International Journal of Infectious Diseases 125 (2022) 51–57

Contents lists available at ScienceDirect

International Journal of Infectious Diseases

journal homepage: www.elsevier.com/locate/ijid

Comparison between a rapid diagnostic test and dried blood

spot-based immunoassay for hepatitis B surface antigen testing:
Performance and cost implications in a population-based serosurvey
in Vietnam
Sumiyo Okawa a,∗, Kenichi Komada a, Yasunori Ichimura a, Masaya Sugiyama b,
Hung Thai Do c, Huy Xuan Le d, Thanh Tien Hoang e, Trieu Bao Nguyen f, Mai Kim Huynh f,
Hang Thi Hai Hoang e, Nhu Anh Thi Tran f, Thieu Hoang Le e, Quyet Thi Ngo f,
Shinsuke Miyano a, Yuta Yokobori a, Yosuke Inoueg, Tetsuya Mizoueg, Masahiko Hachiya a
a
Bureau of International Health Cooperation, National Center for Global Health and Medicine, Sumiyo Okawa, 1-21-1, Toyama, Shinjuku-ku, Tokyo
162-8655, Japan
b
Genome Medical Science Project, The Research Center for Hepatitis and Immunology, National Center for Global Health and Medicine, Chiba, Japan
c
Pasteur Institute in Nha Trang, Nha Trang, Khánh Hòa, Viet Nam
d
Medical Health Service Center, Pasteur Institute in Nha Trang, Nha Trang, Khánh Hòa, Viet Nam
e
Department of Infectious Disease Control and Prevention, Pasteur Institute in Nha Trang, Nha Trang, Khánh Hòa, Viet Nam
f
Department of Microbiology and Immunology, Pasteur Institute in Nha Trang, Nha Trang, Khánh Hòa, Viet Nam
g
Department of Epidemiology and Prevention, National Center for Global Health and Medicine, Tokyo, Japan

a r t i c l e i n f o a b s t r a c t

Article history: Objectives: This study aimed to determine the agreement between a rapid diagnostic test (RDT) and
Received 8 March 2022 a dried blood spot (DBS)-based electrochemiluminescence immunoassay (ECLIA) of hepatitis B surface
Revised 1 October 2022
antigen and to compare the costs of conducting serosurveys using RDTs and DBS in a field setting.
Accepted 6 October 2022
Methods: A serosurvey was conducted in the South Central Coast region of Vietnam in May 2019. Partic-
ipants aged 1-39 years were recruited using a four-stage random sampling method and tested for hepati-
Keywords: tis B surface antigen using an RDT kit (Alere Determine) and a DBS-based ECLIA. The agreement between
Hepatitis B the RDT and the DBS-based ECLIA was assessed using cross-tabulation and Cohen kappa. Cost data were
Seroprevalence survey categorized by input (personnel, transportation, field consumables, laboratory consumables, and capital
Cost comparison
item/overhead) and survey phase (survey preparation, data/biospecimen collection, laboratory testing, and
Rapid diagnostic tests
coordination).
Dried blood spots
Vietnam Results: A total of 2072 participants were analyzed. There was a 99% agreement between the RDT and
the DBS-based ECLIA results, with a Cohen kappa of 0.9. The estimated cost of conducting a serosurvey
by DBS was UD $75,291, whereas RDT was $53,182.
Conclusion: RDTs and DBS-based ECLIA provide test results with high agreements. RDTs are a better
option in terms of cost, whereas the DBS-based ECLIA may be useful when evaluating multiple infectious
diseases.
© 2022 The Authors. Published by Elsevier Ltd on behalf of International Society for Infectious Diseases.
This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/)

Introduction by infections acquired within the first 5 years of life, particu-

larly at birth (World Health Organization [WHO], 2020). Although
Hepatitis B virus (HBV) infection causes acute and chronic the introduction of the hepatitis B (HB) vaccination program ex-
liver disease. Most HBV-related deaths among adults are caused panded worldwide in the late 1990s (Namgyal, 2003) and has
averted millions of chronic HBV infections and deaths in chil-
dren (WHO, 2018), the disease burden associated with HBV in-
∗
Corresponding author.
fection remains high; an estimated 1.5 million people were newly
E-mail address: [email protected] (S. Okawa).

