness of the treatment but did not have the 7th month smear result was reported as
treatment completed.
the results to changes in the cost values. We varied one cost
doi:10.1371/journal.pone.0009158.t001
variable at a time, repeating the analysis for the cost items. We
based the uncertainty analyses on the minimum and maximum
values of mean travel time, transport and total cost in our study. Costs
We used the 95% confidence interval of the effectiveness for the Programme costs: the health service and health extension
treatment outcome. workers costs. The health service invested US$73.5 for
HFDOT and US$7.9 for CDOT. The cost of anti-TB drugs for
Ethical Clearance a patient was US$22.1. The cost of training was US$10.0 in
The Ethical Review Committee of Southern Nations, Nation- HFDOT and US$5.1in CDOT. Similarly, the cost of supervision
alities and Peoples’ Regional Health Bureau approved the study. was US$10.9 in HFDOT and US$5.9 in CDOT.
We first discussed the aim of the study with the TB programme The travel time (estimated travel cost) for HEWs was 19.7 hours
managers and kebele leaders about community- based TB care (US$5.1). The transport and food costs were US$0.9 and US$2.8,
and obtained permission to proceed. Then we explained the aim respectively. Therefore, the HEWs total cost per patient was
of the study to the study participants and enrolled them after US$8.8, accounting for 14.3% of the total cost per patient for
obtaining informed consent. The study participants were also CDOT.
informed about the right to refuse or withdraw from the study. Patient costs: the patient and caregiver costs. The
The Ethical Review Committee approved verbal consent, in patient costs are described as follows. The mean and standard
adherence to NTLCP recommendations. deviation (SD) of travel time (estimated travel cost) was 27.6
hours (US$4.3, SD = 1.9) in CDOT and 68.9 hours (US$11.9,
Results SD = 5.2) in HFDOT (p,0.05). The transport cost was US$0.6
(SD = 1.2) in CDOT and US$3.7 (SD = 10.5) in HFDOT
Two hundred and twenty-nine smear-positive patients were (p = 0.013). Similarly, the associated food cost was (US$3.5,
enrolled. We interviewed 161smear-positive patients and 113 care SD = 2.9) in CDOT and US$8.8 (SD = 5.2) in HFDOT. The
givers in the CDOT and 68 smear-positive patients and 97 direct patient cost was lower in CDOT (US$4.1, SD = 3.0) than
caregivers in HFDOT. More women were enrolled in CDOT HFDOT (US$12.1, SD = 10.7) (p,0.05). The total patient cost
62% (99 of 161 patients) than HFDOT 37% (25/68). Regarding was lower in CDOT (US$8.4, SD = 3.9) than HFDOT (US$24.4,
literacy, 63% of the patients (93/148) from community and 55% SD = 12.2) (p,0.05). The total cost in CDOT was lower than HF
(33/60) from facility were literate. Regarding marital status, 63% DOT by 63.9% (Figure 1).
patients (93/148) from community and 71% (47/66) from facility The caregiver costs are described as follows. The travel time
were married (Table 1). (estimated travel cost) was 9.9 hours (US$1.6, SD = 1.5) in CDOT
Cost-Effectiveness
The cost per successfully treated patient was US$161.9 and
US$60.7 in HFDOT and CDOT, respectively. CDOT reduced
the total cost per successfully treated patient by 62.6% (Table 2
and figure 3). Based on the cost and effect estimates of no
intervention (US$0, 17%), HFDOT (US$161.9, 83.1%) and
CDOT (US$60.7, 89.3%), the incremental cost-effectiveness ratio
of running HFDOT and CDOT from a do-nothing alternative
was 2.4 and 0.8, respectively. The incremental cost-effectiveness
ratio of HFDOT to CDOT was -16.3.
Sensitivity Analysis
The sensitivity analysis showed CDOT to be a more effective
Figure 1. Tuberculosis patient costs under DOT options. Blue and less costly approach compared to HFDOT on varying
bar - Health facility DOT. Red bar - Community DOT.
estimates of the main cost items (Table 3).
doi:10.1371/journal.pone.0009158.g001
Discussion
and 16.5 hours (US$4.7, SD = 5.7) in HFDOT (p,0.05). The
transport cost was lower in CDOT (US$0.1, SD = 0.9) than Interpretations and Overall Evidence
HFDOT (US$14.2, SD = 43.6) (p = 0.006). Similarly the associat- In Ethiopia, DOTS coverage was reported to be 100%
ed food cost was lower in CDOT (US$0.8, SD = 1.2) than (implemented in all hospitals and health centres). However, case
HFDOT (US$2.1, SD = 2.8) (p,0.05). The total caregiver cost finding and TSRs are below the WHO targets [1]. In such a
was lower in CDOT (US$2.5, SD = 2.7) than HFDOT (US$21.1, setting, it is relevant to ask how this could be improved.
SD = 50.6) (p = 0.002). The total care giver cost in CDOT was Improvement may be achieved by more efficient intervention for
lower than HF DOT by 88.2% (Figure 2). identifying TB cases, including providing treatment at a lower cost
[24]. In our study, the cost of treating a patient in health facilities
The total cost (patient and programme cost) per successfully
was 2.7 times higher than the cost of treating a patient in the
treated smear-positive patient was higher in HFDOT (US$161.9)
community-based approach inclusive of the initial investment for
compared to CDOT (US$60.7). The total cost in CDOT was
implementation, training and supervision of CDOT. CDOT
lower than HF DOT by 62.6% (Table 2 and figure 3).
improved the TSR by 6.2% and reduced the cost of treating a
patient by 62.6%. This shows that more patients could be
Effectiveness successfully treated with the same amount of resources by using
In the CRT, smear-positive patients received DOT, 230 under CDOT instead of HFDOT.
HEWs in the community and 88 under GHWs in health facilities. The main reason for the reduction in cost of the community-
In the community-based approach, of the 230 patients, 172 based approach was the reduction in the travel distance and
(74?8%) were cured and 33 (14?3%) completed treatment. Of the related costs as the patients visited the HEWs in the health post,
88 patients treated in the health facilities, 60 (68?2%) were cured which was located nearer to where the patients lived. The
and 14 (15?9%) completed treatment. The mean TSR was higher reduction in caregiver and patient costs results in a slight increase
in CDOT (89.3%) than HFDOT (83?1%). The mean and its over the health service cost. However, from a societal perspective,
difference being 6.2% (1.4% - 10.9%, p = 0?012). The details are the gain in terms of cost and health benefits is huge. Thus,
given elsewhere [17]. Based on the reported 17.5% TB-HIV co- involving HEWs in TB treatment is an attractive economic option
infection rate in smear-positive patients in Southern Ethiopia [21] to the health service and to the patients and their caregivers.
and using the formula to calculate self-cure, cure without Decentralization of the DOTS programme improves the TSR
treatment (given above in the methods section under ‘effective- [25,26]. A community-based approach is found to be more
ness’), we found TSR of 80.8% for HFDOT and 86.9% for effective and cost-effective as it overcomes the limitation of
CDOT. reliance on health facilities in providing access to TB care
[27,28,29]. It also consistently reduces the cost of treatment even
in a decentralized health service [16]. In our study, the cost per
successfully treated patient was low (US$61) compared with studies
from Malawi (US$201) and Botswana (US$1657). Similarly, the
reduction in average cost per patient treated in our study was 63%
compared to those reported for South Africa (36%) and Kenya
(65%) [11,12,13,14]. The main reason could be that Ethiopia is a
low-cost country with low salaries. Also, we did our study in a rural
setting as opposed to an urban setting.
