Despite strong international mobilization and recommendations, human immunodeficiency virus (HIV) counseling and testing remains insufficient to control the epidemic in West Africa. In 2017, only 48% of people living with HIV knew their status despite UNAIDS' goal of reaching 90% by 2020.1 In order to reduce the proportion of people who doesn't know their HIV status, WHO recommends, since 2007, routine Provider-Initiated HIV Testing and Counselling (PITC) regardless of the consultation reasons in countries where HIV prevalence is above 1%.2
In 2009, with an estimated 2.9%3 HIV prevalence, Côte d'Ivoire introduced a routine PITC policy for all patients attending health facilities.4 After several years of implementation, routine PITC remains low.5 In addition, the current stabilization of HIV prevalence at around 2.8%3 and the international donors policies' shift to targeted approaches are directing testing activities towards “high-yield” and “high-impact” sites and populations.6 In this context of disinterest and disengagement regarding routine PITC, it is important to ensure that an HIV testing is at least routinely offered at specific life events clinically recommended (pregnancy, sexually transmitted infection [STI]) or legally recommended (premarital check-up).4,7 Maintaining these screening opportunities are all the more crucial since only 54% of HIV-infected people knew their status in this country in 2017.1
Here, our objective is to identify to what extent the occurrence of certain life events leads to HIV testing in the general population of Côte d'Ivoire. Four events recommended at national level for HIV testing were studied: pregnancy (for women), pregnancy of a partner (for men), any type of STI episode and marriage (premarital testing is not legally required but the Ivorian government recommends a test offer at this occasion). Different events were considered since PITC implementation and practice may vary according to the type of consultation.8 For each event, we want to consider the different steps between the occurrence of an event and an effective HIV testing, i.e., the “HIV testing cascade.” The concept of “cascade” is commonly used to present the different steps of a process such as the “HIV care cascade.”1
MATERIALS AND METHODS
Between February and November 2017, a cross-sectional survey on HIV testing experiences was conducted among a nationally representative sample of the population aged 16–59 in Côte d'Ivoire. The sample was obtained from a 2-stage telephone survey. An 8-digit phone number list was randomly generated, using prefixes attributed to national phone companies. Then trained interviewers dialed all telephone numbers; the numbers were dialed between 8 a.m. and 8 p.m., Monday to Saturday. In case of contact, the study was introduced to the respondent and the eligibility criteria were checked (eg, aged between 16 and 59 years and living in Côte d'Ivoire for at least 6 months per year). After obtaining verbal consent, the respondent was asked to list all household members living with him/her, their age, residency status and the number of telephone numbers owned by each member (it is common in Côte d'Ivoire to have several numbers). Then one of the eligible household members was randomly selected and invited to participate in the survey. The survey was interviewer-administered. In case of noncontact, a number was called back at least 15 times before being considered unsuccessful. Different recall times and days were implemented to increase the chances of contact. This study, part of the ANRS 12323 DOD-CI project, obtained approval from the National Ethics Committee of Côte d'Ivoire (N/Ref: 058/MSHP/CNER-kp).
Data collected were self-provided (no other member of the household provided information by proxy). Data about the occurrence of certain events such as a marriage (civil or customary), a childbirth and an STI episode were collected over the last 5 years. For each of these events, it was asked (i) whether the person had consulted a health care professional on that occasion, (ii) whether an HIV test offer was made by the health care professional during that consultation and (iii) whether that test offer was accepted by the participant. For the last child, men were asked if they accompanied the mother of their child to at least 1 antenatal visit and if they received and accepted a test offer on that occasion. The survey did not ask if the participant returned for the results.
Explanatory variables included: socio-demographic characteristics (age, sex, level of education, religion, occupation, mobility over the past 12 months), information (media exposure), individual perceptions (perceived HIV exposure, self-perceived health status, perceived benefit of antiretroviral treatment), individual resources and autonomy (index of economic well-being, reporting having a health insurance, decision-making process within the couple) and individual environment (having an HIV-positive relative and residence department). The index of economic well-being, the media exposure and the perception of the ARV treatment benefit were built from questions presented in the Supplemental digital content (Supplemental Text S1, http://links.lww.com/OLQ/A424).