https://doi.org/10.1016/j.ijid.2022.10.011
1201-9712/© 2022 The Authors. Published by Elsevier Ltd on behalf of International Society for Infectious Diseases. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
S. Okawa, K. Komada, Y. Ichimura et al. International Journal of Infectious Diseases 125 (2022) 51–57

infected with chronic HBV infection, and 0.82 million died from reference in designing a serosurvey, such as an evaluation survey
HBV-related disease in 2019 (WHO, 2021a). of an immunization program.
In 2016, the WHO called for the elimination of HBV infection
by 2030 (WHO, 2020). In the Western Pacific region, nearly 70% of Methods
the countries had achieved an HB surface antigen (HBsAg) preva-
lence of <1% among children aged 5 years by December 2017 Study design
(Woodring et al., 2019), whereas the prevalence in African and
Eastern Mediterranean regions in 2015 were 3.0% and 1.6%, respec- This field-based serosurvey was conducted in the South Cen-
tively (WHO, 2017a). In these regions with high prevalence, evalu- tral Coast region of Vietnam in May 2019. The region consisted of
ating the magnitude of HBV infection is particularly important to the independent municipality of Da Nang and seven provinces, of
control the disease burden, and WHO recommends HBsAg serosur- which we selected four provinces (Quang Ngai, Binh Dinh, Khanh
veys as a priority action (WHO, 2016, 2017b, 2017c). However, HB- Hoa, Ninh Thuan) to include urban and rural areas as study sites.
sAg serosurveys have not been conducted regularly (Allison et al.,
Study participants
2016; Breakwell et al., 2017).
The accuracy of a serosurvey result is important to evaluate im-
The participant inclusion criteria were being a resident of a
munization activities, which is generally affected by various fac-
study site and aged 1-39 years as of May 20, 2019. The sample
tors, such as assay formats. For example, laboratory-based enzyme
size was calculated based on an expected HBsAg seroprevalence of
immunoassay (EIA) using serum or plasma specimens is the rec-
10%. A sample size of 208 for each 5-year age group was needed
ommended standard for detecting HBV infection because of its
based on a 5% level of significance with ± 5% precision and an as-
high sensitivity and specificity (WHO, 2011). In field-based sero-
sumed design effect of 1.5. We targeted eight 5-year age groups,
surveys, however, drawing venous blood is often not feasible, par-
resulting in a required sample size of 1664. We assumed a poten-
ticularly in serosurveys involving children; thus, rapid diagnostic
tial response rate of 70% due to the absence or lack of an indi-
tests (RDTs) or EIAs that use dried blood spot (DBS) samples have
vidual in the target age group being at home at the time of our
been widely used as an alternative (WHO, 2011). High sensitiv-
visit. Further, we considered the population pyramid of Vietnam
ity and specificity of HBsAg RDT (90.0%, and 99.5%, respectively)
in 2014 (General Statistics Office and United Nations Population
(Amini et al., 2017) and DBS-based HBsAg testing (98%, and 100%,
Fund, 2016), in which the number of household members aged 1-
respectively) (Lange et al., 2017) compared with laboratory-based
39 years per household was estimated to be 2.16, with an addi-
EIAs and venous blood-based testing have been shown in pre-
tional 20% to make up shortfalls due to variability in HBsAg preva-
vious meta-analysis studies. However, no studies have examined
lence among age groups. Based on these assumptions, the target
the level of agreement between the test results of an RDT and a
sample size was 1887 households.
DBS-based electrochemiluminescence immunoassay (ECLIA) using
Study participants were recruited using the four-stage random
finger-prick blood samples from the same participants obtained by
sampling. We defined the provinces as strata and villages as clus-
cluster sampling in such a setting.
ters. In the first stage, three districts per province were randomly
In addition to the accuracy of estimated seroprevalence, the
selected according to probability proportional to size, based on the
cost and logistical challenges in collecting samples are critical is-
population data from the 2015 national census of Vietnam. Follow-
sues in serosurveys (Metcalf et al., 2016) because biospecimen col-
ing the same procedure, two communes per district, two villages
lection entails an additional cost for trained fieldworkers, supplies,
per commune, and 40 households per selected village were ran-
storage, transportation, laboratory processing, and testing (Carcelen
domly selected by probability proportional to size. All household
et al., 2020; MacNeil et al., 2014). For example, one study con-
members who met the age criteria in the selected households were
ducted in Zambia examined the cost of a field-based serosurvey
recruited for the study.
of antimeasles immunoglobulin G and antirubella immunoglobulin
G using DBS-based EIA (Carcelen et al., 2020). The Zambian sero- Questionnaire survey
survey cost a total of $68,558 to collect and analyze 658 samples
(i.e., $104.2 per sample), of which the costs for planning, personnel, Trained fieldworkers conducted face-to-face interviews with
mapping, enumeration, and transportation were saved by nesting study participants in their homes to obtain data on their sociode-
it in a large postcampaign measles and rubella vaccine coverage mographic characteristics and children’s birth and vaccination pro-
study (Carcelen et al., 2020). file using a structured questionnaire.
HB is endemic in Vietnam, with an estimated prevalence of
10.8% among the general population (Schweitzer et al., 2015). Hepatitis B surface antigen testing
The country introduced universal infant HB vaccination in 2003
(Nguyen et al., 2008) and issued a national plan for 2015-2019 that After the interview, the fieldworkers collected finger-prick cap-
outlined the core areas in prevention, testing, and treatment of vi- illary blood samples using a safety lancet and glass capillary tube
ral hepatitis (Chan et al., 2021). Although some studies have exam- and applied them to the sample pad of the RDT kit (Alere De-
ined HB prevalence among specific generations or populations in termine HBsAg; Alere Medical Co. Ltd., Chiba, Japan). The RDT kit
Vietnam (Duong et al., 2020; Ishizaki et al., 2017; Lam et al., 2013; has demonstrated a high clinical performance, with a finger-prick
Nguyen et al., 2014; Ngoc Thanh et al., 2020; Viet et al., 2012), no whole-blood sensitivity and specificity of 97.2% and 98.5%, respec-
studies have reported seroprevalence of HBV infection among the tively (Avellon et al., 2020). The fieldworkers read the test result 15
general population or agreement of test results and cost of the RDT minutes after dropping the chase buffer on the test strip. If no ref-
and the DBS-based ECLIA. erence bar appeared on the test strip, the result was considered in-
According to our literature review, estimating seroprevalence valid, and the sample was retested. Participants who tested HBsAg-
accurately and the cost of the survey are key elements to con- positive on the RDT were advised to have another test at a health
sider in the design of a serosurvey. This study, therefore, aimed facility to confirm their infection status.
to determine the level of agreement between an RDT and a DBS- Concurrently (i.e., without additional finger prick), capillary
based ECLIA and to compare the cost of using an RDT and a DBS- blood was collected onto filter cards of DBS (Whatman 903TM ,
based ECLIA in a field-based serosurvey, which may be useful as a Protein Saver Card; Whatman, Maidstone, Kent, UK) for testing

