Open access Original research
HIV testing and counselling among 
women in Benin: a cross-s ectional 
analysis of prevalence and predictors 
from demographic and health 
survey data
Collins Adu   ,1,2 Leticia Akua Adzigbli,3 Abdul Cadri,4,5 Paa Akonor Yeboah,3 
Aliu Mohammed,6 Richard Gyan Aboagye   7
To cite: Adu C, Adzigbli LA, ABSTRACT
Cadri A, et al.  HIV testing and Objective To examine the uptake of HIV testing and STRENGTHS AND LIMITATIONS OF THIS STUDY
counselling among women in counselling (HTC) and its associated factors among ⇒ The use of nationally representative data and the 
Benin: a cross-s ectional analysis women in Benin. use of rigorous statistical analysis to estimate the 
of prevalence and predictors 
Design We performed a cross- sectional analysis of data prevalence and predictors of HTC uptake ensures 
from demographic and health 
from the 2017–2018 Benin Demographic and Health the generalisability of the findings.survey data. BMJ Open 
2023;13:e068805. doi:10.1136/ Survey. A weighted sample of 5517 women was included ⇒ Findings in this study aid to bridge literature gap on 
bmjopen-2022-068805 in the study. We used percentages to present the results prevalence and predictors of HIV testing and coun-
of the uptake of HTC. Multilevel binary logistic regression selling among women in Benin.
► Prepublication history and analysis was used to examine the predictors of HTC ⇒ This study used secondary data and the analysis 
additional supplemental material 
uptake. The results were presented using adjusted odds was limited to only the variables present in the data-
for this paper are available 
ratios (aORs), with 95% confidence intervals (CIs). set. Hence, the interpretation of the findings and any online. To view these files, 
please visit the journal online Setting Benin. inferences made should be based on the variables 
(http://dx.doi.org/10.1136/ Participants Women aged 15–49. used.
bmjopen-2022-068805). Outcome measure Uptake of HTC. ⇒ Demographic Health Survey employs a cross- 
Results The overall uptake of HTC among women in sectional design and this limits the study’s ability to 
Received 30 September 2022 Benin was found to be 46.4% (44.4%–48.4%). The odds of draw causal inferences.
Accepted 03 February 2023 HTC uptake was higher among women covered by health ⇒ The study’s variables were assessed based on the 
insurance (aOR 3.04, 95% CI 1.44 to 6.43) and those with women’s self- reports, which raised the possibility of 
comprehensive HIV knowledge (aOR 1.77, 95% CI 1.43 to recall bias and other social desirability biases.
2.21). The odds of HTC uptake increased with increasing 
level of education, with the highest odds among those in 
the secondary or higher level (aOR 2.06, 95% CI 1.64 to with HIV (PLHIV) are aware of their status 
2.61). Also, the age of the women, mass media exposure, and nearly 19 million people do not know of 
region of residence, high community literacy level, and their serostatus, even though this figure has 
high community socioeconomic status were associated dropped to about 7.1 million in 2019.1 The 
with higher odds of HTC uptake. Women residing in rural 
areas were less likely to use HTC. Religious affiliation, HIV burden in sub-S aharan Africa (SSA) 
number of sexual partners, and place of residence were is high, with West and Central Africa being 
associated with lower odds of HTC uptake. the most afflicted regions.
1 There is a high 
Conclusion Our study has shown that the uptake of prevalence of PLHIV in SSA, with women 
HTC among women in Benin is relatively low. There is a being disproportionately affected.1–3 In 2019, 
need to enhance efforts to empower women, as well as 4.9 million people in West and Central Africa 
© Author(s) (or their reduce health inequities as they all have a substantial were living with HIV, 240, 000 were newly 
employer(s)) 2023. Re- use impact on HTC uptake among women in Benin, taking into infected, and 140, 000 died of AIDS-r elated 
permitted under CC BY- NC. No consideration the factors identified in this study.
commercial re- use. See rights illnesses.
3
and permissions. Published by HIV testing and counselling (HTC) is a key 
BMJ. public health intervention that entails coun-
For numbered affiliations see INTRODUCTION selling, testing, and treatment to lower HIV/
end of article. The global burden of HIV/AIDS currently AIDS transmission and its associated health 
Correspondence to stands at about 38 million with approximately burden.
4 The relevance of voluntary coun-
Richard Gyan Aboagye;  36.2 million being adults and 1.8 million chil- selling and testing (VCT) sessions cannot be 
1
a boagyegyan94@ gmail.c om dren.  Also, an estimated 81% of people living overstated. VCT has long been a significant 
Adu C, et al. BMJ Open 2023;13:e068805. doi:10.1136/bmjopen-2022-068805 1
BMJ Open: first published as 10.1136/bmjopen-2022-068805 on 13 April 2023. Downloaded from http://bmjopen.bmj.com/ on May 5, 2023 by guest. Protected by copyright.
Open access 
policy strategy for controlling HIV transmission.5 HIV elsewhere.16 17 Pretested and structured questionnaires 
testing is an important first step in HIV prevention, care, were used to collect data from the respondents on several 
and treatment because it allows for early diagnosis and health and demographic characteristics, including 
treatment, which helps to prevent new infections in HTC uptake.16 17 We included a weighted sample of 
women.2 5,517 women of reproductive age (15–49 years) in the 
According to data from the Joint United Nations study. This sample consisted of women with complete 
Programme on HIV/AIDS (UNAIDS), 68% of PLHIV in observations on all the variables of interest included in 
West and Central Africa knew their serological status as the study. We relied on the Strengthening the Reporting 
of 2019.3 The UNAIDS announced an ambitious 95- 95-9 5 of Observational Studies in Epidemiology guidelines to 
target for ending the HIV epidemic by 2030, with the goal draft this paper (online supplemental table 1).18 The 
of having 95% of PLHIV know their status, 95% of those dataset is freely accessible to download at https://dhspro-
diagnosed with HIV infection receive sustained antiret- gram.com/data/dataset/Benin_Standard-DHS_2017. 
