de Cuevas et al.

Infectious Diseases of Poverty (2016) 5:24

DOI 10.1186/s40249-016-0117-x

RESEARCH ARTICLE Open Access

Patients direct costs to undergo TB

diagnosis
Rachel M. Anderson de Cuevas1, Lovett Lawson2,3, Najla Al-Sonboli4, Nasher Al-Aghbari5, Isabel Arbide6,
Jeevan B. Sherchand7, Emenyonu E. Nnamdi2, Abraham Aseffa8, Mohammed A. Yassin1,9,
Saddiq T. Abdurrahman10, Joshua Obasanya11, Oladimeji Olanrewaju2, Daniel Datiko12, Sally J. Theobald1,
Andrew Ramsay13, S. Bertel Squire1 and Luis E. Cuevas1*

Abstract
Background: A major impediment to the treatment of TB is a diagnostic process that requires multiple visits.
Descriptions of patient costs associated with diagnosis use different protocols and are not comparable.
Methods: We aimed to describe the direct costs incurred by adults attending TB diagnostic centres in four countries
and factors associated with expenditure for diagnosis. Surveys of 2225 adults attending smear-microscopy centres in
Nigeria, Nepal, Ethiopia and Yemen. Adults >18 years with cough >2 weeks were enrolled prospectively. Direct costs
were quantified using structured questionnaires. Patients with costs >75th quartile were considered to have high
expenditure (cases) and compared with patients with costs <75th quartile to identify factors associated with
high expenditure.
Results: The most significant expenses were due to clinic fees and transport. Most participants attended the centres
with companions. High expenditure was associated with attending with company, residing in rural areas/other towns
and illiteracy.
Conclusions: The costs incurred by patients are substantial and share common patterns across countries. Removing
user fees, transparent charging policies and reimbursing clinic expenses would reduce the poverty-inducing effects of
direct diagnostic costs. In locations with limited resources, support could be prioritised for those most at risk of high
expenditure; those who are illiterate, attend the service with company and rural residents.
Keywords: Tuberculosis, Costs, Access to healthcare, Ethiopia, Nepal, Nigeria, Yemen

Multilingual abstract [1], that no TB affected household should experience

Please see Additional file 1 for translations of the catastrophic costs due to TB [2].
abstract into the six official working languages of the Access to services is also a major problem. Of the
United Nations. 9.5 million incident cases estimated by the WHO in
2013, only 6 million were reported and the rest were
either not diagnosed, or diagnosed but not reported to
Background national TB programmes (NTPs) [3]. Over 77 million
TB is a disease of poverty mostly affecting populations smear investigations are conducted every year in the top
with limited resources and restricted access to health twenty high burden countries, most of these in poor
services in low and middle income countries (LMICs). people [4]. Although drugs for first-line treatment for TB
For this reason the WHO has included a specific finan- are provided for free by most NTPs [5], the financial costs
cial risk protection target in the new End TB Strategy incurred by the patient and their family for diagnosis re-
main significant [6, 7] and further costs may include hos-
* Correspondence: [email protected] pitalisation, treatment of side effects, nutrition and others.
1
Liverpool School of Tropical Medicine, Pembroke Place, Liverpool L3 5QA, The onset of disease often exacerbates poverty, as the
UK
patient enters a lengthy period of expenditure on
Full list of author information is available at the end of the article

© 2016 de Cuevas et al. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
de Cuevas et al. Infectious Diseases of Poverty (2016) 5:24 Page 2 of 9

healthcare and productivity losses [8]. Direct costs start Tribhuvan University Teaching Hospital, a governmental
with the patient undertaking consultations to reach a referral hospital in Kathmandu; in Abuja Nigeria, pa-
diagnosis; yet these costs had been overlooked by policy- tients were enrolled in Wuse District Hospital, a govern-
makers until recently [1, 9]. Although TB diagnostics ment general hospital and in Yemen, patients were
tests are free of charge in most countries, public health selected from the National Tuberculosis Institute (NTI),
services often charge patients for their initial clinical a government referral centre and the main centre for the
consultation, non-specific medications and other diagnos- diagnosis of TB in Sana’a. These centres were selected as
tic tests, such as x-rays [10, 11]. Patients must also meet they were integral to the NTP services and large num-
the cost of displacement from home, including domestic bers of patients received a diagnosis of TB. The approxi-
responsibilities, subsistence and loss of earnings. Costs are mate number of patients routinely screened for TB in
augmented by the need for patients to spend several days each centre was 260, 1000, 250 and 780 per month in
near the health facility to complete the tests and meet Ethiopia, Nepal, Nigeria and Yemen, respectively. Individ-
health staff for clinical management decisions [12, 13]. uals attending the centres self-referred or were referred by
The cost of diagnosis is not uniform for all patients government, private or informal health care providers. A
and some have higher expenses than others. Identifying minimum of 500 patients were expected to be recruited
individuals at risk of unusually high expenditure would from each study site to obtain a representative sample of
be of value to health programmes and policy makers to patients attending the centre and to have at least 80 %
inform the development of interventions to support pa- power to identify risk factors with an Odds Ratio > 1.5.
tients with limited resources and high expenditure. Few Participants were interviewed using standard question-
studies have isolated the patients’ direct and indirect naires to obtain demographic and clinical information,
costs of attending NTP health facilities for diagnosis and to establish whether they had travelled with company
even fewer have compared expense patterns across and the direct costs associated with attending the centre.
countries. The questionnaire was structured and administered face
This study examined the direct financial costs incurred to face. Variables were defined using quantitative scales.
by adults attending smear-microscopy based TB diag- The questionnaire had been developed over several studies
nostic centres in Ethiopia, Nepal, Nigeria and Yemen preceding the study and developed jointly by all co-authors
and the risk factors for high expenditure for diagnosis. to allow for local sensitivities and pooling of expertise.
This information was used to identify the population Questionnaires were administered by local research staff
groups that are more likely to experience higher costs to specifically trained and employed for the study, using the
undergo the TB diagnostic process. local languages and took about 30 min per patient. Partici-
pants who were illiterate were supported by local staff to
Methods clarify the questions and to estimate costs. Very few pa-
Ethical approvals for the study were obtained from the tients refused to participate, and >95 % of patients con-
research ethics committees of the Liverpool School of sented on first approach. Patients who refused often did
Tropical Medicine, the Institutional Review Boards of because they were in a hurry or (in some settings) they pre-
the World Health Organisation and all the participating ferred to be interviewed with a partner who was not
institutions in Ethiopia, Nepal, Nigeria, and Yemen. All present at the time of the interview. A pilot study was run
participants gave written informed consent. in all countries before the main surveys and questions were
The study comprised four cross-sectional surveys adjusted as needed. Direct costs were defined as medical
using the same study protocol among patients undergo- (clinic fees or registration costs, cost of investigation or
ing routine smear microscopy in Ethiopia, Nepal, Nigeria consultation, diagnostic tests and medication) and the costs
and Yemen [14]. These countries were selected because of transportation, food and accommodation for patient and
the team was conducting a multi-country study to companion/s). Direct costs also included travel, overnight
optimise the use of smear microscopy and represented stays, expenditure needed to attend the second day of diag-
high burden countries across three World Health nosis and others related to the current visit. Indirect costs,
Organization regions [14]. Adults over 18 years with defined as the loss of income for patient and household
chronic cough of more than 2 weeks duration attending due to lost work days, incapacity to work or forced chan-
selected health facilities were invited to participate. ged of occupation and substitution costs (the cost of re-
Participants were selected using systematic random placing the patient in their duties) and costs for previous
sampling, with a maximum of 10 patients interviewed health service encounters were not measured.
each day. In Ethiopia, participants were enrolled in Data were entered in a database using Epi-Info. The
Bushullo Major Health Centre, a not-for-profit mission participants’ characteristics by country were described
hospital located in the outskirts of Hawassa; capital of using summary statistics. Proportions were compared
the Southern Region; in Nepal, patients were enrolled at using chi square tests and means were compared using
de Cuevas et al. Infectious Diseases of Poverty (2016) 5:24 Page 3 of 9