For each event, all explanatory variables were analyzed at each step of the “HIV testing cascade.” Bivariate analysis was conducted using the χ2 test. Then variables associated with the 0.20 threshold were included in multivariate binary logistic regression models.9 When applicable, sex and age were forced in all the models. Wald tests were used for calculation of the P value in the multivariate models.
The analyses considered the study design. First, survey weights were computed to take into account the probability to be included into the sample, depending on the number of eligible members and the number of telephone numbers in the household. Then, an iterative poststratification was used to calibrate the sample on the marginal distribution of the national population per gender, age, education level and region, based on data from the last general population census conducted in Côte d'Ivoire in 2014. Analyses were performed using R version 3.3.5 and survey package.10
Populations and Events Description
Overall, 9,600 telephone numbers were dialed, 2,076 numbers did not lead to a contact and 662 led to an ineligible person (Supplemental Table S1, http://links.lww.com/OLQ/A425). Of the remaining 6,862 contacted respondents, 10.2% refused to participate in the survey (9.6% refused before the enumeration of household members and 0.6% after it) and 21.0% accounted for execution failures (19.5% could not be reached after agreeing on an appointment and 1.5% dropped out before completing the questionnaire). Thus, 3,867 individuals were interviewed (1,984 men and 1,883 women) for an overall response rate of 56.4% among eligible individuals.
HIV Testing Cascade During Last Pregnancy Among Women
In our sample, 792 women (42.1%) reported a live birth in the 5 years preceding the survey (Supplemental Table S2, http://links.lww.com/OLQ/A426). Those less educated and those 25–34 years old were more likely to have a child in the past 5 years.
Among those who reported a child, 94.9% consulted a health care professional for antenatal care (Fig. 1A). In bivariate analysis, women with an income-generating activity had a lower probability of consulting a health care professional (Supplemental Table S2, http://links.lww.com/OLQ/A426). This association was still found in the multivariate model (adjusted odds ratio [aOR], 0.15 [95% confidence interval, 0.03–0.66], P < 0.05) (Fig. 2 and Supplemental Table S3, http://links.lww.com/OLQ/A427).
Among those who consulted, 70.1% were offered the HIV test. Those less educated, Muslim people, and those who don't know an HIV-infected relative were less likely to be offered the test. In the multivariate analysis, having a secondary level of education (vs. none education) and being Catholic (vs. Muslim) were associated with a higher probability of being offered the test (aOR, 3.59 [1.47–8.77] and aOR, 6.79 [1.88–24.59], respectively).
When proposed, the test was accepted by almost all women (95.6%). Knowing an HIV-infected relative was associated with a higher test acceptance in both bivariate and multivariate analysis (aOR, 13.52 [1.25–145.92]).
HIV Testing Cascade During Partner's Last Pregnancy Among Men
Among the 1984 surveyed men, 794 (40.0%) reported a childbirth within the last 5 years (Supplemental Table S4, http://links.lww.com/OLQ/A428). Men with an income-generating activity and less educated men were more likely to report a birth in the past 5 years.
Among those who reported a child, 58.3% accompanied the mother of the child to at least 1 antenatal consultation (Fig. 1B). In the bivariate analysis, not living in Abidjan (the economic capital of Côte d'Ivoire) and, surprisingly, not having a health insurance were associated with being present in an antenatal visit (Supplemental Table S4, http://links.lww.com/OLQ/A428). In the multivariate model, living in urban or rural inland departments (vs. Abidjan: aOR, 2.17 [1.15–4.07] and aOR, 1.83 [1.09–3.08] respectively) and not having a health insurance (aOR, 2.23 [1.18–4.00]) were significantly associated with accompanying his child's mother for an antenatal visit (Fig. 3 and Supplemental Table S5, http://links.lww.com/OLQ/A429).
Among those who visited an antenatal clinic, 33.1% were offered the test. Older men and those with a higher economic well-being index were more likely to be offered to test. These 2 factors were still found significantly associated in the multivariate analysis.
In case of test offer, test acceptance was 86.0%. No variables were significantly associated with test acceptance in either bivariate or multivariate analysis.