The gain in effectiveness of the CDOT was mainly due to the
reduced travel distance that reduced the cost, and time lost on
travel to receive treatment. In settings with low health service
coverage the significance of CDOT is high. CDOT could
complement the existing health service to improve the access
and success rate TB programmes in countries like Ethiopia where
Figure 2. Caregiver costs under DOT options. Blue bar - Health CDOT has not yet been implemented on a national scale.
facility DOT. Red bar - Community DOT. The strength of the study is that the data was prospectively
doi:10.1371/journal.pone.0009158.g002 collected in CRT. We adhered to the routine care for treatment and
Table 2. Average cost per patient for treatment options of smear-positive tuberculosis patients in Southern Ethiopia 2006/07.
quantity mean unit price in US$ quantity mean unit price in US$
Programme costs
Running TB programme 1 7.9 1 73.5
Training and review meeting 1 5.1 1 10.0
Drugs and supplies 1 22.1 1 22.1
Supervision 8 5.9 3 10.9
Health extension workers cost
Direct cost of visit 8 3.7
Indirect cost 8 5.1
Total programme cost 49.8 116.5
Patient costs
TB patient
Direct cost of visits* 66 4.1 66 12.5
Indirect cost 4.3 11.9
Caregiver
Direct cost 0.9 16.3
Indirect cost 1.6 4.7
Total patient costs 10.9 45.4
Total costs 60.7 161.9
*
Patients visited the health facilities in health facility DOT and health posts in community DOT. Health extension workers visited the health facilities monthly to collect
drugs. Direct cost implies out-of- pocket expenses and indirect cost implies travel time cost.
doi:10.1371/journal.pone.0009158.t002
outcome measures as recommended by the NTLCP, which did not training new HEWs and providing refresher training to the already
require extra visits by the patients because of the community-based trained HEWs should be considered. This also applies to the general
approach. We used HEWs living in the kebeles, who were employed health workers involved in TB control in health centres and hospitals
to provide a health service that favoured the sustainability of the that have higher drop outs. Therefore, the estimated cost required for
community approach in the Ethiopian health system as opposed to CDOT will still remain lower than HFDOT for similar outcomes.
other community approaches whereby the community health The government of Ethiopia has already increased the uptake
workers have been used for only short periods. We conducted our and training of HEWs in the country to ensure and deploy two
intervention under approved programme conditions and prospec- HEWS per kebele. Thus, doubling HEWs per kebele has already
tively collected cost data that reduced the chance of recall bias. We started at the end of the first year of the intervention in 2007
included all cost categories in the sensitivity analysis that reduced the before drop outs occur at least in our study area. Therefore, it only
chance of selection bias. Moreover, the long period of the observation requires training HEWs which was two days in case of our study to
(September 2006 to April 2008, i.e., 20 months) may have enrol them in community based TB control activities to achieve
contributed to the consistency of the data [18]. In our study we did the outcomes reported in our intervention.
not have drop outs of HEWs due to the fact that HEWs were selected A major limitation of the study was that we based our estimation
from the kebeles they live in. However, in the future, the possibility of on the time converted into monetary value for which there is no
Figure 3. Costs per successfully treated smear-positive tuberculosis patient. Blue bar - Health facility DOT. Red bar - Community DOT.
doi:10.1371/journal.pone.0009158.g003
Input variables Input values CER* CDOT Input values CER HFDOT
TB patient time cost 0.51 11.16 51.12 55.95 3.16 27.21 125.29 150.33
Care giver time cost 0 7.17 54.08 59.19 0 18.95 133.87 160.63
HEW{ time cost 3.63 8.94 55.04 60.24
TB patient cost 0.65 17.37 55.06 60.26 12.32 94.88 127.52 153.01
Caregiver cost 0 13.19 55.43 60.67 0 177.27 131.56 157.86
HEW cost 3.63 30.05 55.62 60.88
Discount rate 0 3% 59.83 61.68 0 3% 154.21 158.98
TSR-N{ 82.22% 91.74% 77.94 86.97
TSR-N 77.79% 89.85% 207.50 259.71
TSR1 85.11 93.15 55.74 61.00
TSR 76.45 91.73 136.27 163.51
*
Cost-effectiveness ratio.
{
Health extension worker.
{
Treatment success rate excluding self-cure.
1
Treatment success rate.
doi:10.1371/journal.pone.0009158.t003
agreement among experts [18]. It was also difficult to get reliable money in health facilities, at least two smear-positive patients
income data for rural areas. Therefore, we based our estimate on could be treated under a community-based approach. There are
the wage of unskilled labour. This might have underestimated the both economic and public health reasons to consider involving
time cost. In treatment outcomes like deaths and defaulters the cost HEWs in TB treatment by the NTLCP of Ethiopia. However, due
was not captured for the total follow up period due to the nature of attention should be paid to ensuring initial start up investment to
the outcome. In such cases, the distance from the health institutions implement CDOT, training and supervision.
and the related high cost could be the plausible explanations for
such outcomes mainly in HFDOT. Therefore, the cost of treating Supporting Information
smear-positive patient could be on the lower side, an underestimate,
in both arms but mainly in HFDOT where distance and related cost Checklist S1
was high. Using one-way sensitivity analysis, where we varied one Found at: doi:10.1371/journal.pone.0009158.s001 (0.06 MB
cost item at a time, might not have captured the interaction between DOC)
cost items. The economic and public health benefit of treating TB Flowchart S1
patients in terms of disease transmission, averting death or
Found at: doi:10.1371/journal.pone.0009158.s002 (0.03 MB
increasing productivity was not the scope of the study.
DOC)
Generalizability Protocol S1
Our study area was a densely populated agrarian community in Found at: doi:10.1371/journal.pone.0009158.s003 (0.44 MB
Ethiopia. This area is typical of the rural population of Ethiopia, DOC)
representing 85% of the total population where, high treatment
success rates are not achieved because of the limited health service Acknowledgments
coverage and shortage of health workers. With health posts in each
We are grateful to the Regional Health Bureau, Sidama Zone Health
kebele and the huge number of HEWs, more cost-effective
Department, Dale and Wonsho Woreda Health Office, and TB
approaches are needed. As opposed to the study period where programme coordinators for their technical and material support. We
there was only one HEW per kebele, now two HEWs are deployed are grateful to the health workers and laboratory technicians in the health
to rural kebeles in Ethiopia. Thus, we believe that our findings are facilities and HEWs in the intervention kebeles who participated in the
applicable in similar settings. For example our approach could be study. We also thank the TB patients who voluntarily participated in the
adopted in other regions or countries where two HEWs work in study.
each rural kebele. In addition, the Federal Ministry of Health of
Ethiopia has assigned full-time public health nurses as supervisors Author Contributions
of HEWs that favour implementation of CDOT. We presented the
Conceived and designed the experiments: DGD BL. Performed the
results of the study at a NTLCP review meeting. experiments: DGD BL. Analyzed the data: DGD BL. Contributed
In conclusion, community DOT costs only 37% of what reagents/materials/analysis tools: DGD. Wrote the paper: DGD BL.
HFDOT costs for similar outcomes. For the same amount of
References
1. World Health Organization (2008) Global Tuberculosis Control, surveillance, 2. Yassin MA, Datiko DG, Shargie EB (2006) Ten-year experiences of the
planning and financing. Geneva, Switzerland: World Health Organization. pp tuberculosis control programme in the southern region of Ethiopia. Int J Tuberc
1–304. Lung Dis 10: 1166–1171.