52
S. Okawa, K. Komada, Y. Ichimura et al. International Journal of Infectious Diseases 125 (2022) 51–57

Figure 1. Framework for estimating serosurvey cost per input category and study phase
Note: The framework was created based on the original framework developed by Carcelen, et al. (2020).

HBsAg. The DBS samples were airdried for at least 60 minutes, preparation, the fieldworkers and supervisors from Ninh Thuan,
sealed in a plastic bag with silica gel, and transported to the Pas- Binh Dinh, and Quảng Ngai received a salary for 2 days (1 day for
teur Institute in Nha Trang (PINT) at ambient temperature within traveling and another day for training), and those from Khanh Hoa
4 days from the sample collection, and stored at -20°C for 15 received a salary for 1 day (1 day for training, with day travel).
months. The DBS samples were extracted from four punched disks For the data collection, all fieldworkers and supervisors across the
of 4 mm in diameter on a filter paper and eluted with 0.6 ml four sites received a salary for 8 days. The survey had two princi-
of phosphate-buffered saline (pH 7.4) at 4°C overnight. The su- pal investigators, but they did not receive a salary from the survey
pernatant of the eluted specimen was tested for HBsAg using an budget. Laboratory technicians were not hired, but the cost of the
ECLIA (Roche Elecsys HBsAg II Cobas e 602; Roche Diagnostics In- laboratory work was included in the cost of extraction and testing
ternational Ltd, Rotkreuz, Switzerland) as used in previous studies of biospecimens.
(Flores et al., 2021; Villar et al., 2019). We followed the manual of
the HISCLTM HBsAg reagent, in which the cut-off point for the HB- Transportation
sAg positivity was ≥0.03 IU/ml, which corresponded with that of One vehicle was rented per province, which the supervisors
serum and plasma samples. Participants who tested HBsAg-positive and researchers used for traveling between PINT and the survey
on ECLIA were not informed of the results. venues. The unit cost of the rental vehicle was adjusted based on
the distance from PINT to the survey site, ranging between $75 and
Cost data $120 per day. As the DBS samples were brought back to PINT by
the supervisors using the vehicles, no additional cost was incurred
Cost data were retrieved from payment receipts, which were for transporting the samples. The fieldworkers used public trans-
initially indicated by Japanese or Vietnamese currency. We con- port to travel to the survey districts and traveled to the survey vil-
verted the cost data into US $, according to the currency exchange lages within the district by motorbike. The cost of public transport
rates at the time when payment was made; 110.6 Japanese Yen and motorbike fuel was included in the transportation costs.
per US $1 in February 2019, 107.45 Japanese Yen per US $1 in
April 2020, 23,357.5 Vietnamese Dong per US $1 in May 2019, and Field consumables
23,244.0 Vietnamese Dong per US $1 in March 2020. Before pur- Field consumables included the fees for the ethical approval;
chase, we obtained quotations from three companies (i.e., three items used in training, pretesting, and interviewing participants us-
quotations) and chose the one with the lowest cost. Based on ing the structured questionnaire (e.g., printing cost, stationary, and
the costing framework used in a previous study (Carcelen et al., small gifts to participants); the fees for making lists of selected
2020), we categorized cost according to five types of input (person- households and data entry of questionnaire; and biospecimen col-
nel, transportation, field consumables, laboratory consumables, and lection items (e.g., RDT kit, DBS filter paper, lancets).
capital items/overhead) and four survey phases (survey prepara-
tion, data/biospecimen collection, laboratory testing, coordination) Laboratory consumables
(Figure 1 and the supplementary material). The detailed items in- Laboratory items included consumable materials used in the ex-
cluded in each category are described next. traction of DBS samples and the fee for extracting and testing the
DBS specimens by laboratory technicians, which was charged per
Personnel participant. All laboratory work was performed in the PINT labo-
The personnel working for the survey were a training lecturer, ratory. Because participants were not informed of the ECLIA result,
an interpreter, fieldworkers, supervisors, administrative staff, and a there was no cost incurred for informing participants with positive
field coordinator. The personnel were hired by PINT. For the survey ECLIA results.