HIV Testing Cascade Following an STI Episode
Of the 3867 individuals surveyed, 172 (4.5%) reported an STI episode in the 5 years preceding the survey (Supplemental Table S6, http://links.lww.com/OLQ/A430). Men and higher educated respondents were more likely to report an STI.
Among those who reported an STI, 70.3% reported having a subsequent consultation with a health care provider following this infection (Fig. 1C). Those who perceived themselves exposed to HIV (vs. not exposed) were more likely to consult a health care provider (Supplemental Table S6, http://links.lww.com/OLQ/A430). In the multivariate analysis, men and people younger than 24 years (vs. 24–35 years) were less likely to consult a health care professional (aOR, 0.45 [0.21–0.98] and aOR, 0.39 [0.18–0.85] respectively), whereas those exposed to media were significantly more likely to consult (aOR, 2.53 [1.19–5.38]) (Fig. 4 and Supplemental Table S7, http://links.lww.com/OLQ/A431).
Among those who consulted, 28.1% reported having received a test offer. Muslim respondents and those who did not know their exposure to HIV were less likely to receive a test offer. In the multivariate analysis, being Catholic or Evangelical (vs. Muslim: aOR, 3.75 [1.25–11.28] and aOR, 3.87 [1.20–12.42] respectively) was associated with receiving a test offer. Those who perceived themselves as exposed to HIV (vs. not exposed: aOR, 0.42 [0.19–0.92]) and those older than 35 years (vs. 25–35 years; aOR, 0.37 [0.13–1.04]) had fewer test offers.
In case of test offer, test acceptance was 67.6%. In the bivariate analysis, no factors were found significantly associated with test acceptance. Due to the small sample, it was not possible to converge a multivariate model for test acceptance.
HIV Testing Cascade on the Occasion of Marriage
In our sample, 14.0% reported getting married in the past 5 years (Supplemental Table S8, http://links.lww.com/OLQ/A432). Those less educated and young people were more likely to report a marriage in the past 5 years.
Among those who reported a marriage, 19.1% reported a health check-up on this occasion (Fig. 1D). In bivariate analysis, Catholic people, those with health insurance, and those who knew an HIV-infected relative were more likely to undergo a premarital health check-up (Supplemental Table S8, http://links.lww.com/OLQ/A432). In the multivariate model, having a premarital health check-up was significantly associated with being Catholic (vs. Muslim: aOR, 2.91 [1.29–6.58]) and possession of health insurance (aOR, 2.80 [1.09–7.19]) (Fig. 5 and Supplemental Table S9, http://links.lww.com/OLQ/A433).
During this check-up, 78.8% were offered the test for HIV. Knowing an HIV-infected relative and living in rural department was associated with being offered to test in the bivariate analysis. In the multivariate model, being 35 to 59 years old (vs. 25–34 years old) was significantly associated with being offered to test (aOR, 12.53 [1.34–116.79]).
When the test was offered, it was accepted by 96.3% of the respondents. No factors were found significantly associated with test acceptance. The small number of people who refused the test did not allow a multivariate model to converge.
Our study highlighted the low HIV testing coverage at events where testing is clearly recommended in national policies.11 These levels of HIV testing are mainly driven by lack of medical consultation (except for pregnant women) and lack of testing offer (except for premarital check-ups). The testing acceptance proportions were high.
Lack of medical consultation constitutes the first obstacle to HIV testing.