roviral therapy, and 95% of those on treatment have cfm?flag=1.19
suppressed viral loads by 2030.6
In SSA, HTC uptake remains low, especially among Study variables
women.7–9 Meanwhile, adolescent girls and women of HTC uptake is the outcome variable in the study. To assess 
reproductive age account for over 60% of new HIV infec- this variable, the women were asked the question ‘Have 
tions occurring each day globally.7 10 Thus, to achieve you ever tested for HIV?’. The response options were ‘no’ 
the global target of ending the HIV epidemic by 2030, and ‘yes’. We coded those who responded ‘no’ as ‘0’ and 
countries in SSA need to implement strategies that could ‘yes’ as ‘1’ and used it in the analysis.20–22
enhance HIV testing uptake among the populace, espe- A total of 17 explanatory variables were considered for 
cially women. Although Benin has implemented some this study. The variables were selected not based on parsi-
strategies to enhance HIV testing uptake, such as a manda- mony but based on their significant association with HTC 
tory testing and counselling for all pregnant women uptake from the literature14 21–23 as well as their avail-
during antenatal visits,11 the level of HIV testing uptake ability in the DHS dataset. We grouped the variables into 
remains low. As of 2019, 75,000 people in Benin were individual level and contextual level. The individual-l evel 
living with HIV, with 3,500 new infections.1 Meanwhile, variables consisted of women’s age, level of education, 
HIV testing is the initial step in the HIV care process and marital status, current working status, religion, ethnicity, 
the gateway to treatment and support.11 health insurance coverage, presence of sexually trans-
Age, gender, marital status, educational status, socio- mitted infections in the last 12 months preceding the 
economic position, area of residence and wealth index survey, condom use during the last sex with most recent 
have been identified as independent predictors of HIV partner, comprehensive HIV knowledge, number of 
testing among women in Ethiopia and other regions of sexual partners excluding spouse in the last 12 months 
Africa.12–14 Previous research in Benin focused on the prior to the survey, wealth index, and exposure to mass 
HIV prevention and treatment pathways among female media. Place of residence, region, community literacy 
sex workers.11 To date, no current study has looked at the level, and community socioeconomic status were the 
prevalence and determinants of HTC uptake in Benin contextual- level variables included in the study. Detailed 
using the current nationally representative data. Find- description of the explanatory variables have been high-
ings from this study could provide useful information lighted in the literature.20–22
to aid policy- makers in the design and implementation 
of targeted strategies and interventions, as well as guide Statistical analyses
public health professionals and policymakers on how to We used Stata software V.17.0 to perform all the anal-
distribute resources fairly to improve HTC uptake among yses. Percentages were used to present the results of the 
women in Benin. In this study, we examined the preva- uptake of HTC and its distribution across the explanatory 
lence and correlates of HTC uptake among women in variables. We employed binary logistic regression anal-
Benin. ysis to select the significant variables for the multilevel 
regression analysis. All the variables with p<0.05 were 
included in the multilevel model. Later, we employed a 
METHODS multivariable multilevel binary logistic regression anal-
Data source and study design ysis to examine the predictors of HTC uptake among 
We performed a secondary data analysis of the 2017–2018 the women in Benin using four models (models O–III). 
Benin Demographic and Health Survey (DHS). The Model O consisted of only the outcome variable, with no 
DHS is a nationally representative survey conducted in explanatory variable with the results indicating the vari-
over 85 countries globally.15 DHS used a cross-s ectional ation in HTC uptake attributed to the clustering at the 
design to collect data from the respondents: women, primary units. Models I and II consisted of the individual- 
men and children. A two-s tage cluster sampling method level and contextual-l evel variables, respectively. Model 
was adopted by DHS to recruit respondents for the III contained all the explanatory variables. The results 
survey. Detailed sampling technique has been published were presented using adjusted odds ratios (aORs), with 
2 Adu C, et al. BMJ Open 2023;13:e068805. doi:10.1136/bmjopen-2022-068805
BMJ Open: first published as 10.1136/bmjopen-2022-068805 on 13 April 2023. Downloaded from http://bmjopen.bmj.com/ on May 5, 2023 by guest. Protected by copyright.
Open access
their respective 95% CIs. The Akaike’s information crite- 1.63, 95% CI 1.07 to 2.51). Women covered by health 
rion (AIC) was used to evaluate model fitness, or how well insurance were more likely to test for HIV compared to 
various models fitted the data, with the best- fitted model those not covered by health insurance (aOR 3.04, 95% CI 
being the model with the smallest AIC values. Model III 1.44 to 6.43). Relative to women with no HIV knowledge, 
was chosen as the best-f itted model since it has the least those with comprehensive HIV knowledge had higher 
AIC value. We performed the multilevel regression using odds of testing for HIV (aOR 1.77, 95% CI 1.43 to 2.21). 
Stata’s ‘mlogit’ function. We also weighted all the analyses Women aged 20–49 years were more likely to test for HIV 
and the Stata’s ‘svyset’ command was used to adjust for as compared to those aged 15–19 years. The odds of HTC 
disproportionate sampling and non- response. uptake was higher among women from communities 
with high literacy level (aOR 1.76, 95% CI 1.28 to 2.41), 
Patient and public involvement those from communities with high socioeconomic status 
In this study, patients and the public were not included in (aOR 1.43, 95% CI 1.08 to 1.90), and those belonging to 
the study’s design and conduct. the Adja ethnic group (aOR 1.84, 95% CI 1.01 to 3.35). 
Women from the Atlantic, Littoral, Couffo, Mono, and 
Zou regions were more likely to test for HIV compared to 
RESULTS those in the Alibori region.
Background characteristics of the respondents However, women belonging to the African Traditional 
Table 1 presents the background characteristics of religion (aOR 0.73, 95% CI 0.54 to 0.98) were less likely 
the respondents included in the study. Majority of the to test for HIV compared to their Christian counterparts. 
respondents were aged 25–29 (22.6%), attained no Increasing the number of sexual partners was associ-
formal education (57.4%), were married (62.9%), were ated with a reduction in the odds of HTC uptake with 
working (81.6%) and belonged to the Fon ethnic group the lowest odds among those with two or more partners 
(37.4%). Most of the respondents were not covered by (aOR 0.35, 95% CI 0.18 to 0.68). Women residing in rural 
health insurance (98.9%), had no STIs in the last 12 areas (aOR 0.73, 95% CI 0.57 to 0.92) were less likely to 
months preceding the survey (92.1%), belonged to the undergo HTC compared to those in urban areas.
richest wealth quintile (22.8%), and resided in rural areas 
(56.9%).