parametric and non-parametric tests for normal and separated (2 %). Between 10 % (Nepal) and 47 % (Yemen)
skewed data, respectively. Costs were described using of patients resided in rural areas and the majority of urban
quartiles and expressed as the median and 25–75th inter- patients resided in the town where the study health centre
quartile range (25–75 % IQR). Costs were calculated by was located; except in Nepal, where almost half of the pa-
adding the figures in the local currencies for each day of tients from urban areas came from other cities. The mean
attendance and converting the day’s total to US dollars (SD) number of residents per household was 6.1 (4), ran-
using the exchange rate at the beginning of the study. ging from 4.1 (3) in Nigeria to 8.6 (5) in Yemen.
Median expenditures were described stratified by se- The majority of patients were accompanied by another
lected patient characteristics (e.g. rural/urban residency) person in all countries except Nigeria, as shown in
and compared using non parametric tests. Patients with Table 2. The accompanying person was a relative other
costs >75th quartile were considered to have high ex- than the spouse in 77 % of patients, the spouse in 15 %
penditure (cases) and were compared with patients with and 8 % were accompanied by other people. Few pa-
costs <75th quartile, to identify factors associated with tients walked to the clinics and the most common trans-
high expenditure. Odds Ratios (OR) with 95 % confi- port methods were buses in Nigeria, Nepal and Ethiopia
dence intervals (95%CI) were calculated and variables and cars or taxis in Yemen. Horse cart and motorbike
with p values <0.2 were entered into backward and travel were also common in Ethiopia (26 %) and Nepal
forward logistic regressions to identify factors inde- (17 %), respectively. The median (25–75 % IQR) travel
pendently associated with high expenditure (Adjusted time to the health centre was 40 (20–66) minutes. Travel
OR, AOR). time was shortest in Nepal (25; 15–30 min) and longest
in Yemen (60, 30–150 min). The majority of patients
Results and discussion intended to spend the night at home the first day of con-
A total of 2225 patients were enrolled, of whom 504 sultation (66 %), except in Nepal where 46 % planned to
were enrolled in Ethiopia, 619 in Nepal, 502 in Nigeria stay with a relative. Very few patients paid for hotel ac-
and 600 in Yemen. Patients had a mean (SD) age of 39 commodation. Transport patterns for the second day of
(17) years and were more likely to be male than female consultation where similar, except in Yemen where pa-
in all countries except Nigeria (Table 1). The majority of tients realised that there was a low cost taxi service and
patients (67 %) were married or had partners; 27 % were a lower number decided to travel by private car the
single and few were widowed (5 %) or divorced/ second day (data not shown).

Table 1 Patients’ demographic characteristics by country

Variablesa Ethiopia N = 504 Nepal N = 619 Nigeria N = 502 Yemen N = 600 All N = 2225
Age (Mean, ±SD) 33.2 (14.5) 43.8 (17.8) 34.4 (11) 41.8 (18.3) 38.7 (16.6)
Gender male: female (% male) 279: 225 (55.4) 395: 224 (63.8) 241: 258 (48.0) 329: 271 (54.8) 1244: 978 (55.9)
Marital status Single 176 (34.9) 107 (17.3) 183 (36.5) 124 (20.7) 590 (26.5)
With partner/married 302 (59.9) 507 (81.9) 259 (51.6) 426 (71) 1494 (67.1)
Divorced/separated 12 (2.4) 0 (0) 13 (2.6) 10 (1.7) 35 (1.6)
Widowed 14 (2.8) 5 (0.8) 44 (8.8) 40 (6.7) 103 (4.6)
Residence Rural 182 (36.1) 63 (10.2) 82 (16.3) 279 (46.5) 606 (27.2)
Same town 216 (42.9) 284 (45.9) 410 (81.7) 268 (44.7) 1178 (52.9)
Other town 106 (21) 272 (43.9) 7 (1.4) 53 (8.8) 438 (19.7)
Literate: Illiterate (% literate) 247: 257 (49) 475: 137 (76.7) 439: 57 (87.5) 264: 336 (44.0) 1425: 787 (64)
Education Nil 257 (51) 197 (31.8) 41 (8.2) 311 (51.8) 806 (36.2)
Primary incomplete 117 (23.2) 37 (6) 42 (8.4) 126 (21) 322 (14.5)
Primary complete 43 (8.5) 105 (17) 63 (12.5) 56 (9.3) 267 (12)
Secondary 62 (12.3) 127 (20.5) 220 (43.8) 61 (10.2) 470 (21.1)
Tertiary 24 (4.8) 153 (24.7) 133 (26.5) 46 (7.7) 356 (16)
Missing 1 (0.2) 0 (0) 3 (0.6) 0 (0) 4 (0.2)
Working: not working (% working) 235: 269 (46.6) 202: 417 (32.6) 368: 131 (73.3) 164: 436 (27.3) 969: 1253 (43.6)
Mean (SD) residents in household 5.7 (2.9) 5.5 (3) 4.1 (3) 8.6 (5) 6.1 (4)
a
Data represent frequency (%), unless otherwise specified. Sex, residency, marital and work status were missing for 3 patients and literacy for 13 patients
de Cuevas et al. Infectious Diseases of Poverty (2016) 5:24 Page 4 of 9

