Article Text
Abstract
Objective Our study aimed at estimating the prevalence of inconsistent condom use and at identifying its determinants in steady partnerships among people living with HIV/AIDS (PLWHA) in Cameroon.
Methods Analyses were based on data collected during the national cross-sectional multicentre survey EVAL (ANRS 12-116), which was conducted in Cameroon between September 2006 and March 2007 among 3151 adult PLWHA diagnosed HIV-positive for at least 3 months. The study population consisted of the 907 survey participants who reported sexual activity during the previous 3 months, with a steady partner either HIV-negative or of unknown HIV status. Logistic regression was used to identify factors associated with individuals' report of inconsistent condom use during the previous 3 months.
Results Inconsistent condom use was reported by 35.3% of sexually active PLWHA. In a multivariate analysis adjusted for socio-demographic characteristics, not receiving antiretroviral therapy (OR (95% CI): 2.28 (1.64 to 3.18)) was independently associated with inconsistent condom use.
Conclusions The prevalence of unsafe sex remains high among sexually active PLWHA in Cameroon. Treatment with antiretroviral therapy is identified as a factor associated with safer sex, which further encourages the continuation of the national policy for increasing access to HIV treatment and care, and underlines the need to develop counselling strategies for all patients.
- Africa
- condoms
- HIV
- risk behaviours
Statistics from Altmetric.com
Introduction
HIV transmission through unsafe heterosexual intercourse remains predominant in Sub-Saharan Africa. However, socio-economic determinants of inconsistent condom use among people living with HIV and AIDS (PLWHA) in this region have been rarely studied using representative samples, and never in the context of a large scaling-up programme.1–4 Moreover, the degree of impact of access to ART on sexual behaviours in developing countries still remains a matter of debate.1 5–8 In 2006, a cross-sectional survey was conducted on a representative sample of PLWHA attending HIV-treatment centres in Cameroon, which is among those western and central African countries most affected by the AIDS epidemic.9 This survey was set up in the context of the national programme for improving access to ART through the decentralisation of HIV care. On the basis of data collected, we aimed at determining factors associated with PLWHA's report of inconsistent condom use with their steady partner either HIV-negative or of unknown HIV status during the previous 3 months.
Patients and Methods
Study setting
According to the WHO Report on the global AIDS epidemic,9 500 000 adults were living with HIV in Cameroon in 2007. The national cross-sectional survey EVAL (ANRS 12-116) was conducted in Cameroon between September 2006 and March 2007 among a random sample of HIV-infected individuals aged 21 or older and diagnosed HIV-positive for at least 3 months, followed up in 27 HIV-treatment centres located in six provinces around the country. The complete design of the EVAL survey is detailed elsewhere.10
Sample size and study population
The survey was presented to 3488 eligible patients attending healthcare structures, among whom 3170 (91%) agreed to participate, and 3151 effectively completed the survey.
We focused our analysis on patients who reported sexual activity in a regular partnership during the 3 months prior to the survey, with a steady partner who was either HIV-negative or of unknown HIV status (study population).
Data collection and definition of variables
Socio-economic and psychosocial data (including condom use) were collected using face-to-face interviews with patients. Clinical data were obtained from medical records. The survey also involved a qualitative research section.
Definition of inconsistent condom use
As the pattern of factors associated with unsafe sex may be different in regular and in casual partnerships, analysis was restricted to condom use with the steady partner. Inconsistent condom use was defined as reporting to have used condoms ‘almost always,’ ‘sometimes’ or ‘never’ with one's steady partner (either HIV-negative or of unknown HIV status) during the previous 3 months.
Variables
Socio-demographic and economic characteristics
Patients' gender, social conditions, age, area of habitation and level of education were included in the analyses. Participation in an association of PLWHA was also considered, as well as alcohol abuse (binge drinking), characterised by having drunk more than three large bottles of beer and/or more than six glasses of other alcohol drinks on one occasion. Monthly household income per adult equivalent calculated from patients' reports was taken as an economic indicator of their standard of living.
Characteristics related to living in a couple, sexual activity and parenthood
Living in a couple was defined as marriage or free union, and sharing or not the same home. Disclosure of seropositivity to the main partner, frequency of sexual relationships, casual partnership, history of forced sexual relationships, parenthood (having at least one biological child, the desire to have a/another child) and the presence of children in the household were also considered.