53
S. Okawa, K. Komada, Y. Ichimura et al. International Journal of Infectious Diseases 125 (2022) 51–57

Capital items and overhead Table 1

Participant characteristics (N = 2072)
No capital items were purchased for this survey. All laboratory
work was performed using equipment owned by the PINT, so no n %
additional cost was incurred. Overhead costs were paid to PINT, Province
provincial and district health offices, and commune health centers Binh Dinh 494 23.8
to coordinate survey preparation and fieldwork. Khanh Hoa 531 25.6
Ninh Thuan 521 25.1
Quang Ngai 526 25.4
Analysis Sex
Male 803 38.8
Participant characteristics were described using frequencies and Female 1267 61.2
Not specified 2 0.1
percentages. The level of agreement between the RDT and DBS-
Age group (years)
based ECLIA HBsAg results was calculated by cross-tabulation, and 1-4 289 14.0
the interassay agreement was measured using the Cohen kappa co- 5-9 357 17.2
efficient. 10-14 310 15.0
Cost data were compiled using Microsoft Excel (the supplemen- 15-19 230 11.1
20-24 169 8.2
tary material), and a cost-minimization analysis was performed. 25-29 236 11.4
The overall cost of the serosurvey was calculated based on the 30-34 238 11.5
subtotals, stratified by the input category and study phase. The de- 35-39 241 11.6
nominator to calculate the cost per participant was 2072, based on Unknown 2 0.1
the number of participants whose capillary blood samples for both Note: The values of percentages are
the RDT and the DBS were collected. In addition, marginal cost rounded to one decimal place; the sum
per participant, which means the cost of adding one participant, of the percentage does not necessarily
add to 100.0 due to rounding.
was estimated by summing the cost of field consumables, labora-
tory consumables, personnel for data collection, and their transport
cost. Further, we categorized the overall survey cost into the cost
(7.6%), and capital items and overheads (5.3%) (Table 3). Regard-
common to both the RDT and the DBS-based ECLIA, that was spe-
ing the cost per serosurvey phase, data and biospecimen collec-
cific to the RDT, and the DBS-based ECLIA and estimated the cost
tion accounted for the highest cost (61.7%), followed by laboratory
of conducting a serosurvey using an RDT or DBS.
testing (28.7%), coordination (5.3%), and survey preparation (4.2%)
(Figure 2). The common cost of the two testing methods, the costs
Ethics statement specific to the RDT and costs specific to the DBS-based ECLIA, were
$48,388, $4795 and $26,904, respectively. Thus, an HBsAg serosur-
This study received approval from the ethics committees of vey using DBS ($75,292 in total, $36.3 per participant) would cost
PINT (approval number was not issued) and the National Center 1.4 times as much as a serosurvey of the same design using an
for Global Health and Medicine, Japan (NCGM-3091). During the RDT ($53,183 in total, $25.7 per participant). Furthermore, focusing
survey, the fieldworkers explained the study objectives and proce- on the test-specific costs, the DBS-based ECLIA (a total of $26,904)
dures to each participant and obtained written informed consent was 5.6 times more expensive than the RDT (a total of $4795).
from them or parents of participating children aged <18 years. Re-
searchers from PINT supervised the informed consent, interview,
and blood sample collection. Discussion