Table 2 Characteristics of patients attending the clinics

Variablesa Ethiopia N = 504 Nepal N = 619 Nigeria N = 502 Yemen N = 600 All N = 2225
Alone: With company (% accompanied) 129:373 (74) 181:435 (70.3) 320:177 (35.3) 91:509 (84.8) 721:1494 (67.1)
Accompanying person Spouse 61 (16.4) 62 (14.3) 50 (28.2) 51 (10) 224 (15)
Other relative 276 (74) 336 (77.2) 94 (53.1) 439 (86.2) 1145 (76.6)
Friend 13 (3.5) 17 (3.9) 19 (10.7) 17 (3.3) 66 (4.4)
Neighbour 7 (1.9) 10 (2.3) 1 (0.6) 0 (0) 18 (1.2)
Other 16 (4.3) 9 (2.1) 13 (7.3) 2 (0.4) 40 (2.7)
Travel time, median (25–75 IQR), min 45 (27–120) 25 (15–30) 45 (30–60) 60 (30–150) 40 (20–66)
Transport day 1 Walking 67 (13.3) 92 (14.9) 12 (2.4) 12 (2) 183 (8.2)
Carried 6 (1.2) 0 (0) 2 (0.4) 0 (0) 8 (0.4)
Bicycle 14 (2.8) 7 (1.1) 0 (0) 0 (0) 21 (0.9)
Cart/horse 129 (25.6) 5 (0.8) 0 (0) 4 (0.7) 138 (6.2)
Motorbike 1 (0.2) 104 (16.8) 2 (0.4) 3 (0.5) 110 (4.9)
Bus 148 (29.4) 303 (48.9) 303 (60.4) 127 (21.2) 881 (39.6)
Car 33 (6.5) 14 (2.3) 76 (15.1) 203 (33.8) 326 (14.7)
Taxi 103 (20.4) 83 (13.4) 101 (20.1) 250 (41.7) 537 (24.1)
Other 1 (0.2) 8 (1.3) 1 (0.2) 1 (0.2) 11 (0.5)
Transport day 2 Walking 118 (23.4) 104 (16.8) 13 (2.6) 12 (2) 247 (11.1)
Carried 3 (0.6) 2 (0.3) 2 (0.4) 1 (0.2) 8 (0.4)
Bicycle 13 (2.6) 9 (1.5) 0 (0) 0 (0) 22 (1)
Cart/horse 181 (35.9) 5 (0.8) 0 (0) 5 (0.8) 191 (8.6)
Motorbike 1 (0.2) 104 (16.8) 1 (0.2) 3 (0.5) 109 (4.9)
Bus 44 (8.7) 298 (48.1) 309 (61.6) 143 (23.8) 794 (35.7)
Taxi 108 (21.4) 72 (11.6) 73 (14.5) 327 (54.5) 580 (26.1)
Other 5 (1) 9 (1.5) 38 (7.6) 0 (0) 52 (2.3)
Accommodation Home 358 (71) 267 (43.1) 437 (87.1) 407 (67.8) 1469 (66)
Relative 60 (11.9) 282 (45.6) 5 (1) 123 (20.5) 470 (21.1)
Hotel 66 (13.1) 40 (6.5) 0 (0) 68 (11.3) 174 (7.8)
Street 2 (0.4) 0 (0) 1 (0.2) 0 (0) 3 (0.1)
Shift 1 (0.2) 6 (1.0) 0 (0) 0 (0) 7 (0.3)
Hospital 15 (3.0) 20 (3.2) 54 (10.8) 2 (0.3) 91 (4.1)
a
Data are frequency (%), unless specified. Patients’ data missing for whether they attended with company (i10,) person accompanying (1), accommodation (11 and
transport (18). Min = minutes. P values for all comparisons between countries < 0.05, except for accompanying person being the spouse (p = 0.26) and Being
carried (p = 0.69)

Costs of attending the clinic all countries and the median (25 % & 75 % IQR) expend-
Costs associated with the two-day clinic attendance are iture was 0 in all countries.
summarised in Table 3. Second day expenses were simi- The costs of attending the clinic stratified by the pa-
lar to first day expenses. The median costs were higher tients’ characteristics are described in Table 4. Older pa-
in Yemen ($11.89) and Nepal ($8.22) and lower in tients, those attending the clinic with company, residing
Nigeria ($5) and Ethiopia ($1.47). The most significant in rural areas or coming from other towns and patients
expenses were due to the clinic costs (except Nigeria) not working, had higher median costs than younger pa-
and transport. Miscellaneous expenses were frequently tients, those attending alone, residing in the same town
reported in Nepal (83 % of patients) and were rare in or working. Farmers, students and housewives had higher
Yemen (12 %), Nigeria (3 %) and Ethiopia (0 %). Al- median costs than patients with other occupations. Pa-
though rare in Yemen, these expenses were significant. tients with a confirmed TB diagnosis (smear and culture
Very few patients paid for overnight accommodation in positive) had higher costs than patients without laboratory
de Cuevas et al. Infectious Diseases of Poverty (2016) 5:24 Page 5 of 9