DISCUSSION
Prevalence of HTC uptake among women in Benin This study sought to examine the prevalence and 
The overall prevalence of HTC uptake among women correlates of HTC uptake among women in Benin. The 
in Benin was found to be 46.4%. Across the various age study found the prevalence of HTC uptake to be 46.4% 
range, women in the age range 30–34 years reported among women in Benin, which is relatively low. The 
the highest prevalence (56.6%) of HTC uptake whereas prevalence found in this study is lower than the prev-
women in the age range 15–19 years reported the least alence found in Ethiopia (53%),24 Zambia (80%),25 
prevalence (21.8%). Also, women with secondary educa- Nigeria (58.4%),26 and Uganda (75.6%).26 Since HTC is 
tion or higher reported a higher prevalence (60.7%) of an important strategy to ending the HIV epidemic, this 
HTC uptake than those with primary (51.2%) and no finding suggests the need to improve the uptake of HTC 
education (39%). Also, the prevalence of HTC uptake among women in Benin.
was higher in women with comprehensive knowledge We found that women who had health insurance were 
of HIV (60.6%) than in women without comprehensive more likely to utilise HTC services. This result corrobo-
knowledge of HIV. The prevalence of HTC uptake was rates with the findings from previous studies conducted 
also found to be higher in women who were covered by in Malawi27 and Ghana.21 Health insurance coverage has 
health insurance (87.1%) than in women without health been found to be an important determinant of access 
insurance coverage. Similarly, women with a higher expo- to health services in most countries.28–30 This is mainly 
sure to mass media (three) reported a higher prevalence because health insurance subscription offers subscribers 
of HTC uptake (70.5%). The results are presented in some level of protection against health expenditure,31 
table 1. thus subscribers do not pay for certain health services. 
This could be the case in the context of HTC in Benin, 
Predictors of HTC uptake among women in Benin as women who subscribed to health insurance might have 
Table 2, model III, presents the results of the predictors had free access to HTC services.
of HTC among women in Benin. The results showed that Also, we found that women with higher educational 
the odds of HTC uptake increases with an increasing level attainment and women with comprehensive knowledge 
of education, with the highest odds among those with of HIV had an increased likelihood of HTC uptake. This 
secondary or higher education (aOR 2.06, 95% CI 1.64 is similar to findings from other studies in Zambia25 and 
to 2.61). Similarly, increasing the number of mass media Ghana.32 Attaining a higher educational level has the 
exposure was associated with higher likelihood of HTC potential to impact women’s level of knowledge on HIV 
uptake, with those exposed to three mass media having which could have eventually improve their HIV- related 
higher odds relative to those with no exposure (aOR behaviour especially in the context of HTC uptake.33–35 
Adu C, et al. BMJ Open 2023;13:e068805. doi:10.1136/bmjopen-2022-068805 3
BMJ Open: first published as 10.1136/bmjopen-2022-068805 on 13 April 2023. Downloaded from http://bmjopen.bmj.com/ on May 5, 2023 by guest. Protected by copyright.
Open access 
Table 1 Bivariate analysis of HIV testing and counselling (HTC) across the explanatory variables
HTC
Variable Weighted N (%) Yes (%) No (%) cOR (95% CI)
Prevalence 46.4 (44.4–48.4) 43.6 (51.6–55.6)
Women’s age (years)
 1 5–19 620 (11.2) 135 (21.8) 485 (78.2) 1.00
  20–24 1090 (19.8) 476 (43.7) 614 (56.3) 2.78*** (2.18 to 3.55)
  25–29 1247 (22.6) 652 (52.3) 595 (47.7) 3.93*** (3.04 to 5.06)
  30–34 862 (15.6) 487 (56.6) 375 (43.4) 4.66*** (3.57 to 6.10)
  35–39 772 (14.0) 412 (53.4) 360 (46.6) 4.10*** (3.18 to 5.30)
  40–44 506 (9.2) 240 (47.4) 266 (52.6) 3.23*** (2.48 to 4.22)
  45–49 420 (7.6) 156 (37.1) 264 (62.9) 2.12*** (1.57 to 2.86)
Level of education
  No formal education 3169 (57.4) 1234 (39.0) 1935 (61.0) 1.00
  Primary 1066 (19.3) 546 (51.2) 520 (48.8) 1.65*** (1.38 to 1.96)
 S econdary or higher 1282 (23.3) 778 (60.7) 504 (39.3) 2.42*** (2.05 to 2.85)
Marital status
  Never married 773 (14.0) 223 (28.8) 550 (71.2) 1.00
 M arried 3469 (62.9) 1653 (47.7) 1816 (52.3) 2.25*** (1.85 to 2.74)
  Cohabiting 1057 (19.1) 567 (53.7) 490 (46.3) 2.86*** (2.19 to 3.74)
  Previously married 218 (4.0) 116 (53.4) 102 (46.6) 2.84*** (2.02 to 3.98)
Current working status
  No 1015 (18.4) 426 (41.9) 589 (58.1) 1.00
  Yes 4502 (81.6) 2134 (47.4) 2368 (52.6) 1.25* (1.05 to 1.49)
Religion
  Christianity 3040 (55.1) 1610 (53.0) 1430 (47.0) 1.00
 I slamic 1630 (29.5) 625 (38.4) 1005 (61.6) 0.55*** (0.46 to 0.66)
 A frican traditional 528 (9.6) 215 (40.8) 313 (59.2) 0.61*** (0.49 to 0.76)
  No religion/others 319 (5.8) 108 (34.0) 211 (66.0) 0.46*** (0.34 to 0.61)
Ethnicity
  Adja 729 (13.2) 364 (50.0) 365 (50.0) 1.00
 B ariba 660 (12.0) 260 (39.3) 400 (60.7) 0.65* (0.46 to 0.91)
 D endi 322 (5.8) 131 (40.8) 191 (59.2) 0.69 (0.44 to 1.07)
  Fon 2064 (37.4) 1163 (56.4) 901 (43.6) 1.29* (1.03 to 1.62)
 Y oa, Lokpa 147 (2.7) 57 (38.7) 90 (61.3) 0.63 (0.39 to 1.01)
  Betamaribe 298 (5.4) 102 (34.2) 196 (65.8) 0.52*** (0.36 to 0.75)
 P eulh 343 (6.2) 57 (16.5) 286 (83.5) 0.20*** (0.13 to 0.30)
  Yoruba 692 (12.6) 312 (45.1) 380 (54.9) 0.82 (0.62 to 1.09)
  Other beninois 195 (3.5) 76 (38.7) 119 (61.3) 0.63 (0.39 to 1.03)
  Other nationalities 67 (1.2) 37 (56.2) 29 (43.8) 1.29 (0.73 to 2.25)
Covered by health insurance
  No 5457 (98.9) 2507 (45.9) 2950 (54.1) 1.00
  Yes 60 (1.1) 52 (87.1) 8 (12.9) 7.96*** (3.98 to 15.92)
Had sexually transmitted infections in the last 12 months
 N o 5080 (92.1) 2349 (46.2) 2731 (53.8) 1.00
 Y es 437 (7.9) 210 (48.2) 227 (51.8) 1.08 (0.83 to 1.40)
Condom used during last sex with most recent partner
 N o 5218 (94.6) 2423 (46.4) 2795 (53.6) 1.00
  Yes 299 (5.4) 136 (45.4) 163 (54.6) 0.96 (0.73 to 1.25)
Comprehensive HIV knowledge
  No 4459 (80.8) 1918 (43.0) 2541 (57.0) 1.00
  Yes 1058 (19.2) 641 (60.6) 417 (39.4) 2.04*** (1.72 to 2.41)
Continued
4 Adu C, et al. BMJ Open 2023;13:e068805. doi:10.1136/bmjopen-2022-068805
BMJ Open: first published as 10.1136/bmjopen-2022-068805 on 13 April 2023. Downloaded from http://bmjopen.bmj.com/ on May 5, 2023 by guest. Protected by copyright.