3. Federal Ministry of Health of Ethiopia (2007) Health and health related 15. Wandwalo E, Robberstad B, Morkve O (2005) Cost and cost-effectiveness of
indicators. Ministry of Health of Ethiopia. community based and health facility based directly observed treatment of
4. Maher D, Harries A, Getahun H (2005) Tuberculosis and HIV interaction in tuberculosis in Dar es Salaam, Tanzania. Cost Eff Resour Alloc 3: 6.
sub-Saharan Africa: impact on patients and programmes; implications for 16. Maher D, Hausler HP, Raviglione MC, Kaleeba N, Aisu T, et al. (1997)
policies. Trop Med Int Health 10: 734–742. Tuberculosis care in community care organizations in sub-Saharan Africa:
5. Federal Ministry of Health of Ethiopia (2008) Tuberculosis, Leprosy and TB/ practice and potential. Int J Tuberc Lung Dis 1: 276–283.
HIV Prevention and Control Manual, Ministry of Health, Ethiopia. Addis 17. Datiko DG, Lindtjorn B (2009) Health extension workers improve tuberculosis
Ababa. case detection and treatment success in southern Ethiopia: a community
6. Federal Ministry of Health of Ethiopia (2005) Essential health Service Package randomized trial. PLoS ONE 4: e5443.
for Ethiopia. Addis Ababa: Artistic printing Enterprise. 18. Drummond MF, Sculpher MJ, Torrance GW, O’Brien BJ, Stoddart GL (2005)
7. Wandwalo E, Kapalata N, Egwaga S, Morkve O (2004) Effectiveness of Methods for the Economic Evaluation of Health Care Programmes. Oxford,
community-based directly observed treatment for tuberculosis in an urban setting UK: Oxford University Press. 379 p.
in Tanzania: a randomised controlled trial. Int J Tuberc Lung Dis 8: 1248–1254. 19. (2003) Making choices in health: WHO guide to cost-effectiveness analysis:
8. Singh AA, Parasher D, Shekhavat GS, Sahu S, Wares DF, et al. (2004) World Health Organization.
Effectiveness of urban community volunteers in directly observed treatment of 20. Donner A, Klar N (2000) Design and analysis of cluster randomization trials in
tuberculosis patients: a field report from Haryana, North India. Int J Tuberc health research. London: Arnold.
21. Datiko DG, Yassin MA, Chekol LT, Kabeto LE, Lindtjorn B (2008) The rate of
Lung Dis 8: 800–802.
TB-HIV co-infection depends on the prevalence of HIV infection in a
9. Nyirenda TE, Harries AD, Gausi F, van Gorkom J, Maher D, et al. (2003)
community. BMC Public Health 8: 266.
Decentralisation of tuberculosis services in an urban setting, Lilongwe, Malawi.
22. Dye C, Garnett GP, Sleeman K, Williams BG (1998) Prospects for worldwide
Int J Tuberc Lung Dis 7: S21–28.
tuberculosis control under the WHO DOTS strategy. Directly observed short-
10. Kangangi JK, Kibuga D, Muli J, Maher D, Billo N, et al. (2003) course therapy. Lancet 352: 1886–1891.
Decentralisation of tuberculosis treatment from the main hospitals to the 23. Briggs AH, Gray AM (1999) Handling uncertainty when performing economic
peripheral health units and in the community within Machakos district, Kenya. evaluation of healthcare interventions. Health Technol Assess 3: 1–134.
Int J Tuberc Lung Dis 7: S5–13. 24. Murray CJ, Kreuser J, Whang W (1994) Cost-effectiveness analysis and policy
11. Floyd K, Skeva J, Nyirenda T, Gausi F, Salaniponi F (2003) Cost and cost- choices: investing in health systems. Bull World Health Organ 72: 663–674.
effectiveness of increased community and primary care facility involvement in 25. Mangura B, Napolitano E, Passannante M, Sarrel M, McDonald R, et al. (2002)
tuberculosis care in Lilongwe District, Malawi. Int J Tuberc Lung Dis 7: S29–37. Directly observed therapy (DOT) is not the entire answer: an operational cohort
12. Nganda B, Wang’ombe J, Floyd K, Kangangi J (2003) Cost and cost- analysis. Int J Tuberc Lung Dis 6: 654–661.
effectiveness of increased community and primary care facility involvement in 26. Ziad Obermeyer JA-K, Christopher JLMurray (2008) Has the DOTS Strategy
tuberculosis care in Machakos District, Kenya. Int J Tuberc Lung Dis 7: Improved Case Finding or Treatment Success? An Empirical Assessment. PLoS
S14–20. ONE 3.
13. Okello D, Floyd K, Adatu F, Odeke R, Gargioni G (2003) Cost and cost- 27. World Health Organization (2003) Community TB care: Practice and Policy.
effectiveness of community-based care for tuberculosis patients in rural Uganda. 28. Baltussen R, Floyd K, Dye C (2005) Cost effectiveness analysis of strategies for
Int J Tuberc Lung Dis 7: S72–79. tuberculosis control in developing countries. Bmj 331: 1364.
14. Sinanovic E, Floyd K, Dudley L, Azevedo V, Grant R, et al. (2003) Cost and 29. Islam MA, Wakai S, Ishikawa N, Chowdhury AM, Vaughan JP (2002) Cost-
cost-effectiveness of community-based care for tuberculosis in Cape Town, effectiveness of community health workers in tuberculosis control in Bangladesh.
South Africa. Int J Tuberc Lung Dis 7: S56–62. Bull World Health Organ 80: 445–450.
Address: 1Centre for International Health, University of Bergen, Overlege Danielsens Hus, Årstadveien 21, 5009 Bergen, Norway and 2Southern
Nations, Nationalities, and Peoples' Regional Health Bureau, PO Box 149, Hawassa, Ethiopia
Email: Daniel G Datiko* - [email protected]; Bernt Lindtjørn - bernt.lindtjø[email protected]
* Corresponding author
Abstract
Background: Decentralization of DOTS has increased the number of cured smear-positive
tuberculosis (TB) patients. However, the rate of recurrence has increased mainly due to HIV
infection. Recurrence rate could be taken as an important measure of long-term success of TB
treatment. We aimed to find out the rate of recurrence in smear-positive patients cured under
DOTS in southern Ethiopia.
Methods: We did a retrospective cohort study on cured smear-positive TB patients who were
treated from 1998 to 2006. Recurrence of smear-positive TB was used as an outcome measure.
Person-years of observation (PYO) were calculated per 100 PYO from the date of cure to date of
interview. Kaplan-Meier and Cox-regression methods were used to determine the survival and the
hazard ratio (HR).
Results: 368 cured smear-positive TB patients which were followed for 1463 person-years. Of
these, 187 patients (50.8%) were men, 277 patients (75.5%) were married, 157 (44.2%) were
illiterate, and 152 patients (41.3%) were farmers. 15 of 368 smear-positive patients had recurrence.
The rate of recurrence was 1 per 100 PYO (0.01 per annum). Recurrence was not associated with
age, sex, occupation, marital status and level of education.
Conclusion: High recurrence rate occurred among smear-positive patients cured under DOTS.
Further studies are required to identify factors contributing to high recurrence rates to improve
disease free survival of TB patients after treatment.
Page 1 of 5
(page number not for citation purposes)
BMC Public Health 2009, 9:348 http://www.biomedcentral.com/1471-2458/9/348
Figure
Map 1 study area in Sidama zone in the southern Ethiopia
of the
Map of the study area in Sidama zone in the southern Ethiopia.