Table 3 Median and IQR costs incurred for attending the clinic
Variablea Ethiopia ETB Nepal NPR Nigeria NGN Yemen YER
N = 504 N = 619 N = 502 N = 600
Proportion reporting miscellaneous expenses, N (%)b 0 (0) 509 (82.6 %) 13 (2.6 %) 73 (12.2 %)
Clinic costs 8 (5; 8) 300 (300; 300) 0 (0; 150) 450 (250; 450)
Transport 16 (4; 36) 48 (0; 100) 400 (200; 600) 2000 (800; 3400)
Food 0 (0; 12) 0 (0; 0) 200 (0; 400) 0 (0; 400)
Overnight accommodation 0 (0; 0) 0 (0; 0) 0 (0; 0) 0 (0; 0)
Miscellaneous expenditure 0 (0; 0) 300 (200; 300) 500 (200; 1000) 1000 (500; 2000)
Total ETB 24 (11; 58) NPR 604 (500; 900) NGN 760 (420; 1400) YER 2550 (1400; 4850)
Total (US $) $1.47 $8.22 $5.00 $11.89
Proportion of poor/general population (H) using MPIc 0.900 0.647 0.635 0.525
Proportion of poor/general population < $1.25/dayd 0.390 0.551 0.644 0.175
Conversion rate to USD 1 = 0.061 1 = 0.013 1 = 0.007 1 = 0.005
a
Values are medians and 25–75 % interquartile ranges (IQR) and given in the local currency, unless specified otherwise
b
Data missing for 10 patients (2 in Ethiopia, 3 in Nepal and 5 in Nigeria)
c
Figures derived from the Multidimensional Poverty Index [34]
d
Proportion of people living on < $1.25/day, World Development Indicators, World Bank 2009

confirmation and expenses were similar among males and consultations, service fees, travel expenses and lost time
females. and opportunity costs [17]. This analysis confirms that
A total of 538 cases with expenditure >75th centile the direct costs sustained by patients undergoing a diag-
(high expenditure) and 1679 patients with expend- nosis of TB across multiple settings are substantial. A
iture <75th centile (controls) were analysed. Factors asso- large component of these are associated with clinic
ciated with high patient expenditure at the univariate and costs, transport and patients attending the services with
multivariate analysis are shown in Table 5. company. According to the Multidimensional Poverty
Although there were considerable differences be- Index, of the four countries Yemen, ranked as the least
tween countries, residing in a rural area/other town poor at the time of the study, followed by Nigeria, Nepal
(OR (95 % CI) 23.6 (10.2–55); 1.5 (1.0–2.2) and 6.9 (4.2– and Ethiopia [2]. Although it is difficult to compare
11.3), respectively), having low education (2.5 (1.3–4.8), 1.4 costs and expenditure directly between countries, as
(1.0–2.1) and 1.9 (1.1–3.3), respectively) and attending with living costs and income were very different across study
company (19.7 (6.1, 63), 1.5 (1.0–2.2) and 3.4 (1.7–7.0), re- settings, our findings identified remarkable similarities.
spectively) were risk factors for high expenditure in Clinic user fees comprised a common and significant
Ethiopia, Nepal and Yemen at the univariate level. Of cost. These fees are known to have a negative impact on
these, illiteracy (OR (95 % CI) 2.4 (1.4–4.3) and being general health service utilisation and this is likely to be
accompanied (OR (95 % CI) 2.7 (1.8–4.2) were also statis- more prominent in TB patients with limited financial
tically significant in Nigeria. Variables independently asso- resilience [18]. Clinic costs for attending TB clinics in
ciated with high expenditure in the multivariate analysis, the study comprised consultation fees, smear micros-
were attending the clinic with company (Ethiopia (AOR copy, X-rays and blood tests to screen for other diseases
7.5 (2.3–25.2), Yemen (AOR 3.1 (1.5–6.5) and Nigeria [19]. Furthermore, although not captured in the study,
(AOR 2.5 (1.6–3.9)), residing in rural areas and other bacteriologically negative cases may undergo further
towns (Ethiopia (AOR 15.3 (6.5–36.0) and Yemen 6.7 consultation and testing, pay for further visits and have
(4.1–11.0)), illiteracy (Nigeria (AOR 1.9 (1.0–3.4) and higher expenditure than smear-positive cases.
Nepal (AOR 2.5 (1.5–4.0)) and males with bacteriolo- Although some NTPs retrospectively reimburse diag-
gically confirmed TB (Nepal, AOR 2.0 (1.3–3.2)). nostic costs to patients with TB, only 10–20 % of patients
Many illnesses are associated with poverty and individ- receive a TB diagnosis. Patients therefore need to be pre-
uals often consider their financial position before attend- pared to meet expenses up front; in practise 80–90 % will
ing health facilities. TB is no exception. As a disease of not be reimbursed and these fees are likely to be a major
poverty, TB is associated with many patients never at- deterrent for attending diagnostic services. Furthermore,
tending health services, attending late, or dropping out we have documented elsewhere that patients are often
after initiation of the diagnostic process [10, 15, 16]. As overcharged or pay under-the-counter fees to speed up
TB diagnostic services are not available in all health test results or to be seen earlier than others [20]. These
facilities, patients often express concerns about multiple expenses are not documented in their receipts and where
de Cuevas et al. Infectious Diseases of Poverty (2016) 5:24 Page 6 of 9