We conducted a field-based survey of HBsAg seroprevalence

Results
with 2072 participants using an RDT and a DBS-based ECLIA. The
test results of the RDT and the DBS-based ECLIA showed almost
Basic characteristics of study participants
perfect agreement. The total cost of the survey was $80,086, which
was equal to $39 per participant. The total cost of conducting a
Of the 2093 individuals recruited, 21 were excluded from the
serosurvey using a DBS-based ECLIA was 1.4 times higher than us-
analysis: one declined to participate in the study, 10 were outside
ing an RDT. To the best of our knowledge, this is the first study to
the age limits, five had invalid results for the RDT, and five had
compare two aspects, test result and cost, between an RDT and a
insufficient blood collected for DBS testing. A total of 2072 partici-
DBS-based ECLIA in a field-based HBsAg seroprevalence survey in
pants were included in the analysis (Table 1).
a resource-limited setting.
The agreement between the RDT and the DBS-based ECLIA HB-
Test results of hepatitis B surface antigen measured using an RDT and sAg test results was almost perfect according to the Cohen kappa
DBS-based immunoassay coefficient. Although we did not compare the results with an EIA
reference standard, high sensitivity and specificity of the two as-
The level of agreement in the HBsAg test results between the say formats for HBsAg testing with a reference standard have been
RDT and the DBS-based ECLIA was 99.0% (Table 2), and the interas- demonstrated (Amini et al., 2017; Lange et al., 2017). This study
say agreement by the Cohen kappa coefficient was 0.90, which in- demonstrates that an RDT and a DBS-based ECLIA used in field-
dicates almost perfect agreement (P <0.001). based HBsAg in a resource-limited setting can provide test results
in high agreement.
Cost of the hepatitis B surface antigen serosurvey The total cost of serosurvey using the DBS-based ECLIA was 1.4
times higher than that using the RDT. When the cost specific to
The overall cost of the HBsAg serosurvey was $80,086, which testing was compared, the cost of the DBS-based ECLIA was more
was equivalent to $38.7 per participant. The estimated marginal than five times higher than that of the RDT. The smaller differ-
cost was $33.4. Regarding the cost by input category, field con- ence in the total cost of serosurvey between the two methods is
sumables accounted for the highest proportion (36.2%), followed by attributable to the common cost spent in a survey using an RDT
laboratory consumables (28.7%), personnel (22.1%), transportation and a survey using a DBS-based ECLIA, such as personnel and

54
S. Okawa, K. Komada, Y. Ichimura et al. International Journal of Infectious Diseases 125 (2022) 51–57

Table 2
Hepatitis B surface antigen test results using a rapid diagnostic test and dried blood spot-based immunoas-
say among participants with results for both tests (N = 2072)

Dried blood spot-based electrochemiluminescence immunoassay

Positive Negative

n (%) n (%)

Rapid diagnostic test Positive 106 (5.1) 6 (0.3)

Negative 15 (0.7) 1945 (93.9)

Figure 2. Cost of the hepatitis B surface antigen serosurvey according to the phase and test type
Abbreviations: DBS, dried blood spot; RDT, rapid diagnostic test; USD, US $.

Table 3 Table 4
Cost of the hepatitis B surface antigen serosurvey by Comparison of advantages between RDT vs DBS.
input category
RDT vs DBS
Cost (US $) % √ √
Low invasiveness to study participants
√
Personnela 17,695 22.1 Lower serosurvey cost
√
Transportationb 6120 7.6 Rapidity in obtaining test result
√
Field consumablesc 28,977 36.2 Consensus on the standard protocola
√
Laboratory consumablesd 23,024 28.7 No necessity for laboratory resource
√
Capital items/overheade 4270 5.3 Measurement of multiple diseases or biomarkers
√
Total 80,086 100.0 Quantitative measurement
√
Abbreviations: DBS, dried blood spot. Key: indicates the advantages related to the tests
Note: Abbreviations: DBS, dried blood spot; RDT, rapid diagnostic test.
a a
Salary for training lecturer, interpreter, fieldwork- Reference: World Health Organization guidelines on hepatitis B and C
ers, supervisors, administrative staff, and field coordi- testing. (WHO, 2017d)
nator.
b
Cost of rental vehicle, public transportation, and
motorbike fuel.
c In choosing whether to use an RDT or a DBS-based ECLIA for
Cost of ethical approval, items used in training,
pretesting, interviewing participants using the struc- serosurveys, researchers should consider various aspects other than
tured questionnaire, the fee for making lists of se- the performance of the assay format and the cost (Table 4). Focus-
lected households, data entry of questionnaire, and ing on cost alone suggests that using an RDT would be preferable
biospecimen collection items.
d
to using a DBS-based ECLIA. However, other advantages and disad-
Cost of consumable materials used in the extrac-
tion of DBS samples, and technical work including ex- vantages of RDTs and the DBS-based ECLIA need to be considered.
tracting and testing the DBS specimens. First, RDTs, particularly the Alere Determine HBsAg kit used in this
e
Cost of coordination of survey preparation and study, have been widely used as an alternative to EIAs. This al-
fieldwork by Pasteur Institute in Nha Trang, provincial lows monitoring over-time change in seroprevalence and compar-
and district health offices, and commune health cen-
ing seroprevalence in different settings. In contrast, the DBS-based
ters. No capital goods were purchased.
ECLIA has no established standards for the type of filter papers,
minimum performance criteria of DBS specimens, storage condi-
tions, or validation of existing commercial assays with DBS spec-
transportation in the survey preparation and data collection. In imens (WHO, 2017d). Thus, RDTs have been a recommended op-
contrast, the larger difference in the cost specific to the testing tion over DBS. Second, RDTs can identify HBsAg status without a
was as we had expected because the DBS-based ECLIA requires the laboratory, and the results can be returned to the participant on-
cost of extracting and testing samples by a laboratory technician, site, which helps to efficiently link participants who test HBsAg-
whereas the kit constitutes the majority of the RDT cost. positive to treatment programs. In contrast, the DBS-based ECLIA