Page 2 of 5
(page number not for citation purposes)
BMC Public Health 2009, 9:348 http://www.biomedcentral.com/1471-2458/9/348
Case definition, treatment duration and outcome TB were advised to visit health posts for sputum collection
TB patients who had productive cough for two weeks or by health extension workers or to visit diagnostic health
more with at least two positive sputum smears or one pos- institutions for examination.
itive smear and x-ray findings consistent with active PTB
were classified as smear-positive pulmonary TB cases. Results
Of the 397 smear-positive TB patients registered, 368
TB patients received two months intensive phase and six (92.7%) were followed. Incomplete information was
months continuation phase treatment. Follow-up sputum obtained for 29 (7.3%) of which 8(2.0%) had moved to
examination was done at the end of 2nd, 5th and 7th other districts. However, no difference was observed by
months of treatment. A smear-positive TB patient who age, sex, and TB category compared o the patients we
had negative sputum smear result in the last month of enrolled.
treatment and on at least one previous occasion (2nd or 5th
month) was reported as cured. The term recurrence was Of the 368 smear-positive TB patients which were fol-
used to indicate rediagnosis of smear-positive TB in lowed, 187 patients (50.8%) were men, 277 patients
patients who were reported cured [8]. (75.5%) were married, 157 (44.2%) were illiterate, and
152 patients (41.3%) were farmers (Table 1). 368 cured
Data collection smear-positive TB patients were followed for 1463 per-
From unit TB register in health institutions in the two dis- son-years. 15 of 368 smear-positive patients had recur-
tricts, we obtained the list of smear-positive TB patients rence. The mean (median) duration of follow-up was 3.87
who were declared cured from 1998 to 2006. We collected (4.0) years. The rate of recurrence was 1 per 100 PYO
information about unit register number, name, age, sex, (0.01 per annum). Recurrence was not associated with
address, TB category, smear result and the treatment out- age, sex, occupation, marital status and level of education
come. The data was crosschecked with the district TB reg- (Table 2).
ister that contained the list of the patients treated in the
health institutions in the districts. Trained health exten- Discussion
sion workers conducted house-to-house visits, and col- The estimated recurrence rate in our study area was 1 per
lected data from the TB patients or their households. They 100 PYO. This could be explained by HIV infection, MDR
collected the information if the TB patient were alive, had TB, reinfection due to high TB burden and inadequate
symptoms of TB and registered the date of the interview treatment supervision and patient follow-up.
using structured questionnaire. The data collection was
done from September 2007 to February 2008. HIV results HIV infection increases the risk of infection, reinfection,
were not available for TB patients enrolled in our study. recurrence and death. It also increased the workload by
fuelling TB epidemic and affected the performance of TB
Data analysis programme [6]. In southern Ethiopia, the prevalence of
We used SPSS 14 for Windows for data entry and analysis. HIV infection in the general population and TB patients
We described the patients by age, sex, TB category, marital was 3.8% and 17.5% respectively [15]. This could be one
status, level of education and occupation. The outcome of the factors to explain the high recurrence rate in our set-
measure was recurrence of TB. Person-years of observation ting. However, the role of HIV in recurrence requires fur-
(PYO) were calculated from the date of cure to date of ther investigation.
interview.
Higher recurrence rates reported elsewhere, 8.6% in Viet-
We used the Kaplan-Meier method to find out the event- nam after 19 months, 11% in India after two and half
free survival and the log-rank test for the statistical signif- years and 36% in Kazakhstan after 22 months of follow-
icance. Cox-regression method was used to determine the up [16-18] were attributed to MDR TB, poor treatment
hazard ratio (HR) and 95% Confidence interval (95%CI). supervision and inadequate patient follow-up
Recurrence rate was calculated as the number of recur- [8,9,11,12]. Though the prevalence of MDR TB in Ethio-
rences per 100 PYO. P-value less than 0.05 was considered pia was believed to be low (1.6% in new and 12% in pre-
significant. viously treated TB cases), 50% resistance to one or more
drugs in re-treatment cases was reported) [19]. Similarly
Ethical clearance 7.7% resistance to at least one TB drug was reported from
The Ethical Review Committee of the Regional Health our study area [20]. This may also be one of the factors to
Bureau approved this study. After explaining the aim of explain the high recurrence rate in our setting.
the study, we obtained informed consent from the study
participants or head of the household. Patients with Moreover, factors that affect the performance of TB pro-
recent history of cough and other symptoms suggestive of gramme (poor treatment supervision and failure to do fol-
Page 3 of 5
(page number not for citation purposes)
BMC Public Health 2009, 9:348 http://www.biomedcentral.com/1471-2458/9/348
Table 1: Socio-demographic characteristics of cured smear- tionally, failure to conduct follow-up sputum examina-
positive patients in southern Ethiopia from 1998 - 2006 tion reduces the chance of detecting failure cases (smear-
Variables number Percent (%) positive at 5th or 7th month) without affecting the number
of patients that complete treatment under DOTS. Thus, in
Sex routine practice where cure is based on smear microscopy,
Male 187 50.8% treatment failure can be missed.
Female 181 49.2%
Marital status The limitation of the study were using sputum microscopy
Single 90 24.5%
for the diagnosis of recurrence in smear-positive patients
Married 268 73.0%
Divorced/widowed 9 2.5%
that may have underestimated the rates of recurrence and
Missing 1 lack of HIV test result to estimate the role of HIV in recur-
Level of education rence.
Illiterate 157 44.2%
1-4 57 16.1% The significance of this study is more in settings with high
5-8 120 33.8% TB and HIV prevalence. In such settings high disease trans-
9+ 21 5.9%
mission may maintain the burden of TB in the commu-
Missing 13
Occupation
nity. Yet, high recurrence rates in cured smear-positive TB
Student 64 17.4% patients should alert TB programme managers to identify
Farmer 152 41.3% the risk factors. The performance of TB programme could
Housewife 35 9.5% be improved by addressing factors that affect treatment
Merchant 16 4.3% adherence and increase the risk of MDR TB. TB patients
Others 101 27.4% could also benefit from the access to HIV prevention and
Current status
control measures in high-risk patients to reduce recur-
New 364 98.9%
Retreatment 4 1.1% rence and improve their long-term survival.
Conclusion
low-up sputum examination) and the patients' general The rate of recurrence in cured smear-positive TB patients
condition could increase the recurrence rate. Inadequate was high in our setting. Further studies are required to
treatment supervision, more pronounced during continu- identify risk factors for recurrence to improve the disease
ation phase when patients receive unsupervised treat- free survival of cured smear-positive TB patients.
ment, reduces treatment adherence and increases the risk
of treatment failure and MDR TB. This is worsened when Competing interests
the importance of treatment adherence is not well The authors declare that they have no competing interests.
addressed during health education sessions [21]. Addi-
Table 2: Factors predicting recurrence in cured smear-positive tuberculosis patients in southern Ethiopia from 1998 - 2006
Variables Recurrence PYO* Recurrence rate per 100PYO Crude HR (95%CI)† P - value
Yes No
Page 4 of 5
(page number not for citation purposes)
BMC Public Health 2009, 9:348 http://www.biomedcentral.com/1471-2458/9/348
Authors' contributions 18. Vree M, Huong NT, Duong BD, Sy DN, Van LN, Hung NV, Co NV,
Borgdorff MW, Cobelens FG: Survival and relapse rate of tuber-
DGD supervised data collection. DGD and BL analyzed, culosis patients who successfully completed treatment in
interpreted the findings and prepared the drafts. All Vietnam. Int J Tuberc Lung Dis 2007, 11(4):392-397.
authors contributed to the final manuscript 19. Abate G, Miorner H, Ahmed O, Hoffner SE: Drug resistance in
Mycobacterium tuberculosis strains isolated from re-treat-
ment cases of pulmonary tuberculosis in Ethiopia: suscepti-
Acknowledgements bility to first-line and alternative drugs. Int J Tuberc Lung Dis
The authors acknowledge the contribution made by the TB programme 1998, 2(7):580-584.
