Secondary Logo

Journal Logo

Original Article

Adherence to HAART and its principal determinants in a cohort of Senegalese adults

Lanièce, Isabellea; Ciss, Mouniroub; Desclaux, Alicec; Diop, Karimd; Mbodj, Fatoue; Ndiaye, Barrad; Sylla, Omare; Delaporte, Ericf; Ndoye, Ibrahimae

Author Information



Adherence to multi-drug antiretroviral regimens has been a focus of attention since their introduction, owing to their complexity, frequent adverse effects, and chronic nature. Recent cohort follow-up studies in industrialized countries have identified several obstacles to adherence, and shown that a high level of adherence is required to delay disease progression effectively [1–3]. However, adherence is difficult to measure, and is known to vary over time [4]. Few data have been published on adherence to multi-drug antiretroviral therapy (ARV) in the context of African drug access programmes [5].

In industrialized countries, where treatments are usually provided free of charge, measures adopted to improve adherence to treatment include counseling, patient education programmes and telephone hot-lines. These measures are seldom feasible on a large scale in poor countries [6]. Studies are therefore needed to determine the level of adherence and to identify the determinants of high adherence in the African context. This knowledge would help to define relevant, efficient and acceptable adherence support measures for patients within the African health system.

In November 1999, we started a pilot project designed to assess adherence and causes of treatment interruption among patients treated through the Senegalese Antiretroviral Access Initiative (ISAARV). This programme had set up access to treatment conditions and adherence support measures fitted to low-income countries [7].

The project comprised both a quantitative study of descriptive and analytical epidemiology, first results of which are reported here and a qualitative socio-anthropological survey, the preliminary results of which have been published elsewhere [8,9].

Patients and methods


The first 180 adults enrolled in the ISAARV prospective observational cohort were eligible for this study if they had been monitored medically for at least 30 days, during the 24 months of observation (November 1999 to October 2001). Eighty patients were receiving ARV within clinical trials [40 patients in French National Agency for Research on AIDS (ANRS) 12-04 project and 40 patients in ANRS 12-06 project] and the remaining 100 patients were enrolled in the ARV access programme. The two groups differed principally by their date of enrollment, their clinical stage at enrollment, their previous exposure to ARV, the type of ARV regimen prescribed, and their contribution to the cost of their treatment.

The patients were monitored in three Dakar health structures (Internal Medicine Unit of Hôpital Principal, and the Infectious Diseases Unit and Ambulatory Treatment Center of Fann Hospital). They all obtained their drugs from a single dispensing site.

Adherence assessment

Adherence to treatment was assessed during regular monthly visits to the dispensing pharmacy. The dispensing procedures and support measures (counseling, access to discussion groups, social and financial support) were the same for all the patients.

Adherence data were collected by the dispensing pharmacist, who interviewed each patient before dispensing each month's treatment, based on a questionnaire containing mainly closed questions. The pharmacist also discussed with the patient any discrepancy between declared adherence and the number of doses returned unused. Adherence was calculated as the ratio between the stated number of tablets taken and the number of tablets prescribed, expressed as a mean percentage for the different components of each multi-drug regimen.

Treatment interruptions for medical reasons were excluded from the definition of non-adherence.

Some missing information due to irregular visits to the pharmacy was subsequently obtained by the prescriber or a social worker, who contacted the patient concerned.

Statistical analysis

Epi Info 6.04 (Centers for Disease Control and Prevention, Atlanta, Georgia, USA) and Stata 6.0 (Stata Corporation, College Station, Texas, USA) were used to record and analyze the data. Adherence between different subgroups (defined by enrollment in a clinical trial, cost or type of regimen, etc.) at a given time point of follow-up was compared by using the Kruskal–Wallis test for quantitative variables. Adherence between two time points was compared by using the Wilcoxon test for paired data. The level of significance was set at P < 0.05.