Health-related quality of life and self-reported symptoms
Physical and mental health-related quality of life (HRQL) of patients was evaluated using two scores ranging from 0 to 100, with higher values denoting better HRQL. These scores were calculated from patients' answers to the Medical Outcome Study 12-Item Short Form Health Survey.11 Patients were also asked if, during the previous 4 weeks, they had experienced any symptoms from a specific list derived from the self-completed HIV symptom index developed by Justice et al12 and which comprised both general symptoms and symptoms related to changes in body shape.
Knowledge concerning antiretrovirals
Patients were asked a series of five questions about their knowledge concerning antiretrovirals' (ARVs) action on HIV and the need for pursuing these treatments on the long-term (table 1).
Clinical and treatment-related characteristics
The following clinical and treatment-related characteristics were considered: time since HIV diagnosis, CDC clinical stage, treatment received, CD4 cell count at last assessment, information on the patient's initiative to have an HIV test, ART history, healthcare services utilisation during the previous 6 months and adherence to ART (defined as no self-report of ART interruption for more than 2 days in the previous 4 weeks).
Statistical analyses
Logistic regression models were used to identify factors associated with inconsistent condom use among patients who reported sexual activity in a regular partnership during the previous 3 months, with a steady partner either HIV-negative or of unknown HIV status. All characteristics of patients were tested, except those binary variables which concerned less than 5% of the study population. Factors with a p value <0.20 in bivariate models were considered eligible for the multivariate model. A forward stepwise procedure with an entry threshold fixed at 0.05 was used to select statistically significant factors in the multivariate analysis and to determine the final model. A forward selection procedure13 involves starting from an empty model, then computing the χ2 statistic for each variable not in the model and examining the largest of these statistics. If it is significant at the α=0.05 level, the corresponding variable is added to the model. The process is repeated until none of the remaining variables meet the level for entry. With the stepwise approach, variables included in the model do not necessarily remain. Indeed, at each step, results of the Wald test for individual parameters are examined, and the least significant variable that does not meet the α=0.05 level for staying in the model is removed. ORs estimates in the final model were systematically adjusted for the frequency of sexual relationships, as a marker of individuals' exposure to HIV transmission risk. Two sensitivity analyses were performed after consistently grouping the ‘do not know’ answers with ‘yes’ (first sensitivity analysis) or ‘no’ answers (second sensitivity analysis) for variables concerning ART knowledge. Statistical analyses were performed using the SPSS V.15.0 (SPSS, Chicago) and Intercooled Stata V.9 (StataCorp LP, College Station, Texas) software packages.
Results
Characteristics of the study population
Among the 3151 survey participants, 1725 (54.7%) reported no sexual activity during the previous 3 months (of whom 72% had no main partner). Among the remaining 1426 patients, 91 (6.4%) had no main sexual partner or no sexual activity with their main partner during the previous 3 months, and 428 (30.0%) had a seroconcordant main partner. The study population consisted of the remaining 907 patients (63.6%). Socio-demographic, economic, clinical and psychosocial characteristics of these patients are described in table 1.
Factors associated with inconsistent condom use with one's steady partner either HIV-negative or of unknown HIV status
A total of 320 patients (35.3%) engaged in inconsistent condom use with a steady partner either HIV-negative or of unknown HIV status in the previous 3 months. The percentages of patients who reported inconsistent condom use were not significantly different between men and women (33.8% and 36.0%, respectively, p=0.50). Characteristics of the study population according to patients' report of inconsistent condom use are presented in table 1.
Bivariate analyses
In bivariate analyses, not receiving ART at the time of the survey was among the characteristics most significantly (p<0.0001) and strongly (crude OR (95% CI): 2.44 (1.78; 3.33)) associated with patient report of inconsistent condom use (table 1). In close relationships with access to ART, having been diagnosed HIV-positive more recently, doing less than 1 visit to hospital every 2 months, as well as reporting more general symptoms and impaired physical HRQL were also significantly associated with reporting inconsistent condom use. In addition, believing that ARVs stabilise the development of the HIV illness was significantly associated with a lower risk of inconsistent condom use, while believing that ART can prevent HIV transmission or that there is no further need to take the medicine for a treated person who feels better was associated with a higher risk of inconsistent condom use. The other variables significantly associated with inconsistent condom use concerned mainly patients' social status and network (including: being a woman not at the head of the household, not being a member of an association of PLWHA, living in a couple) as well as their sexual activity and couple relationships (including: living in a couple, having a casual sexual partner, having the desire to have a child).