Although antenatal consultations were prevalent among women (95%), only slightly more than half of men accompanied the mother of their future child for an antenatal visit at least once. The presence of men during antenatal consultations varies greatly from 1 country to another.12 These differences are related to social norms regarding gender and male partners' poor reception during antenatal consultations—for example, poor reception by health care professionals and/or lack of dedicated space.12,13 In our sample, when asking women, only 36.5% of women were offered the opportunity to invite their partner to antenatal consultations for HIV testing (data not shown). Routine invitation of the male partner coupled with other interventions to stimulate his presence could improve male partner screening coverage.12,14
Consulting a health care professional for an STI episode was not common in our study (around two thirds of concerned individuals). Men, youth, and those with low-media exposure were less likely to attend consultations; potentially indicating a perceived poor acceptability of the health offer—for example, fear of being judged for having an STI15,16— as well as lack of information about the available health offer.17
Test offer was very low for men accompanying their pregnant partner to antenatal clinics. Improving male partner testing is all the more important as HIV-infected men are less likely to be diagnosed compared with women in Côte d'Ivoire.18
Offer proportion was also very low for individuals seeking a health care professional for an STI. Low offer proportions during STI consultations were found in South Africa due to the lack of integration of HIV testing with other medical activities and difficulties in managing counselling activities.19 In our study, the association between religion/age and testing offer could reflect a difficulty for health care professionals to speak about HIV testing with their patients and/or that they perceive some people at a lower risk of HIV acquisition. For example, health care professionals may have difficulties to address questions related to sexuality with 35 to 59 years old and/or Muslims.20 Older people may also be perceived as being less at risk by health care professionals due to misconceptions about the sexuality of older people.21 This perception is in contradiction with the latest population-based serological survey conducted in 2018 showing that among the 15 to 59 adult population, the highest HIV prevalence was among the 50 to 59 years old.18
Although higher (around 70%), the offer proportion among pregnant women in antenatal care remains relatively low when considering all the efforts to implement the prevention of mother-to-child transmission programs in Côte d'Ivoire over the last 2 decades.
The acceptance of the test remained relatively high in our study regardless of the event studied. While acceptance may have been lower in the past,22 HIV testing offer seems to be a medical “gesture” that is now well accepted by the general population and can be partly due to the availability of treatment free of charge in Côte d'Ivoire. Testing acceptance was slightly lower during STI consultations, but caution should be exercised with this result due to the wide confidence intervals.
Our study has some limitations. In 2016, 91.8% of Ivorian households owned a mobile phone in Côte d'Ivoire.23 Individuals not living in such household were de facto excluded from the survey. They were probably more often rural and poorer and may have had different testing behaviors. However, such bias should have been attenuated due to the calibration of our sample on the population distribution obtained at the last population census.
Overall refusal rate and execution failure rate was similar to other phone survey conducted in Côte d'Ivoire or in France.24,25 The high number of execution failures were probably due to unassigned numbers since we did not have access to the list of the number assigned by the national phone providers. The high number of recalls (i.e., at least 15 attempts), the different recall times or days, and the possibility to schedule an appointment probably limited the number of noncontacts. With the iterative poststratification, which corrected selection biases, our results are potentially generalizable at national level.
Recall biases may have underestimated test offer proportions and occurrence of some events. Regarding test offer, some patients may not have understood the offer made to them. A study carried out in several general medical services showed that some health care professionals prescribed the test without informing the patient orally,5 which could reinforce this phenomenon. This explains the association between educational level and high offer proportion in antenatal consultation; women with a high level of education are more likely to be aware of the proposal made to them. However, the fact some people may have been tested without knowing is not a recommended practice and should be seen as a missed opportunity, considering that learning one's HIV status is the entry door for linkage to HIV care and linkage to prevention. To our knowledge, individuals remain poorly informed about situations where a test is recommended, which would require better communication through health education materials for instance.
While our study considered the demand dimensions at each stage of the cascade, the survey did not consider certain factors such as the characteristics of the professionals consulted and the health facilities visited. The survey only collected the type of structure (private or public) for antenatal visits to women and for consultations for STIs, but the type of structure was not associated with the testing offer in either case (data not shown).
In the HIV testing cascades, return of results were not documented but this process is routinely done. Rapid tests with same-day result are now widespread in Côte d'Ivoire and national health statistics show that almost 100% tested individuals get their results.26
One of the original features of our study is to document, at a national population level, the different steps of the HIV testing cascade following certain events when other studies tend to limit their results to 1 or 2 steps of the HIV testing cascade (eg, testing offer and testing acceptance). Our results provide an overview of the stages where the barriers to HIV testing are greatest by event type.