55
S. Okawa, K. Komada, Y. Ichimura et al. International Journal of Infectious Diseases 125 (2022) 51–57

has several other advantages. First, it can test samples for multiple DBS-based ECLIA may be a useful option if the serosurvey aims to
infectious diseases (e.g., vaccine-preventable diseases, e.g., measles evaluate multiple infectious diseases, particularly in a setting with
or rubella) (Breakwell et al., 2020; Hachiya et al., 2018; Hayford high endemicity of vaccine-preventable diseases.
et al., 2019), biomarkers (e.g., total cholesterol) (Lakshmy et al.,
2010), or exposures to environmental toxicants (Batterman and Declaration of Competing Interest
Chernyak, 2014) in a single survey, whereas RDT kits that have
been prequalified by WHO are available for HIV, HBV and hepatitis The authors have no competing interests to declare.
C virus, malaria, and syphilis only as of 2021 (WHO, 2021b). Sec-
ond, DBS samples are stable under storage conditions at or below
Funding
25°C for at least 5 months (Mayer et al., 2012), and DBS-based as-
says can measure quantitative values of HBsAg and viral load and
This work was supported by the Health and Labour Sciences
analyze the genotype of HBV (Jardi et al., 2004; Mohamed et al.,
Research Grant (grant number JPMH18BA1003), Grants for the
2013). The detailed advantages and disadvantages of RDTs and
National Center for Global Health and Medicine (grant number
the DBS-based ECLIA are described fully elsewhere (WHO, 2011,
19A01, 30-2009, 22A01), and Japan Society for the Promotion of
2017b). Thus, the choice of an RDT or a DBS-based assay should
Science KAKENHI (grant number JP19K10656). However, the fund-
be made depending on the overall study objective.
ing source was not involved in design of the study, collection, anal-
We used an existing framework developed by Carcelen et al.
ysis, and interpretation of data or writing of the manuscript.
(2020) to calculate the cost of the serosurvey. Although the frame-
work was initially developed to ensure comparability with other
Acknowledgments
studies, we found it difficult to simply compare our results with
their results because we were unable to fully synchronize our cost
The authors thank the study participants for understanding the
breakdown to that of the study by Carcelen et al. For example,
significance of the serosurvey and offering their biospecimens. We
our study combined the cost of personnel for the data collec-
also acknowledge the fieldworkers, supervisors, and researchers
tion and biospecimen collection because the same fieldworkers en-
from the Pasteur Institute in Nha Trang and the staff from the
gaged in these activities. In the study by Carcelen et al., these costs
provincial and district Department of Health. The authors also
were calculated separately. Furthermore, our study costed labo-
thank Editage for proofreading the manuscript.
ratory consumables per participant. In contrast, in the study by
Carcelen et al., laboratory technicians’ salaries were costed sepa-
rately. Despite these challenges, estimating cost using a common CRediT authorship contribution statement
framework would be beneficial if the definition of cost items was
described in detail and the results were carefully compared. SO, KK, YI, MS, HTD, HXL, TTH, TBN, and MH conceived the
study design. KK, YI, HTD, HXL, TTH, TBN, MKH, and MH collected
Limitations data. HXL, TBN, MKH, HTHH, NATT, THL, and QTN performed the
laboratory analysis. SO, KK, and MH conducted data analysis and
This study has several limitations. First, we did not use an EIA drafted the manuscript. All authors critically reviewed, edited, and
reference standard because we collected blood samples from a rep- agreed with the final draft for submission.
resentative sample of the population in a resource-limited field