Open access
Table 1 Continued
HTC
Variable Weighted N (%) Yes (%) No (%) cOR (95% CI)
No of sexual partners excluding spouse, in last 12 months
  0 4539 (82.3) 2224 (49.0) 2314 (51.0) 1.00
  1 878 (15.9) 303 (34.5) 575 (65.5) 0.55*** (0.46 to 0.65)
  2 or more 101 (1.8) 32 (31.7) 69 (68.3) 0.48** (0.31 to 0.75)
Exposure to mass media
  None 1864 (33.8) 655 (35.1) 1029 (64.9) 1.00
  1 1669 (30.2) 732 (43.8) 937 (56.2) 1.44*** (1.23 to 1.69)
  2 1627 (29.5) 921 (56.6) 706 (43.4) 2.41*** (2.02 to 2.87)
  3 356 (6.5) 251 (70.5) 105 (29.5) 4.41*** (3.36 to 5.78)
Wealth index
 P oorest 982 (17.8) 273 (27.8) 709 (72.2) 1.00
  Poorer 1056 (19.1) 371 (35.1) 685 (64.9) 1.40** (1.13 to 1.74)
  Middle 1048 (19.0) 437 (41.8) 610 (58.2) 1.86*** (1.47 to 2.36)
  Richer 1173 (21.3) 544 (53.6) 544 (46.4) 3.00*** (2.37 to 3.80)
 R ichest 1258 (22.8) 409 (67.5) 409 (32.5) 5.38*** (4.27 to 6.79)
Place of residence
 U rban 2380 (43.1) 1339 (56.3) 1041 (43.7) 1.00
  Rural 3137 (56.9) 1220 (38.9) 1917 (61.1) 0.49*** (0.42 to 0.59)
Region
  Alibori 767 (13.9) 240 (31.3) 527 (68.7) 1.00
  Atacora 406 (7.4) 125 (30.7) 281 (69.3) 0.97 (0.62 to 1.53)
  Atlantic 715 (13.0) 420 (58.7) 296 (41.3) 3.12*** (2.07 to 4.70)
  Borgou 501 (9.1) 184 (36.7) 317 (63.3) 1.27 (0.80 to 2.02)
  Collines 465 (8.4) 164 (35.2) 301 (64.8) 1.19 (0.78 to 1.83)
  Couffo 356 (6.4) 171 (48.1) 185 (51.9) 2.03** (1.33 to 3.12)
  Donga 294 (5.3) 126 (42.9) 168 (57.1) 1.65* (1.05 to 2.58)
 L ittoral 337 (6.1) 248 (73.7) 89 (26.3) 6.16*** (4.09 to 9.28)
 M ono 264 (4.8) 129 (48.8) 135 (51.2) 2.10** (1.24 to 3.54)
  Ouémé 492 (8.9) 287 (58.2) 205 (41.8) 3.06*** (2.06 to 4.55)
  Plateau 354 (6.4) 153 (43.1) 202 (56.9) 1.66* (13 to 2.45)
  Zou 567 (10.3) 315 (55.5) 252 (44.5) 2.74*** (1.82 to 4.14)
Community literacy level
  Low 1878 (34.0) 665 (35.4) 1214 (64.6) 1.00
  Medium 2073 (37.6) 910 (43.9) 1164 (56.1) 1.44** (1.16 to 1.79)
  High 1565 (28.4) 985 (62.9) 580 (37.1) 2.72*** (2.21 to 3.34)
Community socioeconomic status
  Low 2237 (40.5) 808 (36.1) 1429 (63.9) 1.00
  Medium 1506 (27.3) 676 (44.9) 830 (55.1) 1.43** (1.14 to 1.79)
  High 1774 (32.2) 1075 (60.6) 699 (39.4) 3.10*** (2.50 to 3.84)
*p<0.05, **p<0.01, ***p<0.001; 1.00=reference category.
cOR, crude OR.
This finding suggests that the education of women is very different parts of the world36 37 and is now evident in 
important in efforts towards HIV prevention and ending the context of HTC uptake among women in Benin. 
the HIV epidemic. This is because media outlets have been reported to 
Moreover, women exposed to mass media had an be one of the most effective medium to reach mass 
increased likelihood of HTC uptake. The use of media audience in a short period.36 While reading news-
outlets such as the radio, television, and reading news- paper or magazine is dependent on a person’s literacy 
papers or magazines as a behavioural change interven- level, radio and television on the other hand is not. 
tion has improved access to a lot of health services in This finding suggests the need to enhance the use of 
Adu C, et al. BMJ Open 2023;13:e068805. doi:10.1136/bmjopen-2022-068805 5
BMJ Open: first published as 10.1136/bmjopen-2022-068805 on 13 April 2023. Downloaded from http://bmjopen.bmj.com/ on May 5, 2023 by guest. Protected by copyright.