20. Lemma E, Niemi M, Lindtjorn B, Dubrie G: Bacteriological studies
manager, health workers and health extension workers in the Dale and
of tuberculosis in Sidamo Regional Hospital. Ethiop Med J 1989,
Wonsho districts of Sidama zone. We are also grateful to TB patients who 27(3):147-149.
participated in the study. 21. Tekle B, Mariam DH, Ali A: Defaulting from DOTS and its deter-
minants in three districts of Arsi Zone in Ethiopia. Int J Tuberc
Lung Dis 2002, 6(7):573-579.
References
1. Raviglione MC: The TB epidemic from 1992 to 2002. Tuberculo-
sis (Edinb) 2003, 83(1-3):4-14. Pre-publication history
2. Shargie EB, Lindtjorn B: DOTS improves treatment outcomes The pre-publication history for this paper can be accessed
and service coverage for tuberculosis in South Ethiopia: a
retrospective trend analysis. BMC Public Health 2005, 5(1):62.
here:
3. Xianyi C, Fengzeng Z, Hongjin D, Liya W, Lixia W, Xin D, Chin DP:
The DOTS strategy in China: results and lessons after 10 http://www.biomedcentral.com/1471-2458/9/348/pre
years. Bull World Health Organ 2002, 80(6):430-436.
4. World Health Organization: Global Tuberculosis Control, surveillance, pub
planning and financing Geneva, Switzerland: World Health Organiza-
tion; 2008.
5. Dye C, Hosseini M, Watt C: Did we reach the 2005 targets for
tuberculosis control? Bull World Health Organ 2007,
85(5):364-369.
6. Maher D, Harries A, Getahun H: Tuberculosis and HIV interac-
tion in sub-Saharan Africa: impact on patients and pro-
grammes; implications for policies. Trop Med Int Health 2005,
10(8):734-742.
7. Harries AD, Dermot , Maher , Stephan , Graham : . In TB/HIV A Clin-
ical Manual Volume 329. 2nd edition. WHO, Switzerland, Geneva;
2004.
8. Cox H, Kebede Y, Allamuratova S, Ismailov G, Davletmuratova Z,
Byrnes G, Stone C, Niemann S, Rusch-Gerdes S, Blok L, et al.: Tuber-
culosis recurrence and mortality after successful treatment:
impact of drug resistance. PLoS Med 2006, 3(10):e384.
9. Kang'ombe C, Harries AD, Banda H, Nyangulu DS, Whitty CJ,
Salaniponi FM, Maher D, Nunn P: High mortality rates in tuber-
culosis patients in Zomba Hospital, Malawi, during 32
months of follow-up. Trans R Soc Trop Med Hyg 2000,
94(3):305-309.
10. CT Kang'ombe ADH, Ito K, Clark T, Nyirenda TE, Aldis W, Nunn PP,
Semba RD, Salaniponi FML: Long-term outcome in patients reg-
istered with tuberculosis in Zomba, Malawi: mortality at 7
years according to initial HIV status and type of TB. Int J
Tuberc Lung Dis 2004, 8(7):829-836.
11. Cox HS, Morrow M, Deutschmann PW: Long term efficacy of
DOTS regimens for tuberculosis: systematic review. Bmj
2008, 336(7642):484-487.
12. Lambert ML, Hasker E, Van Deun A, Roberfroid D, Boelaert M, Stuyft
P Van der: Recurrence in tuberculosis: relapse or reinfection?
Lancet Infect Dis 2003, 3(5):282-287.
13. Buccheri S, Reljic R, Caccamo N, Meraviglia S, Ivanyi J, Salerno A, Dieli
F: Prevention of the post-chemotherapy relapse of tubercu-
lous infection by combined immunotherapy. Tuberculosis
(Edinb) 2008, xxx:1-4.
14. Yassin MA, Datiko DG, Shargie EB: Ten-year experiences of the
Publish with Bio Med Central and every
tuberculosis control programme in the southern region of scientist can read your work free of charge
Ethiopia. Int J Tuberc Lung Dis 2006, 10(10):1166-1171. "BioMed Central will be the most significant development for
15. Datiko DG, Yassin MA, Chekol LT, Kabeto LE, Lindtjorn B: The rate
disseminating the results of biomedical researc h in our lifetime."
of TB-HIV co-infection depends on the prevalence of HIV
infection in a community. BMC Public Health 2008, 8:266. Sir Paul Nurse, Cancer Research UK
16. Vijay S, Balasangameswara VH, Jagannatha PS, Saroja VN, Kumar P:
Treatment outcome and two & half years follow-up status of Your research papers will be:
new smear positive patients treated under RNCTP. Indian J available free of charge to the entire biomedical community
Tuberc 2004, 51:199-208.
peer reviewed and published immediately upon acceptance
17. Thomas A, Gopi PG, Santha T, Chandrasekaran V, Subramani R, Sel-
vakumar N, Eusuff SI, Sadacharam K, Narayanan PR: Predictors of cited in PubMed and archived on PubMed Central
relapse among pulmonary tuberculosis patients treated in a yours — you keep the copyright
DOTS programme in South India. Int J Tuberc Lung Dis 2005,
9(5):556-561. Submit your manuscript here: BioMedcentral
http://www.biomedcentral.com/info/publishing_adv.asp
Page 5 of 5
(page number not for citation purposes)
VI
INT J TUBERC LUNG DIS 14(7):866–871
© 2010 The Union
D. G. Datiko,*† B. Lindtjørn*
* Centre for International Health, University of Bergen, Overlege Danielsens Hus, Bergen, Norway; † Southern Nations,
Nationalities and Peoples’ Regional Health Bureau, Hawassa, Ethiopia
SUMMARY
SETTING: The tuberculosis (TB) programme in the Si- R E S U LT S : A total of 725 TB patients were followed for
dama zone of southern Ethiopia. 2602 person-years: 91.1% (659/723) were alive and 8.9%
O B J E C T I V E : To measure excess mortality in successfully (64/723) had died. The mortality rate was 2.5% per an-
treated TB patients. num. Sex, age and occupation were associated with high
D E S I G N : In a retrospective cohort study of TB patients mortality. More deaths occurred in non-farmers (SMR =
treated from 1998 to 2006, mortality was used as an 9.95, 95%CI 7.17–12.73).
outcome measure, and was calculated per 100 person- D I S C U S S I O N : The mortality rate was higher in TB pa-
years of observation (PYO) from the date of completion tients than in the general population. More deaths oc-
of treatment to date of interview if the patient was alive, curred in non-farmers, men and the elderly. Further stud-
or to date of death. Kaplan-Meier and Cox regression ies are required to identify the causes of death in these
methods were used to determine the survival and hazard patients.