Table 4 Median costs incurred according to patients’ characteristics

Variablea Ethiopia N = 504 Nepal N = 619 Nigeria N = 502 Yemen N = 600
N ETB N NPR N NGN N YER
Age (years) <20 52 21 (8; 48) 37 600 (500; 740)*** 13 1100 (850; 2300)**** 54 2240 (1400; 3850)
20–49 367 24 (11; 58) 332 600 (500; 858) 444 725 (400; 1390) 312 2550 (1400; 4795)
≥50 83 26 (12; 60) 250 640 (560; 980) 40 800 (525; 1800) 232 2760 (1445; 5050)
Gender Male 278 24 (11; 56) 395 610 (500; 976) 241 700 (420; 1400) 328 2525 (1305; 4650)****
Female 224 24 (11: 60) 224 600 (500; 849) 256 800 (480; 1400) 271 2600 (1600; 5450)
Residency Rural 182 43 (24; 79)* 63 725 (460; 1890)* 82 700 (480; 1400) 279 3890 (2250; 6250)*
Same town 215 12 (5; 24) 284 598 (500; 749) 408 760 (400; 1400) 267 1650 (890; 2654)
Other town 105 56 (40; 82) 272 648 (572; 895) 7 1660 (850; 1800) 53 4250 (2450; 7360)
Companions Alone 129 9 (5; 14)* 181 600 (500; 750)** 320 645 (400; 1100)* 90 1420 (800; 2450)*
Accompanied 373 38 (24; 72) 435 620 (500; 980) 177 1100 (600; 1900) 509 2850 (1650; 5250)
Diagnosis Completed 488 24 (11; 58) 602 600 (500; 885) 423 750 (400; 1400) 569 2550 (1400; 4890)
$1.47 $8.16 $4.94 $11.89
Incomplete 14 27.5 (8; 48) 16 868 (400; 1188) 47 750 (500; 1400) 30 2675 (1650; 4050)
$1.68 $11.81 $4.94 $12.47
Smear result Positive 111 28 (13; 53) 71 700 (560; 1160)*** 85 800 (480; 1400) 120 2610 (1450; 4850)
Negative 389 24 (11; 58) 545 600 (500; 850) 403 740 (400; 1400) 476 2550 (1400; 4930)
Culture Positive 141 29 (13; 56)**** 73 672 (580; 1160)*** 80 770 (400; 1400) 150 2535 (1400; 4800)
Negative 345 24 (11; 58) 515 600 (500; 840) 399 760 (420; 1400) 430 2555 (1450; 5050)
Illness duration (weeks) 1–2 90 24 (11; 48)**** 63 540 (270; 1200)** 133 750 (420; 1400) 162 2625 (1510; 5050)
3–4 153 24 (9; 52) 153 596 (500; 780) 144 800 (400; 1400) 140 2400 (1325; 4610)
5–11 120 38 (17; 64) 129 600 (540; 950) 91 800 (400; 1440) 131 2850 (1350; 4890)
≥12 137 24 (11; 72) 271 640 (560; 976) 118 700 (500; 1220) 153 2480 (1550; 4750)
Cough duration (weeks) 1–2 95 24 (11; 48)**** 80 565 (300; 1148)** 165 800 (400; 1500) 196 2650 (1455; 5100)
3–4 160 24 (9; 50) 169 600 (500; 800) 134 740 (400; 1380) 132 2450 (1450; 4725)
5–11 115 36 (16; 64) 130 630 (540; 1020) 84 800 (490; 1440) 118 2425 (1250; 4750)
≥12 128 24 (11.5; 72) 238 623 (580; 900) 104 700 (500; 1370) 138 2515 (1610; 4750)
Employment status Working 234 24 (10; 56) 202 598 (500; 780)** 367 700 (400; 1300)** 163 2150 (1090; 4050)**
Not working 268 28 (11.5; 63) 417 640 (540; 900) 130 1000 (520; 1600) 436 2650 (1580; 5100)
Occupation Farmer 125 52 (24; 80)* 133 648 (552; 980) 32 800 (535; 1600)*** 140 3375 (2000; 6050)*
Student 47 24 (11; 52) 66 640 (500; 900) 46 1150 (550; 1900) 58 2250 (1250; 4050)
Housewife 121 42 (24; 73) 146 611.5 (500; 848) 22 690 (400; 1040) 210 2850 (1650; 6050)
Labourer 38 11 (5; 21) 21 600 (480; 1080) 21 1200 (640; 1440) 71 2050 (1120; 3250)
Merchant 32 22.5 (8; 52) 97 600 (536; 720) 1 20000 (−; −) 8 1925 (975; 2950)
Government 45 13 (7; 27) 33 600 (540; 700) 86 720 (400; 1300) 61 2500 (1250; 4250)
Other 92 16 (8; 27.5) 117 600 (500; 975) 289 700 (400; 1320) 51 1300 (800; 2950)
a
Values are medians and 25–75 % interquartile ranges (IQR) and given in the local currency, unless specified otherwise
*P <0.001, **P < 0.01, ***P < 0.05, ****P < 0.2

reimbursement does occur, patients are often only par- and the cost of transport from areas with limited road
tially reimbursed. and public transport infrastructure. In Yemen, collective
Transport costs also contribute significantly to ex- public transport is limited to the main towns and
penditure in Ethiopia, Nigeria and Yemen, where a high women often rely on private vehicles to attend the fairly
number of patients travel from other towns and rural centralised diagnostic services. In Nigeria, buses were
areas and use buses and taxis. In Ethiopia, the costs re- very limited within the metropolitan areas of the Federal
flect the predominantly rural population of the country Capital at the time of the study and people relied on
de Cuevas et al. Infectious Diseases of Poverty (2016) 5:24 Page 7 of 9

Table 5 Risk factors for high patient expenditure, odds ratios (OR), adjusted odds ratios (AOR) with 95 % Confidence intervals (95 % CI)
by country
Ethiopia Nepal Nigeria Yemen
Variable OR AOR OR AOR OR AOR OR AOR
Age ≥50 1.1 (0.6–1.8) 1.3 (0.9–1.8) 1.6 (0.8–3.2) 1.3 (0.9–1.9)
Male 0.9 (0.6–1.4) 1.4 (0.9–2.1) 2.0 (1.3–3.2) 1.1 (0.7–1.6) 0.7 (0.5–1.1)
Married or with partner 1.9 (1.3–3.0) 0.8 (0.5–1.3) 1.0 (0.7–1.5) 1.1 (0.7–1.7)
Rural/other town residency 23.6 (10.2–55) 15.3 (6.5–36.0) 1.5 (1.0–2.2) 1.1 (0.6–1.8) 6.9 (4.2–11.3) 6.7 (4.1–11.0)
Illiterate 1.9 (1.3–2.9) 1.7 (1.1–2.6) 2.5 (1.5–4.0) 2.4 (1.4–4.3) 1.9 (1.0–3.4) 1.7 (1.2–2.5)
Education nil or primary 2.5 (1.3–4.8) 1.4 (1.0–2.1) 1.1 (0.7–1.7) 1.9 (1.1–3.3)
Accompanied 19.7 (6.1, 63) 7.5 (2.3–25.2) 1.5 (1.0–2.2) 2.7 (1.8–4.2) 2.5 (1.6–3.9) 3.4 (1.7–7.0) 3.1 (1.5–6.5)
Smear-positive 0.9 (0.5–1.4) 2.2 (1.3–3.7) 0.8 (0.5–1.5) 1.0 (0.6–1.5)
Culture-positive 1.0 (0.6–1.6) 2.2 (1.3–3.7) 2.3 (1.3–3.8) 1.0 (0.6–1.7) 0.9 (0.6–1.4)
Illness ≥ 12 weeks 1.6 (1.0–2.4) 1.2 (0.8–1.7) 0.9 (0.5–1.5) 0.9 (0.6–1.5)
Cough ≥ 5 to 11 weeks 1.3 (0.8–2.0) 1.3 (0.9–2.1) 1.2 (0.7–2.0) 1.0 (0.6–1.6)
Not working 1.2 (0.8–1.9) 1.1 (0.8–1.7) 1.7 (1.1–2.7) 1.4 (0.9–2.1)
Farmer, student, 3.8 (2.2–6.5) 1.1 (0.8–1.6) 1.4 (0.9–2.3) 2.3 (1.3–4.0)
housewife, labourer