Open access 
Table 2 Mixed- effect analysis of factors associated with HIV testing and counselling among women in Benin
Model I Model II Model III
Variables Model O aOR (95% CI) aOR (95% CI) aOR (95% CI)
Fixed effect
Women’s age (years)
 1 5–19 1.00 1.00
 2 0–24 2.74*** (1.99 to 3.78) 2.74*** (1.99 to 3.77)
 2 5–29 4.04*** (2.90 to 5.63) 3.95*** (2.84 to 5.50)
 3 0–34 4.97*** (3.46 to 7.13) 4.86*** (3.39 to 6.98)
  35–39 4.27*** (2.96 to 6.18) 4.11*** (2.85 to 5.95)
  40–44 3.77*** (2.56 to 5.55) 3.60*** (2.45 to 5.31)
 4 5–49 2.24*** (1.43 to 3.52) 2.15*** (1.37 to 3.38)
Level of education
  No formal education 1.00 1.00
  Primary 1.51*** (1.21 to 1.90) 1.45** (1.15 to 1.82)
  Secondary or higher 2.17*** (1.73 to 2.73) 2.06*** (1.64 to 2.61)
Religion
  Christianity 1.00 1.00
  Islamic 1.09 (0.83 to 1.43) 1.12 (0.85 to 1.48)
  African traditional 0.73* (0.54 to 0.99) 0.73* (0.54 to 0.98)
  No religion/others 0.82 (0.57 to 1.19) 0.84 (0.58 to 1.21)
Current working status
  No 1.00 1.00
 Y es 1.04 (0.83 to 1.31) 1.04 (0.83 to 1.31)
Ethnicity
 A dja 1.00 1.00
 B ariba 1.06 (0.68 to 1.65) 1.84* (1.01 to 3.35)
 D endi 0.64 (0.35 to 1.17) 1.02 (0.49 to 2.12)
  Fon 1.15 (0.84 to 1.58) 1.20 (0.79 to 1.82)
  Yoa, Lokpa 0.74 (0.35 to 1.59) 1.19 (0.52 to 2.73)
  Betamaribe 0.96 (0.52 to 1.76) 1.72 (0.82 to 3.63)
  Peulh 0.31*** (0.17 to 0.58) 0.58 (0.28 to 1.16)
  Yoruba 0.94 (0.62 to 1.43) 1.16 (0.68 to 1.99)
  Other Beninois 0.96 (0.51 to 1.79) 1.59 (0.77 to 3.28)
 O ther nationalities 0.81 (0.41 to 1.62) 1.05 (0.50 to 2.22)
Covered by health insurance
  No 1.00 1.00
  Yes 3.05** (1.45 to 6.41) 3.04** (1.44 to 6.43)
Comprehensive HIV knowledge
 N o 1.00 1.00
  Yes 1.74*** (1.40 to 2.17) 1.77*** (1.43 to 2.21)
No of sexual partners excluding spouse, in last 12 months
  0 1.00 1.00
  1 0.45*** (0.35 to 0.59) 0.44*** (0.34 to 0.57)
  2 or more 0.37** (0.19 to 0.70) 0.35** (0.18 to 0.68)
Exposure to mass media
  None 1.00 1.00
  1 1.26* (1.02 to 1.54) 1.22 (0.99 to 1.49)
 2 1.42** (1.12 to 1.82) 1.39** (1.09 to 1.78)
 3 1.68* (1.09 to 2.57) 1.63* (1.07 to 2.51)
Continued
6 Adu C, et al. BMJ Open 2023;13:e068805. doi:10.1136/bmjopen-2022-068805
BMJ Open: first published as 10.1136/bmjopen-2022-068805 on 13 April 2023. Downloaded from http://bmjopen.bmj.com/ on May 5, 2023 by guest. Protected by copyright.
Open access
Table 2 Continued
Model I Model II Model III
Variables Model O aOR (95% CI) aOR (95% CI) aOR (95% CI)
Wealth index
  Poorest 1.00 1.00
 P oorer 0.91 (0.69 to 1.18) 0.88 (0.67 to 1.15)
  Middle 1.09 (0.81 to 1.47) 1.01 (0.75 to 1.37)
  Richer 1.38* (1.01 to 1.90) 1.19 (0.86 to 1.65)
  Richest 1.90*** (1.32 to 2.73) 1.39 (0.95 to 2.05)
Place of residence
  Urban 1.00 1.00
 R ural 0.55*** (0.45 to 0.69) 0.73** (0.57 to 0.92)
Region
  Alibori 1.00 1.00
  Atacora 0.84 (0.52 to 1.37) 0.89 (0.49 to 1.60)
  Atlantic 2.56*** (1.65 to 3.98) 2.81*** (1.53 to 5.17)
  Borgou 0.82 (0.48 to 1.40) 0.88 (0.50 to 1.56)
  Collines 0.90 (0.55 to 1.46) 1.20 (0.63 to 2.29)
  Couffo 1.90** (1.20 to 3.02) 2.64* (1.25 to 5.60)
  Donga 1.32(0.81 to 2.16) 1.18 (0.64 to 2.17)
  Littoral 2.23** (1.38 to 3.60) 2.32** (1.25 to 4.28)
 M ono 1.50(0.90 to 2.50) 2.23* (1.04 to 4.77)
 O uémé 1.87** (1.22 to 2.87) 1.80 (0.98 to 3.30)
  Plateau 1.57* (1.05 to 2.35) 1.75 (0.95 to 3.20)
  Zou 2.26*** (1.42 to 3.60) 2.90** (1.53 to 5.50)
Community literacy level
  Low 1.00 1.00
 M edium 1.33* (1.05 to 1.68) 1.23(0.97 to 1.58)
  High 2.23*** (1.67 to 2.98) 1.76*** (1.28 to 2.41)
Community socioeconomic status
  Low 1.00 1.00
  Medium 1.44** (1.13 to 1.83) 1.21(0.94 to 1.56)
 H igh 1.91*** (1.49 to 2.45) 1.43* (1.08 to 1.90)
Random effect model
  PSU variance (95% CI) 1.649(1.367, 1.988) 1.037(0.833 to 1.291) 0.853(0.699 to 1.042) 0.872(0.707 to 1.076)
  ICC 0.333 0.240 0.206 0.210
 W ald χ2 Reference 424.16 (<0.001) 343.13 (<0.001) 654.39 (<0.001)
Model fitness
 L og- likelihood −9620.8883 −8840.3917 −9476.8784 −8780.1638
 A IC 19 245.78 17 748.78 18 989.76 17 660.33
 N 5517 5517 5517 5517
  No of clusters 555 555 555 555
*p<0.05, **p<0.01, ***p<0.001; 1=reference category.