ratios. An indirect method of standardisation was used K E Y W O R D S : tuberculosis; mortality; standard mortal-
to calculate the standard mortality ratio (SMR). ity ratio; Ethiopia
THE DECENTRALISATION of the World Health least isoniazid [H] and rifampicin [R]).9–11 High re-
Organization (WHO) recommended DOTS strategy currence rates may also increase TB incidence, reduce
has increased the number of successfully treated tu- treatment success and increase post-treatment mor-
berculosis (TB) patients. TB treatment ensures long- tality or reduce post-treatment survival of success-
term success by reducing transmission, recurrence (re- fully treated TB patients.12,13
lapse or re-infection) and death.1–5 However, in some To our knowledge, no follow-up study has been
countries, mortality in TB patients has increased, conducted in Ethiopia to determine mortality in TB
mainly due to the human immunodeficiency virus patients after completion of treatment. The aim of the
(HIV) epidemic.6,7 TB mortality, as defined by the present study was to measure mortality in TB patients
WHO, includes the number of TB cases dying dur- after treatment completion under the DOTS strategy.
ing treatment, irrespective of cause, and is obtained
from routine reports. This excludes deaths that occur MATERIALS AND METHODS
among TB patients after treatment completion un-
der DOTS, defaulters and transfers out, and under- Study area and population
estimates mortality in TB patients.8 Mortality in suc- The present study was conducted in the Dale and
cessfully treated TB patients could be taken as an Wonsho Districts of the Sidama zone of southern
important measure of the efficacy of treatment. How- Ethiopia. This is a densely populated agrarian com-
ever, there is no routine monitoring of TB patients af- munity with a population of 296 811. The farmers cul-
ter treatment completion to understand what happens tivate cash crops (coffee and ‘khat’) that non-farmers
to them after successful treatment under the DOTS depend on for commercial activities. The DOTS strat-
strategy. egy was introduced in the study area in 1996.14 Six
Post-treatment studies conducted among TB pa- health facilities were providing DOTS by trained gen-
tients reported high mortality (24%) after 22 months eral health workers, using standard recording and re-
of follow-up.9 Mortality was highest among HIV- porting formats. District TB programme coordina-
infected (65%) and multidrug-resistant TB (MDR-TB) tors regularly checked the completeness and accuracy
patients (69%; i.e., cases with strains resistant to at of TB case recording in the unit TB register. The 2007
Correspondence to: Daniel G Datiko, P O Box 1722, Hawassa, Ethiopia. Tel: (+47) 5597 4980. Fax: (+47) 5597 4979.
e-mail: [email protected]
Article submitted 15 September 2009. Final version accepted 22 January 2010.
Mortality in treated TB patients in Ethiopia 867
estimate of mortality among TB patients on treatment Person-years of observation (PYO) were calculated
was 92 per 100 000 population per year.3 from the date of treatment completion to date of in-
terview if the patient was alive, or to date of death.
Study design The study outcomes of participants were censored
This retrospective cohort study was based on TB pa- if they were reported to be alive at the time of
tients who were registered in the unit TB registers in interview.
the health facilities providing DOTS. TB cases who Event-free survival and the log-rank test for statis-
completed treatment under the DOTS programme tical significance were determined using the Kaplan-
from 1998 to 2006 were enrolled in the study. Meier method. The Cox regression method was used
to determine the hazard ratio (HR) and 95% confi-
Treatment regimen, duration and outcome dence intervals (95%CI). P < 0.2 was used as a cut-
Treatment for smear-positive patients consisted of off point to include the variable in the multivariate
8 months of daily supervised streptomycin (S), R, Cox regression model. Mortality was calculated as
H and pyrazinamide (Z) for 2 months, followed by the number of deaths/100 PYO. Excess mortality was
6 months of self-administered ethambutol (E) and H calculated by subtracting age- and occupation-specific
for adults (2SRHZ/6EH), and 4 months of RH for mortality in the reference population from mortality
children (2SRHZ/4RH). Smear-negative and extra- among successfully treated TB patients. P < 0.05 was
pulmonary TB (EPTB) cases received 2 months of considered statistically significant.
RHZ followed by 6 months of EH (2RHZ/6EH). The To ascertain whether more than the expected num-
treatment regimen was the same throughout the study ber of deaths had occurred among our cohort, we
period. Follow-up sputum examination was performed used the indirect method of standardisation to calcu-
at the end of months 2, 5 and 7 of treatment. A smear- late the standard mortality ratio (SMR).15 As such
positive pulmonary TB (PTB) patient who had a neg- reference data for Sidama were not available, we used
ative sputum smear result in the last month of treat- the data from the Demographic and Health Survey of
ment and on at least one previous occasion (month 2 the Butajira Rural Health Programme, an open co-
or 5) was reported as cured. Smear-positive PTB cases hort,16 as the reference to calculate SMR. We believe
without month 7 smear results, and smear-negative that the two areas are comparable as they have simi-
PTB and EPTB cases who finished the full course of lar socio-economic development and are located at
treatment were declared ‘treatment completed’. TB the same altitude. In addition, about 50% of the pop-
cases declared as cured or treatment completed were ulation has access to health services and the DOTS
reported to be successfully treated under the DOTS strategy was implemented in the same year in the
strategy. The treatment success rate is the sum of health centres of the two areas. The SMR was calcu-
the cured and treatment completion rate. A patient lated as the ratio of the number of observed deaths
who died for any reason after treatment was recorded over expected deaths, using age- and occupation-
as ‘death’.9 specific mortalities in the reference population.
Table 2 Factors predictive of mortality in successfully treated tuberculosis patients in southern Ethiopia, 1998–2006
Death
Yes No Mortality Crude HR Adjusted HR
Variables n n PYO /100 PYO (95%CI) P value (95%CI) P value
Age, years 64 657 2586.1 2.5 1.1 (1.0–1.1) <0.01 1.0 (1.0–1.1) <0.01
Sex
Female 24 321 1270.8 1.9 1.0
Male 40 337 1320.5 3.0 1.6 (0.9–2.7) 0.1 2.2 (1.3–3.9) 0.01
Level of education
Illiterate 15 284 1118.5 1.3 1.0
Literate 11 350 1289.9 0.9 0.6 (0.3–1.4) 0.3
Marital status
Never married 5 179 632.3 0.8 1.0
Married 23 476 1861.3 1.2 1.6 (0.6–4.1) 0.4
Occupation
Farmers 19 479 1838.4 1.0 1.0
Non farmers 45 180 757.2 5.9 5.7 (3.3–9.7) <0.01 6.3 (3.6–11.1) <0.01
TB classification
Smear-positive PTB 33 394 1504.8 2.2 1.0
Smear-negative PTB 22 142 606.9 3.6 1.7 (0.9–2.9) 0.1 1.1 (0.6–1.9) 0.8
EPTB 9 122 481.1 1.9 0.9 (0.4–1.8) 0.7 1.1 (0.5–2.2) 0.9
PYO = person-year of observation; HR = hazard ratio; CI = confidence interval; TB = tuberculosis; PTB = pulmonary TB; EPTB = extra-pulmonary TB.