share minibuses or taxis to move from the slums sur- questionnaire when they present to diagnostic centres.
rounding the metropolitan areas. Although transport This is particularly relevant in the context of the Global
costs were incurred for different reasons in each loca- Plan to Stop TB 2016–2020, which aims to eliminate the
tion, they represented more than half of patient expend- number of families facing catastrophic costs due to TB
iture to attend diagnostic centres and are thus a major [1, 22]. In the study context, additional support made
barrier. Addressing these issues would require long term available by NTPs could be channelled to those most at
infrastructural development. However interim solutions risk of high costs, including illiterate patients, those ori-
such as recruiting community volunteers or community ginating from rural areas or attending with company.
workers to organise transport and the provision of sup- This approach could also reimburse expenses to all pa-
port funds that facilitate transport could be explored tients investigated for TB; independently of whether the
[21]. Similar issues occur for other diseases such as diagnosis is confirmed.
HIV/AIDS, the management of accidents and emergen- This analysis, however, has several limitations, beginning
cies and complicated deliveries. with the sampling strategy, which carried a risk of selection
A high proportion of patients attended with company in bias. Participants were recruited using systematic random
all study settings. This was especially prominent in Yemen, sampling, rather than randomly, as this suited the objec-
where 4 out of 5 patients came accompanied, and signifi- tives of the larger clinical trial [14]. Patients arriving at the
cantly increased the costs associated with displacement. In beginning of the day might have had different characteris-
Yemen, gender norms meant that women in particular tics from those arriving later. We can hypothesise that the
were required to be accompanied by a male relative and former might have resided more locally, or conversely,
faced particular challenges in accessing a TB diagnosis [20]. have travelled the previous day from afar and stayed over-
The need for company also reflects cultural practices night. Patients arriving early might have been better pre-
underpinning support to a person perceived to have a mor- pared. Next, second day costs were calculated according to
tal illness - in Ethiopia, for example, TB is equated with patients’ predicted expenses, rather than the actual costs
lung cancer - and the frailty of patients with chronic and incurred. Moreover, all costs were self-reported, rather
debilitating conditions. Rural residents were more likely to than observed by investigators. Actual and predicted ex-
attend with company in all countries, which is probably penditure might have been expressed differently by differ-
the reason why these patients had higher expenditure in ent subgroups of the population and different cultures, as
Ethiopia and Yemen than patients from urban areas [11]. mediated by established social hierarchies, gender roles,
Our findings indicate that despite the differences in economic standing and the distribution of power, to name
the settings, patients across LMICs experience many but a few modifiers. For example, costs which might be
similarities in the type of costs associated with clinic at- overstated by patients in one setting in the hope of finan-
tendance and that it might be feasible to identify patients cial remuneration might be underreported in another out
at risk of high expenditure by conducting a simple of individual pride. Women who do not have access to
de Cuevas et al. Infectious Diseases of Poverty (2016) 5:24 Page 8 of 9