Age, income level, frequency of sexual relationships, adherence to ART, last CD4 count and mental HRQL were not significantly associated with inconsistent condom use.
Multivariate analysis
In the final multivariate model, not receiving ART at the time of the survey was confirmed as a factor significantly associated with patient report of inconsistent condom use (table 2). However, time since HIV diagnosis, frequency of visits to hospital, number of symptoms reported, physical HRQL as well as report of the desire to have a child and variables related to ART knowledge were no longer significant after multivariate adjustment.
The results of the two sensitivity analyses confirmed the stability of the multivariate model, as the pattern of factors identified as being independently associated with inconsistent condom use was essentially the same (except for the variable ‘Someone who takes ARVs cannot transmit the virus,’ which was no longer significant when the ‘do not know’ answers were grouped with the ‘no’ answers).
Discussion
The present study underlines the high frequency of inconsistent condom use among sexually active HIV-infected outpatients in Cameroon. Results of the multivariate analysis primarily highlight that not receiving ART is associated with a higher risk of inconsistent condom use, even after controlling for socio-demographic and economic characteristics, characteristics related to living in a couple, sexual activity and parenthood, clinical characteristics, HRQL and self-reported symptoms, and patients' knowledge concerning ARVs (whatever the coding used for this latter group of variables).
This study is the first in Cameroon and, to our knowledge, in Sub-Saharan Africa, to explore the association between ART and unsafe sexual behaviours among a large sample of PLWHA, representative of HIV-infected outpatients followed up in the Cameroonian healthcare system. However, it is substantially limited by its cross-sectional design, which makes it impossible to evaluate pre- to post-ART changes in patients' sexual behaviours. In addition, it focuses only on clinical attenders, which are not representative of the whole population of PLWHA, contrary to population-based samples.2 Another limitation is that data on sexual behaviours are only based on patients' declarations, which may be affected by social desirability bias.14
Nevertheless, results confirm those of previous studies conducted in Côte d'Ivoire1 and in South Africa,3 15 in which 30–44% of the patients included reported unprotected sex. In addition, the association found between not receiving ART and reporting inconsistent condom use is in accordance with the conclusion of the meta-analysis by Crepaz et al in developed countries, which underlines that treated patients do not exhibit increased sexual risk behaviours.16 In resource-limited settings, most of the studies exploring this issue have also led to the same conclusion.1 3 5 6 8 17 These studies focused either on patients followed-up in a given HIV centre3 6 or on patients living in specifically delineated geographical areas.1 5 17 The present study therefore reinforces these findings on a sample of PLWHA followed up in HIV care centres located in different regions of Cameroon, of which some are decentralised. However, it contradicts the results of the study of Diabate et al conducted in Côte d'Ivoire7 among 312 non-treated and 303 ART-initiating patients. This study shows a significant increase in unsafe sexual behaviours among treated patients after 6 months of ART, as compared with a relative stability of such behaviours among non-treated patients. This is, to our knowledge, the only study to document an increase in unsafe sexual behaviours after ART initiation in a resource-limited setting. Such an increase has already been described in high-resource settings, but most often on specific subpopulations such as men who have sex with men18 or young people.19 Nevertheless, the discrepancy of results between the longitudinal study by Diabate et al and the present cross-sectional study points out the need for further analyses on longitudinal data so as to verify the stability of findings after taking into account the temporal dimension.
In the present study, the lower odds of reporting inconsistent condom use among ART-treated patients may be partly due to better information on prevention strategies among ART-treated patients, as messages of this type are essentially given at the time of diagnosis and at initiation of therapy. We may also hypothesise that the psychosocial support offered to patients during ART, notably in terms of counselling on adherence, creates favourable conditions not only for better adherence to treatments, but also for safer behaviours in general. This is suggested by the results of a recent study in Kenya, which found less sexual risk behaviours among 179 ART-treated patients than among 143 patients only receiving preventive therapy with cotrimoxazole/isoniazid.17 Both groups of patients had the same schedule of visits in HIV centres, the only difference being the presence of intensive counselling on treatment adherence for ART-treated patients.