Another original feature of our study is describing the HIV testing cascade in case of marriage, where the routine offer for testing has been recommended in law since 2014.7 Only 20% of those who got married in the last 5 years have performed a premarital health check-up, but test offer and test acceptance were high in case of premarital check-up. The nature of the marriage (civil or customary) had no effect on the frequency of premarital health check-up (results not shown). Doing a premarital health check-up, and more specifically premarital HIV testing, is commonly encouraged by religious leaders at a local or regional level, which may explain the prevalence of premarital testing among individuals of Catholic faith.27–29 This influence of religious leaders regarding premarital testing could be used in other recommended events, such as the presence and testing of male partners during antenatal consultations.
The factors associated with having consulted a health care professional were very different according to the events, which reflect the singularity of the different events studied. Nevertheless, common geographical (eg, area of residence) and/or accessibility (eg, currently working) barriers were found for most of the events studied. This point highlights the unequal access to health care services as observed in other contexts in Sub-Saharan Africa.30
In Côte d'Ivoire, still 42% of women and 63% of men have never been HIV tested.23 This low coverage of screening in the general population reflects the insufficient testing access (to both PITC and voluntary testing) in Côte d'Ivoire. In fact, a large proportion of people has never been tested in our 4 populations studied (24.2% and 48.4% respectively for women and men with children, 32.0% for those who reported an STI and 44.1% for those who got married), suggesting that access to HIV testing has remained limited. Thus, these missed opportunities for screening during these events are all the more damaging as they could contribute to reducing the large proportion of people in Côte d'Ivoire who do not know their HIV status.1
Regardless of the 4 events studied (pregnancy, STI episodes and marriage), test coverage was mainly dependent on access to health care professionals and an HIV testing offer during consultations.
Improving testing coverage in these relevant events will require improved access to health care services through better communication—especially for patients with STIs— and better accessibility to health care services. A systematic invitation by health care professionals to male partners to attend antenatal consultations is also recommended in order to stimulate their presence.
The HIV testing offer proportions was very low, especially among men present at antenatal consultations and among those consulting for an STI. The test offer seems to be more related to the practices and representations of health care professionals than to the characteristics of patients. Further research is needed to document more accurately the obstacles to PITC among health care professionals.
Considering testing gap in the general population of Côte d'Ivoire, it is crucial not to miss testing opportunities offered by the 4 life events studied.
1. UNAIDS. Knowledge is power
. Geneva: UNAIDS, 2018:32.
2. WHO, UNAIDS. Guidance on Provider-Initiated HIV Testing and Counselling in Health Facilities. Geneva: World Health Organization, 2007.
4. Ministère de la Santé et de l'Hygiène Publique [Côte d'Ivoire]. Document de Politique, Normes et Procedures Des Services de Dépistage Du VIH En Côte d'Ivoire, Édition 2014
. Abidjan: Ministère de la santé, 2014.
5. Carillon S, Bekelynck A, Assoumou N, et al. « Il y a Des Conseillers Communautaires Payés Pour Ça ! » Les Réticences Des Soignants à La Proposition Systématique d'un Test VIH en Consultation de Médecine Générale. Le Cas de La Côte d'Ivoire
. Abidjan: Presented at the International Conference on AIDS and STIs in Africa (ICASA) 2017, 2017.
6. PEPFAR. Côte d'Ivoire, Country Operational Plan (COP) 2014, Strategic Direction Summary
7. Présidence de la République [Côte d'Ivoire]. Article 9. de La Loi N°2014–430 Portant Régime de Prévention, de Protection et de Répression En Matière de Lutte Contre Le VIH et Le Sida
8. Inghels M, Carillon S, Desgrees du Lou A, et al. Effect of organizational models of provider-initiated testing and counseling (PITC) in health facilities on adult HIV testing coverage in sub-Saharan Africa. AIDS Care 2019; 1–7.
9. Berkson J. Application of the logistic function to bio-assay. J Am Stat Assoc 1944; 39:357–365.
10. Lumley T. Analysis of complex survey samples. J Stat Softw 2004; 9:1–19.
11. Programme National de Lutte contre le SIDA [Côte d'Ivoire]. Document de Politique, Normes et Procédures Des Services de Dépistage Du VIH En Côte d'Ivoire, Edition 2016
. Abidjan: Ministère de la santé, 2016.