Mortality in treated TB patients in Ethiopia 869
merchants, former soldiers and government and pri- In our study, overall mortality was 2.5/100 PYO
vate sector employees. In Cox regression analysis, age (2.5% per annum), which is lower than in other re-
(P < 0.01), sex (P < 0.01) and occupation (P < 0.01) ports from Africa (ranging from 3.1/100 PYO in the
were associated with increased mortality (Figure 2, Democratic Republic of Congo [DRC] to 23.7/100
Table 2). PYO in Malawi). This could be explained by the
Overall excess mortality was 13.6/100 PYO, 2.5/ lower TB-HIV co-infection rate in our setting (17.5%)
100 PYO for farmers and 11.5/100 PYO for non- compared to other reports from Africa (ranging from
farmers. Our results suggest that more deaths occurred 21% in the DRC to 77% in Malawi).11 The mortality
than expected after treatment of TB patients. The SMR in our setting was also lower than reports from India
for all patients was 4.5 (95%CI 3.42–5.57). The (5.7/100 PYO in rural Velliyur and 6.8/100 PYO in
SMR was higher in non-farmers (9.95, 95%CI 7.17– Chennai City) and Vietnam (6.6/100 PYO).19,20 The
12.73) than in farmers (1.97, 95%CI 1.10–2.84). lower prevalence of MDR-TB among new and previ-
ously treated cases could explain this difference (re-
spectively 1.6% and 12% in Ethiopia, 2.7% and
DISCUSSION
17% in India and 2.7% and 19% in Vietnam).17 Al-
One of the targets of the Millennium Development though MDR-TB seems to be the least likely explana-
Goals is to reduce TB mortality.17 However, reliable tion in our setting, it should be interpreted with cau-
data on TB mortality are not available.8,18 This could tion in the presence of a reported resistance rate of
be improved by measuring mortality in successfully 7.6% to at least one anti-tuberculosis drug.21
treated cases. We report higher mortality in success- Some population groups are at higher risk of death
fully treated TB cases than in the general population. due to occupational exposure or lifestyle. In our
study, the risk of excess mortality was six times higher
in non-farmers (including merchants, former soldiers
and government or private sector employees) than
farmers. This could be explained by the high mobility
among these groups, and the high prevalence and in-
creased risk of HIV infection, as these groups were
most affected by HIV and were sources of HIV infec-
tion and transmission from urban to rural areas.22
HIV infection in successfully treated non-farmer TB
patients could thus be one of the plausible explana-
tions for the high mortality in our setting. In addi-
tion, the risk of excess mortality was twice as high in
males as in females, with the risk increasing with in-
creasing age. This could be because more male TB pa-
Figure 2 Kaplan-Meier survival curve showing post-treatment tients were reported,17,23 which could also result in
mortality in tuberculosis patients by occupation in southern more deaths. The higher risk of HIV infection in males
Ethiopia. than in females could explain the higher mortality in
870 The International Journal of Tuberculosis and Lung Disease
rural communities.22 Moreover, the traditional risk- tered or undiagnosed by the DOTS programme are
taking role of males, given their role in society, puts beyond the scope of this study.
them at higher risk for excess mortality.24 As ex- In conclusion, post-treatment mortality was higher
pected, mortality was higher in the elderly, possibly in TB patients than in the general population. There is
due to age-related diminution of immunity, and the a need for selective post-treatment follow-up of high-
increased magnitude of chronic illnesses such as dia- risk groups that could be identified at the start of
betes and other comorbidities with age.25–27 Case- treatment. Post-treatment mortality could be used as
fatality rates among TB patients during treatment at additional evidence of case fatality (obtained through
6 months and at 12 and 20 months after treatment routine reports) to better understand mortality in TB
were reported to be increasing, and more deaths oc- patients. Further studies are needed to ascertain the
curred in the elderly and in males.8 The relative lower causes of death in TB patients.
mortality in our setting could be because of the lower
Acknowledgements
prevalence of HIV infection, and the lower burden of
TB and other comorbidities. The authors thank the TB programme manager, health workers
and health extension workers in the study area. They are grateful
Mortality rates in our setting were similar among to Professor Y Berhane for supplying the mortality data for the ref-
smear-positive, smear-negative and EPTB cases, pos- erence population.
sibly due to similar rates of HIV infection in TB pa-
tients.19 This is in contrast to many studies that have References
reported higher mortality in smear-negative and EPTB 1 Shargie E B, Lindtjorn B. DOTS improves treatment outcomes
cases. In such settings, the mortality was higher in TB and service coverage for tuberculosis in South Ethiopia: a retro-
patients, mainly due to HIV infection.11 spective trend analysis. BMC Public Health 2005; 5: 62.
Generally, mortality was higher in TB patients af- 2 Xianyi C, Fengzeng Z, Hongjin D, et al. The DOTS strategy in
ter successful treatment. We report excess mortality China: results and lessons after 10 years. Bull World Health
Organ 2002; 80: 430–436.
in our study participants over the general population
3 World Health Organization. WHO report 2008. Global tuber-
and an SMR of 4.5, similar to the SMR of 4.2 re- culosis control: surveillance, planning, financing. WHO/HTM/
ported in rural settings of India.26 In our setting, mor- TB/2008.393. Geneva, Switzerland: WHO, 2008.
tality was higher in non-farmers than in farmers for 4 Dye C, Hosseini M, Watt C. Did we reach the 2005 targets for
similar reasons: HIV infection and related opportun- tuberculosis control? Bull World Health Organ 2007; 85: 364–
369.
istic infections, chronic illnesses such as diabetes and
5 Raviglione M C. The TB epidemic from 1992 to 2002. Tuber-
other comorbidities such as malaria, malignancies culosis (Edinb) 2003; 83: 4–14.
and others. Mortality was lower in our study partici- 6 Maher D, Harries A, Getahun H. Tuberculosis and HIV inter-
pants than the reported SMR of 6.1 in urban settings action in sub-Saharan Africa: impact on patients and pro-
in India. grammes; implications for policies. Trop Med Int Health 2005;
10: 734–742.
Gaps in the performance of the TB programme,
7 World Health Organization. TB/HIV, a clinical manual. 2nd ed.
such as inadequate patient follow-up, and the patients’ WHO/HTM/TB/2004.329. Geneva, Switzerland: WHO, 2004.
general condition could lead to increases in mortality 8 Kolappan C, Subramani R, Karunakaran K, Narayanan P R.
during treatment and put them at higher risk of ex- Mortality of tuberculosis patients in Chennai, India. Bull World
cess mortality after completion of treatment. Inade- Health Organ 2006; 84: 555–560.
quate treatment supervision, mainly during the con- 9 Cox H, Kebede Y, Allamuratova S, et al. Tuberculosis recur-
rence and mortality after successful treatment: impact of drug
tinuation phase, when treatment is taken unsupervised, resistance. PLoS Med 2006; 3: e384.
could also reduce treatment adherence and increase 10 Kang’ombe C, Harries A D, Banda H, et al. High mortality
the recurrence rate of TB (recurrence rates as high as rates in tuberculosis patients in Zomba Hospital, Malawi, dur-
1/100 PYO have been reported in the study area28), ing 32 months of follow-up. Trans R Soc Trop Med Hyg 2000;
which could contribute to increased mortality. This 94: 305–309.
11 Kang’ombe C T, Harries A D, Ito K, et al. Long-term outcome
could be worse where the importance of treatment in patients registered with tuberculosis in Zomba, Malawi:
adherence is not well addressed during health educa- mortality at 7 years according to initial HIV status and type of
tion sessions.29 TB. Int J Tuberc Lung Dis 2004; 8: 829–836.