Results of the present study show lower odds of reporting inconsistent condom use among patients who were members of PLWHA's associations. This emphasises the need for widespread information programmes on HIV transmission and highlights the importance of coupling ART with counselling programmes aimed at encouraging patients to adopt and maintain safer sexual practices. On the contrary, results show higher odds of reporting inconsistent condom use among both women who were not the head of their household and individuals who were living in a couple. This confirms preliminary results from qualitative interviews of PLWHA included in the EVAL survey which show the social pressure experienced by some women (fear of being exposed to violence from their partner, or being disowned). This also underlines women's vulnerability and lack of power to negotiate condom use with their partner.20 More generally, this highlights the difficulties associated with negotiating condom use in the context of a stable relationship.4 Results also confirm that a low educational level and rural area of habitation are associated with a higher risk of inconsistent condom use,21 22 and enlighten the negative impact of alcohol abuse, as already pointed out in other studies in developing countries.1 3 7 23–27 Finally, results show that PLWHA who reported believing that ART-treated individuals cannot transmit HIV also reported more frequently inconsistent condom use. Interestingly, such beliefs were significantly more frequent among non-treated PLWHA (results not shown), which suggests a lack of information about the mechanisms of actions of ARVs in this latter group of patients. However, interpretation must remain cautious, as this finding was not confirmed in the sensitivity analysis, after grouping the ‘do not know’ answers with the ‘no’ answers.
In conclusion, the present study brings an encouraging message in the context of the Cameroonian national programme for improving access to ART. However, findings need to be confirmed on longitudinal data, such as those being currently collected in the STRATALL (ANRS 12-110) research project. Findings suggest the need for developing counselling interventions at each step of the follow-up of PLWHA, especially for patients not receiving ART, who often have fewer contacts with healthcare structures. Behavioural interventions could also be planned to help PLWHA increase their skills in negotiating safer sex, controlling sexual situations, and communicating effectively with sexual partners. Finally, as hospitals in resource-limited settings face a large increase in their HIV caseload each day, attention should be paid to involving external resources such as patients' associations when developing both medical and psychosocial support of PLWHA.
Key messages
Unsafe sex remains frequent among sexually active people living with HIV/AIDS in Cameroon.
Receiving antiretroviral therapy is associated with less frequent report of inconsistent condom use.
Counselling interventions should be developed at each step of the follow-up of HIV-infected patients, especially for non-treated patients, who have fewer contacts with healthcare structures.
Acknowledgments
We would like to thank the French National Agency for Research on AIDS and viral hepatitis (ANRS) which funded this research and the Cameroonian Ministry of Health for its support in the implementation of the EVAL survey. We would like to thank also all the patients, who agreed to participate in the survey, as well as the healthcare providers, for their strong involvement in the data-collection stage. The EVAL Study Group: S. Koulla-Shiro (Central Hospital, Yaoundé, Cameroon), P. Ongolo-Zogo (Ministry of Public Health - Division of Health Operations Research, Yaoundé, Cameroon), J. Blanche, A-D Bouhnik, S. Boyer, M-P Carrieri, A. Dia, F. Eboko, S. Loubière, F. Marcellin, J-P Moatti, Y. Obadia, C. Protopopescu, B. Spire (INSERM, IRD, University of the Mediterranean UMR 912, Marseilles, France), S-C Abega, C. Abé, P. C. Bilé, C. Bios, R-C Bonono, Y. Mehe, M. T. Mengue, H. Mimcheu, F. Mounsade, L. M. Ngaba, J. Ngo Mbog, S. Ngo Yebga, H. Nkwidjan (IRSA, catholic University of Central Africa, Yaoundé, Cameroon), R. Nantchouang (GERCIS, catholic University of Central Africa, Yaoundé, Cameroon).
References
Footnotes
Funding The French National Agency for Research on AIDS and viral hepatitis (ANRS)101, rue de Tolbiac, 75013 Paris—France funded the EVAL survey (ANRS 12-116) research programme and the Cameroonian Ministry of Public Health supported its execution.
Competing interests None.
Ethics approval Ethics approval was provided by the Cameroonian national ethics comittee.
Provenance and peer review Not commissioned; externally peer reviewed.
Linked Articles
- Whistlestop tour