household finances might also be unaware of the full cost was approved by the WHO Ethics Review Committee,
of attendance. Calculation of costs as a proportion of indi- the Liverpool School of Tropical Medicine Ethics
vidual income would have provided a more complete pic- Research Committee, and the national and institutional
ture of the economic burden for the individual and their ethics committees of the four countries. Consent and in-
family, however including the many variables involved in formation sheets were translated, and informed witnessed
these calculations was not possible within the confines of written or oral consent was obtained.
a short addition to a large survey and asking participants
to disclose their income was considered unreliable. Additional file
The costs described here represent only a proportion of
the economic barriers faced by individuals with symptoms Additional file 1: Multilingual abstracts in the six official working
languages of the United Nations. (PDF 314 kb)
of TB, as the cost of previous healthcare, visits to the pri-
vate sector and indirect costs - such as loss of employment
Abbreviations
and work time – which may constitute a major part of pa- AOR: adjusted odds rations; CI: Confidence Intervals; DFiD: Department for
tients’ outgoings, were not documented [19, 20, 23–26]. International Development; ESRC: Economic and Social Research Council;
Opportunity costs tend to be higher for people living in IQR: interquartile range; NTI: National Tuberculosis Institute; NTPs: national TB
programmes; OR: Odds Ratios; SD: standard deviation; TB: tuberculosis;
poverty, who for the most part work in the informal sector SDGs: Sustainable Development Goals.
and are vulnerable to loss of income or dismissal from
work [27]. In countries with high HIV prevalence, this vul- Competing interests
The authors declare that they have no competing interests.
nerability is heightened, as the population perceives that
patients with TB are likely to be co-infected with HIV [28]. Authors’ contributions
Furthermore, a large proportion of symptomatic adults do RMA, MAY, SJT and LEC conceived the study, lead the studies and wrote the
not attend diagnostic centres and these individuals often study protocol with AR. RMA and LEC conducted data analysis and wrote
the initial manuscript. LL, OO, NA-S, NA-A, IA, DG and JBS coordinated the
have fewer financial and social resources at their disposal studies in Nigeria, Yemen, Ethiopia and Nepal. EEN conducted laboratory
[20]. Financial barriers could be alleviated at the point of tests for participants in Nigeria. STA and JO and were the State and National
care through further decentralisation of health services TB control programme managers at the time of the study, facilitate the study
inception into their control programmes and advised on study design.
[29], the provision of free diagnostic services, transport for DG coordinated the studies in Ethiopia. AR and SBS provided advised on
remote populations and/or mobilisation of services to re- TB control policies and how to focus the study within international priorities,
mote communities [28, 30, 31]. supported the development of protocols and commented on earlier drafts
of the manuscript. All authors contributed to the final manuscript. All authors
read and approved the final manuscript.
Conclusion
The costs incurred by patients are substantial and share Acknowledgements
This research was funded by the Bill and Melinda Gates Foundation and the
common patterns across countries. Removing user fees, United States Agency for International Development through grants
transparent charging policies and reimbursing clinic ex- awarded to the UNICEF/UNDP/World Bank/WHO Special Programme for
penses would reduce the poverty-inducing effects of direct Research and Training in Tropical Diseases (Grant Nos. 3636/9900727 and
AAGG-00-99-00005-31) and by a joint research programme of the Economic
diagnostic costs. In locations with limited resources, sup- and Social Research Council (ESRC) and the Department for International
port could be prioritised for those most at risk of high ex- Development (DfID) (Grant Reference: RES-167-25-0387). The funders had no role
penditure; those who are illiterate, attend the service with in study design, data collection and analysis, decision to publish, or preparation
of the manuscript. The authors of this paper alone are responsible for the views
company and rural residents. expressed in this publication which do not necessarily represent the decisions or
Poverty constitutes a major access barrier for symptom- policies of the World Health Organization and their institutions. We are also
atic adults in low income countries, as recently recognised grateful to all laboratory staff and TB control management personnel who
supported the enrolment of patients to this and allied studies. Without their help
by the Sustainable Development Goals (SMGs) [32]. Pov- this study would not have been possible.
erty is often compounded by low education and health in-
formation, leading to misconceptions of the disease and Author details
1
Liverpool School of Tropical Medicine, Pembroke Place, Liverpool L3 5QA,
disempowerment. Rural and urban residence also deter- UK. 2Zankli Medical Center, Abuja, Nigeria. 3Bingham University, Nassarawa
mines access to diagnostic services. Addressing poverty is State, Abuja, Nigeria. 4Medical Faculty, Sana’a University, Sana’a, Yemen.
5
likely to be the most crucial factor in determining the suc- National Tuberculosis Institute, Sana’a, Yemen. 6Bushullo Major Health
Centre, Awassa, Ethiopia. 7Tribhuvan University Institute of Medicine,
cess or failure of the Global Health Plan to Stop TB 2016– Kathmandu, Nepal. 8Armauer Hansen Research Institute, Addis Ababa,
2020 as a significant public health problem [1] and recog- Ethiopia. 9The Global Fund to Fight AIDS, Tuberculosis and Malaria, Geneva,
nised as major impediment for the achievement of the TB Switzerland. 10Federal Capital Territory Tuberculosis and Leprosy Control
Programme, Abuja, Nigeria. 11Nigeria Tuberculosis and Leprosy Control
component of the SMGs [17, 33]. Programme, Abuja, Nigeria. 12REACH Ethiopia, Hawassa, Ethiopia. 13UNICEF/
UNDP/World Bank/WHO Special Programme for Research and Training in
Ethics statement Tropical Diseases, World Health Organization, Geneva, Switzerland.
The protocol (International Standard Randomized Received: 23 November 2015 Accepted: 14 March 2016
Controlled Trial Number Register ISRCTN53339491)
de Cuevas et al. Infectious Diseases of Poverty (2016) 5:24 Page 9 of 9

References 23. Tanimura T, Jaramillo E, Weil D, Raviglione M, Lonnroth K. Financial burden