AIC, Akaike’s information criterion; aOR, adjusted OR; ICC, intraclass correlation; PSU, primary sampling unit.
radio, especially as a medium of behavioural change rural areas have reduced odds of HTC uptake. This 
communication. finding could be due to geographical disparity in 
We also found that women residing in rural areas access to healthcare services in most parts of SSA 
have lower likelihood of HTC uptake compared to including Benin where women in urban areas usually 
women in urban areas. This result aligns with finding have higher access to quality and needed health 
from a previous study in Ghana32 where women in services due to some socioeconomic privilege in urban 
Adu C, et al. BMJ Open 2023;13:e068805. doi:10.1136/bmjopen-2022-068805 7
BMJ Open: first published as 10.1136/bmjopen-2022-068805 on 13 April 2023. Downloaded from http://bmjopen.bmj.com/ on May 5, 2023 by guest. Protected by copyright.
Open access 
dwellings. This finding suggests the need to improve departments in Benin.52 Perhaps, the low socioeconomic 
universal health coverage in Benin.38 status of women in the region couple with reduced 
Additionally, women aged 20–49 years were more likely access to healthcare services53 contributed to the low 
to test for HIV compared to those aged 15–19 years. HIV screening uptake. Thus, improving availability and 
This finding affirms the reports from previous studies in ensuring easy accessibility to HIV screening services with 
Zambia25 and Gambia.1 The high uptake of HTC among limited or no cost to the women could improve screening 
women aged 20 years and above could be attributed to uptake among women in the Alibori region.
their increased exposure to HIV-r elated information 
and increased awareness. For instance, women who are Strength and limitation
20 years and older are more likely to get pregnant and The major strength of our study is the use of nationally 
thus receive HIV education, counselling and testing representative data to examine the prevalence and predic-
through the prevention of mother-t o-c hild transmission tors of HTC among women. Also, the use of weighting 
programmes.39 40 Meanwhile, evidence suggests that and rigorous statistical analysis to a complex sample 
younger women (15–19 years) have higher risk for HIV helped to reduce the undersampling and oversampling 
due to their increased exposure to multiple sexual part- of the DHS dataset. For the limitations, causal inferences 
ners.41 Like the findings of a previous study in South cannot be drawn from the study due to the cross- sectional 
Africa,42 our result suggests that having higher number of nature of the DHS dataset. In addition, there is a hint of 
sexual partners was associated with a reduced likelihood recall bias in the study as the variables were assessed using 
of HTC uptake, perhaps due to fear of HIV diagnosis and a self- reporting method. Further, our study is limited to 
limited confidentiality.42 The high risk of HIV exposure only the variables present in the dataset. As a result, the 
coupled with a reduced propensity to test increases the findings should be interpreted in light of the studied 
risk of HIV infection among adolescents. This emphasises variables.
the need for adolescent-t argeted HIV education, counsel-
ling and testing programmes to improve testing uptake.
We also found that the likelihood of HTC uptake was CONCLUSION
higher among women from communities with high We found a relatively low uptake of HTC in Benin. There 
literacy levels. Women in communities with high literacy is the need to enhance efforts to empower women, as well 
rates are more likely to have higher education and thus as reduce health inequities as they all have a substantial 
comprehensive HIV/AIDS knowledge,43 which increases impact on HTC uptake among women in Benin, taking 
their propensity to uptake HTC.44 45 In addition, evidence into consideration the factors identified in this study.
shows that higher education is associated with increased 
healthcare decision- making capacity and HIV screening Author affiliations
uptake among women in SSA.22 Our findings again 1College of Public Health, Medical and Veterniary Sciences, James Cook University, 
revealed that women living in communities with higher Townsville, Queensland, Australia
2
socioeconomic status were more likely to uptake HTC. Center for Social Research in Health, University of New South Wales, Sydney, New 
South Wales, Australia
Similar findings were reported in previous studies.25 46 47 3Department of Epidemiology and Biostatistics, Fred N. Binka School of Public 
Higher socioeconomic status had been associated with Health, University of Health and Allied Sciences, Ho, Ghana
increased empowerment and autonomy in healthcare 4Department of Social and Behavioural Science, University of Ghana, Legon, Ghana
5
decision- making and healthcare seeking behaviour Department of Family Medicine, McGill University Montreal, Montréal, Quebec, 
among women in SSA22 48 and thus increasing the likeli- Canada6Department of Health, Physical Education and Recreation, University of Cape Coast, 
hood of HTC uptake. Cape Coast, Ghana
Also, our finding suggests that women belonging to 7Department of Family and Community Health, Fred N. Binka School of Public 
the African Traditional religion were less likely to test for Health, University of Health and Allied Sciences, Ho, Ghana
HIV compared to their Christian counterparts. Similar 
findings were reported in Burkina Faso47 and in Ghana.49 Acknowledgements We are grateful to the MEASURE DHS for making the DHS 
Women who belong to the traditional religion have a dataset freely accessible to use for the study.
strong belief in supernatural powers and faith, which Contributors CA, LAA and RGA conceived the study. RGA, AC, LAA and CA wrote 
50 the methods section and performed the data analysis. CA, PAY, AM and AC were reduces their wiliness to uptake HTC.  For instance, the responsible for the initial draft of the manuscript. All the authors reviewed and 
high rate of Voodooism adherents in Benin had been approved the final version of the manuscript. RGA is the guarantor and accepts full 
associated with reduced uptake of preventive healthcare responsibility for the work and/or the conduct of the study.
services.51 Perhaps, educating and involving leaders of Funding The authors have not declared a specific grant for this research from any 
the traditional religion in designing and implementing funding agency in the public, commercial or not- for- profit sectors.
targeted screening programmes could improve HTC Competing interests None declared.
uptake among members. Patient and public involvement Patients and/or the public were not involved in 
Further, we found that women living in Alibori region the design, or conduct, or reporting, or dissemination plans of this research.
were least likely to test for HIV compared to those from Patient consent for publication Not applicable.
Atlantic, Littoral, Couffo, Mono and Zou. Alibori, located Ethics approval Ethical clearance was not sought for the current study since 
in the Northmost part of Benin, is one of the poorest the dataset is publicly available in the public domain. The detailed information 
8 Adu C, et al. BMJ Open 2023;13:e068805. doi:10.1136/bmjopen-2022-068805
BMJ Open: first published as 10.1136/bmjopen-2022-068805 on 13 April 2023. Downloaded from http://bmjopen.bmj.com/ on May 5, 2023 by guest. Protected by copyright.