setting, so we could not assess the sensitivity and specificity of Supplementary materials
the RDT and the DBS-based ECLIA used in this study. However, the
high sensitivity and specificity of HBsAg RDT and DBS-based HBsAg Supplementary material associated with this article can be
testing compared to laboratory-based EIAs and venous blood-based found, in the online version, at doi:10.1016/j.ijid.2022.10.011.
testing have been previously demonstrated. Thus, we believe that
the seroprevalence estimated by the two methods were equally References
and sufficiently accurate. Second, the standard cost for personnel
Allison RD, Teleb N, Al Awaidy S, Ashmony H, Alexander JP, Patel MK. Hepatitis
and transportation varies by country or region, which limits the
B control among children in the Eastern Mediterranean Region of the World
generalizability of the study findings. Third, our cost estimation Health Organization. Vaccine 2016;34:2403–9.
did not account for potential changes in cost related to time and Amini A, Varsaneux O, Kelly H, Tang W, Chen W, Boeras DI, et al. Diagnostic accu-
transport for reaching the survey venues and participants, which racy of tests to detect hepatitis B surface antigen: a systematic review of the
literature and meta-analysis. BMC Infect Dis 2017;17:698.
could be affected by season or variability in the participation rate Avellon A, Ala A, Diaz A, Domingo D, Gonzalez R, Hidalgo L, et al. Clinical perfor-
across study sites. Fourth, the difference in the cost of serosurvey mance of Determine HBsAg 2 rapid test for hepatitis B detection. J Med Virol
between the use of RDT and DBS-based ECLIA for multiple diseases 2020;92:3403–11.
Batterman S, Chernyak S. Performance and storage integrity of dried blood spots for
may not be the same as that for a single disease, as shown in PCB, BFR and pesticide measurements. Sci Total Environ 2014;494–495:252–60.
this study. Despite these limitations, this study showed that using Breakwell L, Tevi-Benissan C, Childs L, Mihigo R, Tohme R. The status of hepatitis B
an RDT or a DBS-based ECLIA are both feasible options in field- control in the African region. Pan Afr Med J 2017;27:17.
Breakwell L, Anga J, Cooley G, Ropiti L, Gwyn S, Wannemuehler K, et al. Seropreva-
based serosurveys of HBsAg and provided an approximate cost of lence of chronic hepatitis B virus infection and immunity to measles, rubella,
conducting a serosurvey in low- or middle-income countries. Our tetanus and diphtheria among schoolchildren aged 6-7 years old in the Solomon
finding, together with the result of HBsAg seroprevalence (Komada Islands, 2016. Vaccine 2020;38:4679–86.
Carcelen AC, Hayford K, Moss WJ, Book C, Thuma PE, Mwansa FD, et al. How much
et al., 2020), will help design a population-based serosurvey.
does it cost to measure immunity? A costing analysis of a measles and rubella
serosurvey in southern Zambia. PLoS One 2020;15.
Conclusion Chan PL, Le LV, Ishikawa N, Easterbrook P. Regional progress towards hepati-
tis C elimination in the western Pacific Region, 2015-2020. Glob Health Med
2021;3:253–61.
The RDT and the DBS-based ECLIA showed an almost perfect Duong MC, Nguyen VTT, Otsu S, McLaws ML. Prevalence of hepatitis B and C virus
agreement with the HBsAg test results. The cost of conducting a infections in hemodialysis patients in Vietnam: a systematic review and meta–
serosurvey using a DBS-based ECLIA was estimated to be 1.4 times analysis. JGH Open 2020;4:29–38.
Flores GL, Barbosa JR, Cruz HM, Miguel JC, Potsch DV, Pilotto JH, et al. Dried blood
higher than that of a serosurvey using an RDT. Thus, the use of RDT spot sampling as an alternative for the improvement of hepatitis B and C diag-
is a better option for conducting a serosurvey of HBsAg. In contrast, nosis in key populations. World J Hepatol 2021;13:504–14.