1. Uplekar M, Weil D, Lonnroth K, Jaramillo E, Lienhardt C, Dias HM, et al. for tuberculosis patients in low- and middle-income countries: a systematic
WHO’s new end TB strategy. Lancet. 2015;385(9979):1799–801. review. Eur Respir J. 2014;43(6):1763–75. doi:10.1183/09031936.00193413.
doi:10.1016/s0140-6736(15)60570-0. 24. Nguyen HT, Hickson RI, Kompas T, Mercer GN, Lokuge KM. Strengthening
2. Lonnroth K, Glaziou P, Weil D, Floyd K, Uplekar M, Raviglione M. Beyond tuberculosis control overseas: who benefits? Value Health. 2015;18(2):180–8.
UHC: monitoring health and social protection coverage in the context of doi:10.1016/j.jval.2014.11.008.
tuberculosis care and prevention. PLoS Med. 2014;11(9):e1001693. 25. Oliwa JN, Karumbi JM, Marais BJ, Madhi SA, Graham SM. Tuberculosis as a
doi:10.1371/journal.pmed.1001693. cause or comorbidity of childhood pneumonia in tuberculosis-endemic
3. WHO. Global tuberculosis report 2014. 2014. areas: a systematic review. Lancet Respir Med. 2015;3(3):235–43.
4. Kik SV, Denkinger CM, Chedore P, Pai M. Replacing smear microscopy for doi:10.1016/S2213-2600(15)00028-4.
the diagnosis of tuberculosis: what is the market potential? Eur Respir J. 26. Hu H, Chen J, Sato KD, Zhou Y, Jiang H, Wu P, et al. Factors that associated
2014;43(6):1793–6. doi:10.1183/09031936.00217313. with TB patient admission rate and TB inpatient service cost: a cross-sectional
5. Lönnroth K, Jaramillo E, Williams B, Dye C, Raviglione M. Tuberculosis: the study in China. Infect Dis Poverty. 2016;5(1):4. doi:10.1186/s40249-016-0097-x.
role of risk factors and social determinants. In: Blas E, Sivasankara Kurup A, 27. Bates I, Fenton C, Gruber J, Lalloo D, Lara AM, Squire SB, et al. Vulnerability
editors. Equity, social determinants and public health programmes. Geneva: to malaria, tuberculosis, and HIV/AIDS infection and disease. Part 1:
World Health Organization; 2010. p. 219–41. determinants operating at individual and household level. The Lancet
6. Kemp JR, Mann G, Simwaka BN, Salaniponi FML, Squire SB. Can Malawi’s Infectious Diseases. 2004;4:267–77.
poor afford free tuberculosis services? Patient and household costs 28. WHO. Addressing poverty in TB control: options for national TB control
associated with a tuberculosis diagnosis in Lilongwe. Bulletin of the World programmes. Geneva: World Health Organization; 2005.
Health Organization. 2007;85(8):580–5. 29. Tulloch O, Theobald S, Morishita F, Datiko DG, Asnake G, Tesema T, et al.
7. Yan F, Thomson R, Tang S, Squire SB, Wang W, Liu X, et al. Multiple perspectives Patient and community experiences of tuberculosis diagnosis and care
on diagnosis delay for tuberculosis from key stakeholders in poor rural China: within a community-based intervention in Ethiopia: a qualitative study.
Case study in four provinces. Health Policy. 2007;82:186–99. BMC Public Health. 2015;15:187. doi:10.1186/s12889-015-1523-x.
8. Benatar SR, Upshur R. Tuberculosis and poverty: what could (and should) be 30. Yassin MA, Datiko DG, Tulloch O, Markos P, Aschalew M, Shargie EB, et al.
done? Int J Tuberc Lung Dis. 2010;14(10):1215–21. Innovative community-based approaches doubled tuberculosis case
9. Pantoja A, Kik SV, Denkinger CM. Costs of novel tuberculosis diagnostics– notification and improve treatment outcome in Southern Ethiopia.
will countries be able to afford it? J Infect Dis. 2015;211 Suppl 2:S67–77. PLoS One. 2013;8(5):e63174. doi:10.1371/journal.pone.0063174.
doi:10.1093/infdis/jiu820. 31. Arinaminpathy N, Dowdy D. Understanding the incremental value of
10. Cambanis A, Ramsay A, Yassin MA, Cuevas LE. Duration and associated factors novel diagnostic tests for tuberculosis. Nature. 2015;528(7580):S60–7.
of patient delay during tuberculosis screening in rural Cameroon. Trop Med Int doi:10.1038/nature16045.
Health. 2007;12(11):1309–14. doi:10.1111/j.1365-3156.2007.01925.x. 32. Nations U. Sustainable Development Goals. 2015. https://sustainabledevelopment.
11. Ramsay A, Al-Agbhari N, Scherchand J, Al-Sonboli N, Almotawa A, Gammo M, un.org/?menu=1300. Accessed 4 Jan 2016.
et al. Direct patient costs associated with tuberculosis diagnosis in Yemen and 33. WHO. Global Tuberculosis Report 2015. Geneva: WHO; 2015. http://www.
Nepal. Int J Tuberc Lung Dis. 2010;14(2):165–70. who.int/tb/publications/global_report/en/. Accessed 4 Jan 2016.
12. Cambanis A, Yassin MA, Ramsay A, Squire SB, Arbide I, Cuevas LE. A one-day 34. Alkire S, Santos ME. Multidimensional Poverty Index: Oxford Poverty and
method for the diagnosis of pulmonary tuberculosis in rural Ethiopia. Int J Human Development Initiative. 2010.
Tuberc Lung Dis. 2006;10(2):230–2.
13. Hirao S, Yassin MA, Khamofu HG, Lawson L, Cambanis A, Ramsay A, et al.
Same-day smears in the diagnosis of tuberculosis. Trop Med Int Health.
2007;12(12):1459–63. doi:10.1111/j.1365-3156.2007.01952.x.
14. Cuevas LE, Yassin MA, Al-Sonboli N, Lawson L, Arbide I, Al-Aghbari N, et al.
A multi-country non-inferiority cluster randomized trial of frontloaded smear
microscopy for the diagnosis of pulmonary tuberculosis. PLoS Med.
2011;8(7):e1000443. doi:10.1371/journal.pmed.1000443.
15. Cambanis A, Yassin MA, Ramsay A, Bertel Squire S, Arbide I, Cuevas LE. Rural
poverty and delayed presentation to tuberculosis services in Ethiopia. Trop
Med Int Health. 2005;10(4):330–5. doi:10.1111/j.1365-3156.2005.01393.x.
16. Storla DG, Yimer S, Bjune GA. A systematic review of delay in the
diagnosis and treatment of tuberculosis. BMC Public Health. 2008;8:15.
doi:10.1186/1471-2458-8-15.
17. Foster S. The economics of tuberculosis diagnosis and treatment. In: Porter J,
Grange J, editors. Tuberculosis: an interdisciplinary perspective. London:
Imperial College Press; 1999.
18. Lagarde M, Palmer N. The impact of user fees on access to health services
in low- and middle-income countries. Cochrane Database Syst Rev.
2011(4):Cd009094. doi:10.1002/14651858.cd009094.
19. Sewell J, Capocci S, Johnson J, Solamalai A, Hopkins S, Cropley I, et al.
Expanded blood borne virus testing in a tuberculosis clinic. A cost and yield Submit your next manuscript to BioMed Central
analysis. J Infect. 2015;70(4):317–23. doi:10.1016/j.jinf.2014.10.014. and we will help you at every step:
20. de Cuevas RM A, Al-Sonboli N, Al-Aghbari N, Yassin MA, Cuevas LE,
Theobald SJ. Barriers to completing TB diagnosis in Yemen: services • We accept pre-submission inquiries
should respond to patients’ needs. PLoS One. 2014;9(9):e105194. • Our selector tool helps you to find the most relevant journal
doi:10.1371/journal.pone.0105194.
• We provide round the clock customer support
21. Li Y, Ehiri J, Hu D, Oren E, Cao J. Framework of behavioral indicators
evaluating TB health promotion outcomes: a modified Delphi study of • Convenient online submission
TB policymakers and health workers. Infect Dis Poverty. 2015;4:56. • Thorough peer review
doi:10.1186/s40249-015-0087-4.
• Inclusion in PubMed and all major indexing services
22. The Global Plan to Stop TB 2016–2020: “Bending the Curve: A global
investment framework to win the fight against TB”. Stop TB Partnership. • Maximum visibility for your research
http://www.stoptb.org/global/plan/plan2/. Accessed 22 Mar 2016.
Submit your manuscript at
www.biomedcentral.com/submit