Open access
concerning the dataset and the ethical guidelines can be accessed at http://goo.gl/  16 Corsi DJ, Neuman M, Finlay JE, et al. Demographic and health 
ny8T6X. We adhered to the ethical guidelines regarding the use of secondary data surveys: a profile. Int J Epidemiol 2012;41:1602–13. 
for publication.  17 ICF International. Demographic and health survey sampling and 
household listing manual. Calverton, Maryland, USA: Measure DHS, 
Provenance and peer review Not commissioned; externally peer reviewed. 2012.
Data availability statement Data are available in a public, open access repository.  18 von Elm E, Altman DG, Egger M, et al. The strengthening the 
reporting of observational studies in epidemiology (STROBE) 
The dataset is freely accessible at https://dhsprogram.com/data/dataset/Benin_ statement: guidelines for reporting observational studies. Int J Surg 
Standard-DHS_2017.cfm?flag=1. 2014;12:1495–9. 
Supplemental material This content has been supplied by the author(s). It has  19 DHS. DHS data source. n.d. Available: https://dhsprogram.com/data/ 
not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been dataset/Benin_Standard-DHS_2017.cfm?flag=1
 20 Tetteh JK, Frimpong JB, Budu E, et al. Comprehensive HIV/AIDS 
peer- reviewed. Any opinions or recommendations discussed are solely those knowledge and HIV testing among men in sub-S aharan Africa: a 
of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and multilevel modelling. J Biosoc Sci 2022;54:975–90. 
responsibility arising from any reliance placed on the content. Where the content  21 Seidu AA. Using anderson’s model of health service utilization to 
includes any translated material, BMJ does not warrant the accuracy and reliability assess the use of HIV testing services by sexually active men in 
of the translations (including but not limited to local regulations, clinical guidelines, Ghana. Front Public Health 2020;8:512. 
terminology, drug names and drug dosages), and is not responsible for any error  22 Seidu A- A, Oduro JK, Ahinkorah BO, et al. Women’s healthcare 
and/or omissions arising from translation and adaptation or otherwise. decision- making capacity and HIV testing in sub-S aharan Africa: 
a multi-c ountry analysis of demographic and health surveys. BMC 
Open access This is an open access article distributed in accordance with the Public Health 2020;20:1592. 
Creative Commons Attribution Non Commercial (CC BY- NC 4.0) license, which  23 Takarinda KC, Madyira LK, Mhangara M, et al. Factors associated 
permits others to distribute, remix, adapt, build upon this work non-c ommercially, with ever being HIV- tested in Zimbabwe: an extended analysis of the 
and license their derivative works on different terms, provided the original work is Zimbabwe demographic and health survey (2010-2 011). PLoS ONE 
properly cited, appropriate credit is given, any changes made indicated, and the use 2016;11:e0147828. 
is non- commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.  24 Erena AN, Shen G, Lei P. Factors affecting HIV counselling and 
testing among Ethiopian women aged 15-4 9. BMC Infect Dis 
ORCID iDs 2019;19:1076.  25 Muyunda B, Musonda P, Mee P, et al. Educational attainment as a 
Collins Adu http://orcid.org/0000-0002-1300-6327 predictor of HIV testing uptake among women of child- bearing age: 
Richard Gyan Aboagye http://orcid.org/0000-0002-3498-2909 analysis of 2014 demographic and health survey in Zambia. Front 
Public Health 2018;6:192. 
 26 Center KE, Gunn JKL, Asaolu IO, et al. Contraceptive use and uptake 
of HIV- testing among sub- Saharan African women. PLoS ONE 
REFERENCES 2016;11:e0154213. 
 1 Deynu M, Agyemang K, Anokye N. Factors associated with HIV  27 Mandiwa C, Namondwe B. Uptake and correlates of HIV testing 
testing among reproductive women aged 15- 49 years in the Gambia: among men in malawi: evidence from a national population- based 
analysis of the 2019- 2020 Gambian demographic and health survey. household survey. BMC Health Serv Res 2019;19:203. 
Int J Environ Res Public Health 2022;19:4860.  28 Johar M, Soewondo P, Pujisubekti R, et al. Inequality in access to 
 2 Paulin HN, Blevins M, Koethe JR, et al. HIV testing service health care, health insurance and the role of supply factors. Soc Sci 
awareness and service uptake among female heads of household in Med 2018;213:134–45. 
rural Mozambique: results from a province- wide survey. BMC Public  29 Harris B, Goudge J, Ataguba JE, et al. Inequities in access 
Health 2015;15:132. to health care in South Africa. J Public Health Policy 
 3 UNAIDS. UNAIDS data. 2020. Available: www.unaids.org/sites/ 2011;32 Suppl 1:S102–23. 
default/files/media_asset/2020_aids-data-book_en.pdf  30 Sanogo NA, Yaya S. Wealth status, health insurance, and maternal 
 4 UNAIDS. Global AIDS response progress reporting 2015. Geneva, health care utilization in Africa: evidence from Gabon. Biomed Res Int 
Switzerland UNAIDS; 2015. 2020;2020:4036830. 
 5 Teklehaimanot HD, Teklehaimanot A, Yohannes M, et al. Factors  31 Gobah FK, Zhang L. The national health insurance scheme in 
influencing the uptake of voluntary HIV counseling and testing in rural ghana: prospects and challenges: a cross- sectional evidence. GJHS 
ethiopia: a cross sectional study. BMC Public Health 2016;16:239. 2011;3:90. 
 6 UNAIDS. The GAP report. 2014. Available: https://files.unaids.org/  32 Sambah F, Baatiema L, Appiah F, et al. Educational attainment and 
en/media/unaids/contestassets/documents/unaidspublication/2014/ HIV testing and counselling service utilisation during antenatal care 
UNAIDS_GAP_report_en.pdf in Ghana: analysis of demographic and health surveys. PLoS One 
 7 Asaolu IO, Gunn JK, Center KE, et al. Predictors of HIV testing 2020;15:e0227576. 
among youth in sub- Saharan Africa: a cross-s ectional study. PLoS  33 Merdekios B, Adedimeji AA. Effectiveness of interventions to prevent 
ONE 2016;11:e0164052. mother- to-c hild transmission of HIV in Southern Ethiopia. Int J 
 8 Rosenberg NE, Westreich D, Bärnighausen T, et al. Assessing the Womens Health 2011;3:359–66. 
effect of HIV counselling and testing on HIV acquisition among South  34 Sambah F, Hagan Jr JE, Mintah JK, et al. Determinants of HIV testing 
African youth. AIDS 2013;27:2765–73. and counseling utilization among trainee nurses and midwives in 
 9 Kurth AE, Lally MA, Choko AT, et al. HIV testing and linkage to central region of Ghana. Clin Res Psychol 2019;2:1–0. 
services for youth. J Int AIDS Soc 2015;18:19433.  35 Sambah F, Hormenu T, Ahinkorah BO, et al. Nurses/midwifery trainees’ 
 10 Kharsany ABM, Karim QA. HIV infection and AIDS in sub- Saharan knowledge, attitudes and utilisation of HIV testing/counselling service 
Africa: current status, challenges and opportunities. Open AIDS J in the central region of Ghana. JHS 2018;6:449–60. 