56
S. Okawa, K. Komada, Y. Ichimura et al. International Journal of Infectious Diseases 125 (2022) 51–57

General Statistics Office, United Nations Population Fund. The population sex-age Nguyen TH, Vu MH, Nguyen VC, Nguyen LH, Toda K, Nguyen TN, et al. A reduc-
structure and related socio-economic issues in Viet Nam: evidence from the tion in chronic hepatitis B virus infection prevalence among children in Vietnam
2014 intercensal population and housing survey. Hanoi: Vietnam News Agency demonstrates the importance of vaccination. Vaccine 2014;32:217–22.
Publishing House; 2016. Nguyen VT, Law MG, Dore GJ. An enormous hepatitis B virus-related liver disease
Hachiya M, Miyano S, Mori Y, Vynnycky E, Keungsaneth P, Vongphrachanh P, et al. burden projected in Vietnam by 2025. Liver Int 2008;28:525–31.
Evaluation of nationwide supplementary immunization in Lao People’s Demo- Schweitzer A, Horn J, Mikolajczyk RT, Krause G, Ott JJ. Estimations of worldwide
cratic Republic: population-based seroprevalence survey of anti-measles and prevalence of chronic hepatitis B virus infection: a systematic review of data
anti-rubella IgG in children and adults, mathematical modelling and a stabil- published between 1965 and 2013. Lancet 2015;386:1546–55.
ity testing of the vaccine. PLoS One 2018;13. Viet L, Lan NT, Ty PX, Björkvoll B, Hoel H, Gutteberg T, et al. Prevalence of hepati-
Hayford K, Mutembo S, Carcelen A, Matakala HK, Munachoonga P, Winter A, et al. tis B & hepatitis C virus infections in potential blood donors in rural Vietnam.
Measles and rubella serosurvey identifies rubella immunity gap in young adults Indian J Med Res 2012;136:74–81.
of childbearing age in Zambia: the added value of nesting a serological survey Villar LM, Cruz HM, Deodato RM, Miguel JC, da Silva EF, Flores GL, et al. Usefulness
within a post-campaign coverage evaluation survey. Vaccine 2019;37:2387–93. of automated assays for detecting hepatitis B and C markers in dried blood spot
Ishizaki A, Tran VT, Nguyen CH, Tanimoto T, Hoang HTT, Pham HV, et al. Discrep- samples. BMC Res Notes 2019;12:523.
ancies in prevalence trends for HIV, hepatitis B virus, and hepatitis C virus in Woodring J, Pastore R, Brink A, Ishikawa N, Takashima Y, Tohme RA. Progress toward
Haiphong, Vietnam from 2007 to 2012. PLoS One 2017;12. hepatitis B control and elimination of mother-to-child transmission of hepati-
Jardi R, Rodriguez-Frias F, Buti M, Schaper M, Valdes A, Martinez M, et al. Useful- tis B virus - western Pacific region, 2005–2017. MMWR Morb Mortal Wkly Rep
ness of dried blood samples for quantification and molecular characterization 2019;68:195–200.
of HBV-DNA. Hepatology 2004;40:133–9. World Health Organization. Documenting the Impact of Hepatitis B Immunization:
Komada K, Hachiya M, Ichimura Y, Shimada M. The impact of hepatitis B vacci- best practices for conducting a serosurvey. Geneva: World Health Organization;
nation program in central Vietnam; evaluation by population-based, cross-sec- 2011.
tional seroprevalence survey of HBsAg. Int J Infect Dis 2020;101:477. World Health Organization. Regional Office for the Western Pacific. Regional action
Lakshmy R, Gupta R, Prabhakaran D, Snehi U, Reddy KS. Utility of dried blood spots plan for viral hepatitis in the Western Pacific 2016-2020: a priority action plan
for measurement of cholesterol and triglycerides in a surveillance study. J Dia- for awareness, surveillance, prevention and treatment of viral hepatitis in the
betes Sci Technol 2010;4:258–62. Western Pacific Region. Geneva: World Health Organization; 2016.
Lam ATN, Yamarat K, Chapman RS. HBsAg prevalence and knowledge on hepati- World Health Organization. Global hepatitis report 2017. Geneva: World Health Or-
tis B among Stieng tribe adults in Binh Phuoc Province. Vietnam. J Health Res ganization, 2017a.
2013;27:123–8. World Health Organization. Regional Office for Africa. End hepatitis by 2030: pre-
Lange B, Cohn J, Roberts T, Camp J, Chauffour J, Gummadi N, et al. Diagnostic accu- vention, care and treatment of viral hepatitis in the African Region: framework
racy of serological diagnosis of hepatitis C and B using dried blood spot samples for action, 2016–2020. Geneva: World Health Organization; 2017b.
(DBS): two systematic reviews and meta-analyses. BMC Infect Dis 2017;17:700. World Health Organization. Regional Office for the Eastern Mediterranean. Regional
MacNeil A, Lee CW, Dietz V. Issues and considerations in the use of serologic action plan for the implementation of the global health sector strategy on viral
biomarkers for classifying vaccination history in household surveys. Vaccine hepatitis 2017-2021. Geneva: World Health Organization; 2017c.
2014;32:4893–900. World Health Organization. WHO guidelines on hepatitis B and C testing. Geneva:
Mayer TK, Vargas RL, Knebel AE, Williams SA, Culver SP, Clark DM, et al. Hepatitis World Health Organization; 2017d.
B assays in serum, plasma and whole blood on filter paper. BMC Clin Pathol World Health Organization. Regional Office for the Western Pacific. Regional frame-
2012;12:8. work for the triple elimination of mother-to-child transmission of HIV, hepatitis
Metcalf CJE, Farrar J, Cutts FT, Basta NE, Graham AL, Lessler J, et al. Use of sero- B and syphilis in Asia and the pacific, 2018-2030. Geneva: World Health Orga-
logical surveys to generate key insights into the changing global landscape of nization; 2018.
infectious disease. Lancet 2016;388:728–30. World Health Organization. Prevention of mother-to-child transmission of hepatitis
Mohamed S, Raimondo A, Pénaranda G, Camus C, Ouzan D, Ravet S, et al. Dried B virus: guidelines on antiviral prophylaxis in pregnancy. Geneva: World Health
blood spot sampling for hepatitis B virus serology and molecular testing. PLoS Organization; 2020.
One 2013;8:e61077. World Health Organization. Global progress report on HIV, viral hepatitis and sex-
Namgyal P. Impact of hepatitis B immunization, Europe and worldwide. J Hepatol ually transmitted infections, 2021. Accountability for the global health sector
2003;39:S77–82. strategies 2016-2021: actions for impact. Geneva: World Health Organization;
Ngoc Thanh P, Thi Thi Tho N, Dac Phu T, Dai Quang T, Thuy Duong N, Chinh 2021a.
Chien V, Trong Lan P. Prevalence and factors associated with chronic hepati- World Health Organization. The List of WHO prequaified in vitro Diagnostic prod-
tis B infection among adults in the Central Highland, Vietnam. AIMS Med Sci ucts. Geneva: World Health Organization; 2021b.
2020;7:337–46.

57