We report mortality in successfully treated TB cases 12 Cox H S, Morrow M, Deutschmann P W. Long term efficacy of
through home-based follow-up under programme con- DOTS regimens for tuberculosis: systematic review. BMJ 2008;
336: 484–487.
ditions in Ethiopia. No baseline difference was found
13 Lambert M L, Hasker E, Van Deun A, Roberfroid D, Boelaert
between the study participants and those who were M, Van der Stuyft P. Recurrence in tuberculosis: relapse or re-
not included in the study, which increases the repre- infection? Lancet Infect Dis 2003; 3: 282–287.
sentativeness of our study and its generalisability to 14 Yassin M A, Datiko D G, Shargie E B. Ten-year experiences of
the study area. However, using the general popula- the tuberculosis control programme in the southern region of
Ethiopia. Int J Tuberc Lung Dis 2006; 10: 1166–1171.
tion may have underestimated the mortality in TB
15 Brand M. An introduction to medical statistics. 3rd ed. Ox-
patients, while failure to ascertain the causes of death ford, UK: Oxford University Press, 2000.
could also have included other causes of death, thus 16 Berhane Y, Wall S, Fantahun M, et al. A rural Ethiopian popu-
overestimating mortality. TB cases who die unregis- lation undergoing epidemiological transition over a generation:
Mortality in treated TB patients in Ethiopia 871
Butajira from 1987 to 2004. Scand J Public Health 2008; 36: cation data to explore sex differences in case detection. Int J
436–441. Tuberc Lung Dis 2000; 4: 123–132.
17 World Health Organization. WHO report 2009. Global tuber- 24 Grieg A, Kimmel M, Lang J. Men, masculinities and develop-
culosis control: epidemiology, strategy, financing. WHO/HTM/ ment: broadening our work towards gender equality. Gender
TB/2009.411. Geneva, Switzerland: WHO, 2009. in Development Monograph Series #10. Geneva, Switzerland:
18 World Health Organization. Tuberculosis programme. Frame- United Nations Development Programme, 2004.
work for effective tuberculosis control. WHO/TB/94.179. Ge- 25 Raviglione M C, ed. Reichman and Hershfield’s tuberculosis: a
neva, Switzerland: WHO, 1994. comprehensive international approach. Lung biology in health
19 Datiko D G, Yassin M A, Chekol L T, Kabeto L E, Lindtjorn B. and disease, Vol 219. 3rd ed. New York, NY, USA: Informa
The rate of TB-HIV co-infection depends on the prevalence of Healthcare, 2006.
HIV infection in a community. BMC Public Health 2008; 8: 26 Kolappan C, Subramani R, Kumaraswami V, Santha T, Naraya-
266. nan P R. Excess mortality and risk factors for mortality among
20 Korenromp E L, Bierrenbach A L, Williams B G, Dye C. The a cohort of TB patients from rural South India. Int J Tuberc
measurement and estimation of tuberculosis mortality. Int J Lung Dis 2008; 12: 81–86.
Tuberc Lung Dis 2009; 13: 283–303. 27 Rajagopalan S. Tuberculosis and aging: a global health prob-
21 Lemma E, Niemi M, Lindtjorn B, Dubrie G. Bacteriological lem. Clin Infect Dis 2001; 33: 1034–1039.
studies of tuberculosis in Sidamo Regional Hospital. Ethiop 28 Datiko D G, Lindtjørn B. Tuberculosis recurrence in smear-
Med J 1989; 27: 147–149. positive patients cured under DOTS in southern Ethiopia: retro-
22 Shabbir I, Larson C P. Urban to rural routes of HIV infection spective cohort study. BMC Public Health 2009; 9: 348.
spread in Ethiopia. J Trop Med Hyg 1995; 98: 338–342. 29 Tekle B, Mariam D H, Ali A. Defaulting from DOTS and its
23 Borgdorff M W, Nagelkerke N J, Dye C, Nunn P. Gender and determinants in three districts of Arsi Zone in Ethiopia. Int J
tuberculosis: a comparison of prevalence surveys with notifi- Tuberc Lung Dis 2002; 6: 573–579.
RÉSUMÉ
CONTEXTE : Le Programme de tuberculose (TB) dans la R É S U LTAT S : Un total de 725 patients TB ont été suivis
zone Sidama du sud de l’Ethiopie. pendant 2602 années-personne. Etaient en vie 91,1%
O B J E C T I F : Mesurer l’excès de mortalité chez les pa- (659/723) alors que 8,9% étaient décédés (64/723). La
tients TB traités avec succès. mortalité a été de 2,5% par an. Sont en association avec
S C H É M A : Dans une étude rétrospective de cohorte sur une mortalité élevée le sexe, l’âge et la profession. Le
les patients TB traités entre 1998 et 2006, la mortalité a nombre de décès est plus élevé chez les non-fermiers
constitué la mesure de résultats, calculé par 100 années- (SMR = 9,95 ; IC95% 7,17–12,73).
personne d’observation (PYO) à partir de la date d’achève- D I S C U S S I O N : La mortalité est élevée chez les patients
ment du traitement jusqu’à la date de l’entretien, si le pa- TB par comparaison avec la population générale. Les
tient était en vie, ou jusqu’à la date du décès. On a utilisé décès sont plus fréquents chez les non-fermiers, chez les
les méthodes de Kaplan-Meier et de régression de Cox hommes et chez les sujets âgés. Des études complémen-
pour déterminer la survie et le ratio de risque. On a uti- taires s’imposent pour identifier les causes de décès chez
lisé la méthode indirecte de standardisation pour le cal- ces patients.
cul d’un ratio standard de mortalité (SMR).
RESUMEN
MARCO DE REFERENCIA: El programa contra la tuber- R E S U LTA D O S : Se practicó el seguimiento de 725 pa-
culosis (TB) en la región de Sidama, al sur de Etiopía. cientes tuberculosos por 2602 PYO. El 91,1% de los pa-
O B J E T I V O : Medir el exceso de mortalidad en los pacien- cientes (659/723) estaba vivo y el 8,9% (64/723) había
tes que recibieron un tratamiento exitoso contra la TB. fallecido. La mortalidad anual fue 2,5%. Los factores
M É T O D O : Se llevó a cabo un estudio retrospectivo de asociados con una alta mortalidad fueron el sexo, la
cohortes de los pacientes tuberculosos tratados entre edad y la ocupación. Ocurrieron más defunciones en las
1998 y 2006. El criterio de valoración fue la mortalidad personas que no eran agricultores (SMR = 9,95; IC95%
que se calculó en años-persona de observación (PYO) 7,17–12,73).
por 100, a partir de los datos de compleción del trata- C O N C L U S I Ó N : Se observó una alta mortalidad en los
miento hasta la fecha de la entrevista cuando el paciente pacientes tuberculosos, comparados con la población
estaba vivo o hasta la fecha de la defunción. Con el fin general. La mortalidad fue mayor en los pacientes que
de determinar la supervivencia y el cociente de riesgos no eran agricultores, en los hombres y en los ancianos.
instantáneos se aplicaron el método de Kaplan-Meier y Se precisan nuevos estudios con el fin de determinar las
el modelo de regresión de Cox. La razón estandarizada causas de defunción en estos pacientes.
de la mortalidad (SMR) se calculó usando un método
indirecto de normalización.
(WKLFDO
&ODUHQFH
Sidama Zone Health Department
Awassa
We are pleased to inform you that the request on the subject of ethical clearance on the
research entitled "Improving Community based TB care in SNNPR" by Dr. Daniel
Gemechu that is planned to be undertaken in Dale woredas is approved and accepted.
Worth mentioning, however, the Center for Health Research & Laboratories is
requested to strictly monitor and evaluate the ethical implementation of the project as
;(ours
Kine
~6"fM
Chawicha Debessa
fm.c; ,.e"1t-'1"":f·t;~h11A"'H* IIC
9n/~"I/.,
Deputy Head, Health Programs and
Services Division