2016;10:34–48.  36 Saaka M, Wemah K, Kizito F, et al. Effect of nutrition behaviour 
 11 Morin L, Béhanzin L, Guédou FA, et al. HIV prevention and treatment change communication delivered through radio on mothers’ 
cascades among female sex workers in Benin, West Africa. Sex nutritional knowledge, child feeding practices and growth. J Nutr Sci 
Transm Dis 2021;48:654–62. 2021;10:e44. 
 12 Diress G, Ahmed M, Adane S, et al. Barriers and facilitators for HIV  37 Ngigi S, Busolo DN. Behaviour change communication in health 
testing practice among Ethiopian women aged 15- 24 years: analysis promotion: appropriate practices and promising approaches. Ijird 
of the 2016 Ethiopian demographic and health survey. HIV AIDS 2018;7:84–93. 
(Auckl) 2020;12:963–70.  38 Paul E, Deville C, Bodson O, et al. How is equity approached in 
 13 Bekele YA, Fekadu GA. Factors associated with HIV testing among universal health coverage? An analysis of global and country policy 
young females; further analysis of the 2016 Ethiopian demographic documents in benin and Senegal. Int J Equity Health 2019;18:195. 
and health survey data. PLoS One 2020;15:e0228783.  39 Musarandega R, Machekano R, Chideme M, et al. PMTCT service 
 14 Worku MG, Tesema GA, Teshale AB. Prevalence and associated uptake among adolescents and adult women attending antenatal 
factors of HIV testing among reproductive- age women in Eastern care in selected health facilities in Zimbabwe. J Acquir Immune Defic 
Africa: multilevel analysis of demographic and health surveys. BMC Syndr 2017;75:148–55. 
Public Health 2021;21:1262.  40 Tabana H, Doherty T, Swanevelder S, et al. Knowledge of HIV status 
 15 Croft TN, Marshall AMJ, Allen CK. Guide to DHS statistics, DHS-7 . prior to a community HIV counseling and testing intervention in a 
Rockville, Maryland, USA ICF; 2018. Available: https://dhsprogram. rural district of South Africa: results of a community based survey. 
com/pubs/pdf/DHSG1/Guide_to_DHS_Statistics_DHS-7.pdf BMC Infect Dis 2012;12:1–8. 
Adu C, et al. BMJ Open 2023;13:e068805. doi:10.1136/bmjopen-2022-068805 9
BMJ Open: first published as 10.1136/bmjopen-2022-068805 on 13 April 2023. Downloaded from http://bmjopen.bmj.com/ on May 5, 2023 by guest. Protected by copyright.
Open access 
 41 Peltzer K, Matseke G. Determinants of HIV testing among  47 Kirakoya- Samadoulougou F, Jean K, Maheu- Giroux M. Uptake of HIV 
young people aged 18-2 4 years in South Africa. Afr Health Sci testing in burkina faso: an assessment of individual and community- 
2013;13:1012–20. level determinants. BMC Public Health 2017;17:486. 
 42 Helleringer S, Kohler HP, Frimpong JA, et al. Increasing uptake of HIV  48 Worku AG, Yalew AW, Afework MF. Maternal complications and 
testing and counseling among the poorest in sub- Saharan countries women’s behavior in seeking care from skilled providers in North 
through home-b ased service provision. J Acquir Immune Defic Syndr Gondar, Ethiopia. PLoS ONE 2013;8:e60171. 
2009;51:185–93.  49 Nyarko SH, Sparks C. Levels and determinants of HIV testing uptake 
 43 Zegeye B, Anyiam FE, Ahinkorah BO, et al. Women’s decision- among Ghanaian men. Afr J AIDS Res 2020;19:40–7. 
making capacity and its association with comprehensive knowledge  50 Fakoya I, Johnson AM, Fenton KA, et al. Religion and HIV diagnosis 
of HIV/AIDS in 23 sub-S aharan African countries. Arch Public Health 
2022;80:111. among Africans living in London. HIV Med 2012;13:617–22. 
 44 Mwaba K, Mannell J, Burgess R, et al. Uptake of HIV testing among  51 Stoop N, Verpoorten M, Deconinck K. Voodoo, vaccines, and bed 
15–19-y ear-o ld adolescents in Zambia. AIDS Care 2020;32:183–92. nets. Econ Dev Cult Change 2019;67:493–535. 
 45 Paulin HN, Blevins M, Koethe JR, et al. HIV testing service  52 Altaf A, Pouw N. Defining, targeting, and reaching the very poor in 
awareness and service uptake among female heads of household in benin. In: What Works for Africa’s Poorest: programmes and Policies 
rural Mozambique: results from a province-w ide survey. BMC Public for the Extreme Poor. 2017: 17.
Health 2015;15:1–11.  53 Zegeye B, El- Khatib Z, Ameyaw EK, et al. Breaking barriers 
 46 Ajayi AI, Awopegba OE, Adeagbo OA, et al. Low coverage of HIV to healthcare access: a multilevel analysis of individual- and 
testing among adolescents and young adults in Nigeria: implication community- level factors affecting women’s access to healthcare 
for achieving the UNAIDS first 95. PLoS ONE 2020;15:e0233368. services in Benin. Int J Environ Res Public Health 2021;18:750. 
10 Adu C, et al. BMJ Open 2023;13:e068805. doi:10.1136/bmjopen-2022-068805
BMJ Open: first published as 10.1136/bmjopen-2022-068805 on 13 April 2023. Downloaded from http://bmjopen.bmj.com/ on May 5, 2023 by guest. Protected by copyright.