12. Orne-Gliemann J, Balestre E, Tchendjou P, et al. Increasing HIV testing among male partners. AIDS 2013; 27:1167–1177.
13. Byamugisha R, Tumwine JK, Semiyaga N, et al. Determinants of male involvement in the prevention of mother-to-child transmission of HIV programme in Eastern Uganda: a cross-sectional survey. Reprod Health 2010; 7:12.
14. Desgrées-du-Loû A, Brou H, Djohan G, et al. Beneficial effects of offering prenatal HIV counselling and testing on developing a HIV preventive attitude among couples. Abidjan, 2002–2005. AIDS Behav 2009; 13:348–355.
15. Ford CA, Millstein SG, Halpern-Felsher BL, et al. Influence of physician confidentiality assurances on adolescents' willingness to disclose information and seek future health care. A randomized controlled trial. JAMA 1997; 278:1029–1034.
16. Boltena AT, Khan FA, Asamoah BO, et al. Barriers faced by Ugandan university students in seeking medical care and sexual health counselling: a cross-sectional study. BMC Public Health 2012; 12:986.
17. Morreale MC, Kapphahn CJ, Elster AB, et al. Access to health care for adolescents and young adults. J Adolesc Health 2004; 35:342–344.
18. PHIA project. Côte d'Ivoire Population Based HIV Impact Assessment : CIPHIA 2017–2018. Summary Sheet : Preliminary Finding
19. Leon N, Lewin S, Mathews C. Implementing a provider-initiated testing and counselling (PITC) intervention in Cape town, South Africa: a process evaluation using the normalisation process model. Implement Sci 2013; 8:97.
20. Gott M, Galena E, Hinchliff S, et al. “Opening a can of worms”: GP and practice nurse barriers to talking about sexual health in primary care. Fam Pract 2004; 21:528–536.
21. Gott M, Hinchliff S, Galena E. General practitioner attitudes to discussing sexual health issues with older people. Soc Sci Med 2004; 58:2093–2103.
22. Desgrées du Loû A, Brou H, Djohan G, et al. Refusal of prenatal HIV-testing: a case study in Abidjan (Côte d'Ivoire). Sante 2007; 17:133–141.
23. Ministère de la Santé et de la Lutte contre le Sida [Côte d'Ivoire], Institut National de la Statistique (INS), MEASURE DHS. Enquête Démographique et de Santé et à Indicateurs Multiples (EDS-MICS) 2016
. Calverton: INS & ORC Macro, 2017.
24. Beltzer N, Saboni L, Sauvage C, et al. Les Connaissances, Attitudes, Croyances et Comportements Face Au VIH / Sida En Ile-de-France En 2010 : Situation En 2010 et 18 Ans d'évolution
. Observatoire régional de santé d'île-de-France, 2011.
25. Larmarange J, Kassoum O, Kakou É, et al. Faisabilité et représentativité d'une enquête téléphonique avec échantillonnage aléatoire de lignes mobiles en Côte d'Ivoire. Population 2016; 71:121–134.
26. Ministère de la santé et de l'hygiène publique [Côte d'Ivoire]. Rapport Annuel Sur La Situation Sanitaire 2017
. Abidjan: Ministère de la santé, 2018.
27. Rennie S, Mupenda B. Ethics of mandatory premarital HIV testing in Africa: the case of Goma, Democratic Republic of Congo. Dev World Bioeth 2008; 8:126–137.
28. Umar SA, Oche OM. Knowledge of HIV/AIDS and use of mandatory premarital HIV testing as a prerequisite for marriages among religious leaders in Sokoto, North Western Nigeria. Pan Afr Med J 2012; 11:27.
29. Alswaidi FM, O'Brien SJ. Premarital screening programmes for haemoglobinopathies, HIV and hepatitis viruses: review and factors affecting their success. J Med Screen 2009; 16:22–28.
30. Rispel LC, de Sousa CA, Molomo BG. Can Social Inclusion Policies Reduce Health Inequalities in Sub-Saharan Africa?—A Rapid Policy Appraisal. J Health Popul Nutr 2009; 27:492–504.