Ethical considerations

Ethical approval was received from the Committee of the AIDS Control National Program. Patients with poor adherence were referred to the Welfare Committee, and were offered appropriate support.



A total of 167 patients met the eligibility criteria. Adherence data were collected among 158 of them (94.6%) as one patient refused to participate, two did not collect their treatment personally and four did not provide any data due to a short follow-up. Follow-up of these 158 patients during the 24-month study period yielded a total of 2752 patient-months of observation; adherence data were available for 2389 patient-months (86.8%); the total treatment period was 2471 patient-months. Eleven deaths (7%) and three drop-outs (2%) occurred during the study period. The median length of follow-up in the ISAARV programme was 21 months for the 158 patients of whom 80 were included in clinical trials

Baseline characteristics

The study population consisted of 84 men and 74 women (M : F sex ratio 1.1 : 1), with a mean age of 38 years. The CDC stage distribution at the outset of antiretroviral treatment in the ISAARV programme (155 patients) was as follows: 6% stage A, 39% stage B, and 55% stage C. The infection was due to HIV-1 in 97% of cases, HIV-2 in 1% of cases, and HIV-1+2 in 2% of cases. At enrollment, the mean viral load (log10) was 5.34 copies/ml (n = 154) and the mean CD4 cell count was 156 × 106 cells/l (n = 154).

Forty-four per cent of the patients were married and 15% were widowed. The mean number of children per patient was 2.6. Thirty-two per cent of the patients had never been to school, and 41% were not in paid employment. The median monthly income was 15 000 FCFA (20 US$).

Ninety-three per cent of patients were antiretroviral-naive at inclusion. The intent to treat regimen was a dual therapy comprising two nucleoside reverse transcriptase inhibitors (NRTI) in 4% of cases and a three-drug regimen in 96% of cases, including two NRTI and one protease inhibitor (PI) in 43% of cases, and two NRTI and one non-nucleoside reverse transcriptase inhibitor (NNRTI) in 53% of cases. The intent to treat regimen prescribed to the patients participating in clinical trials consisted of two NRTI and one NNRTI [efavirenz (EFZ)].

Treatment regimens prescribed during the study period

Most antiretroviral treatments prescribed during the 24-month study period were multi-drug combinations, including stavudine (D4T)/didanosine (ddI)/indinavir (IDV) (26%) and lamivudine (3TC)/ddI/EFZ (30%). Treatment switches were sometimes necessitated by concurrent antituberculous treatment, an adverse effect, or, in rare cases, by temporary unavailability of a low- dose formulation of stavudine (Zerit® 15 or 20).


Mean adherence rate

The mean adherence among the 158 patients during the 24-month study period was 91% [median, 100%; interquartile range (IQR), 97–100%]. The patients stated that they had taken the entire monthly dose during 69% of the months covered by the study period.

Changes in adherence during the study period

Mean adherence was 90% during the first year (median, 100%; IQR, 97–100%) and 92% in the second year (median,100%;IQR,98–100%). Mean adherence always remained above 80%, oscillating between 83 and 95% according to the month.

Adherence tended to be better, with a smaller dispersion of values, among the 80 patients included in the clinical trials than among the remaining 78 patients (97 versus 87%). This difference diminished with time (Fig. 1), being statistically significant every month during 17 of the 24 study months.

Fig. 1
Fig. 1:
Changes in mean adherence between November 1999 and October 2001 in the different patients' categories.

Adherence improved between October 2000 and April 2001 among patients not included in clinical trials (P = 0.02) and declined between October 2000 and October 2001 among patients participating in trial ANRS 12-04 (P = 0.03).

Determinants of adherence

This analysis of factors influencing adherence focused on treatment-related factors, namely the duration of antiretroviral treatment, the patient participation towards the cost of treatment, and the treatment regimen. This choice was guided by the strategic decisions to be taken in the process of generalizing access to HAART in Senegal.

Treatment duration

A difference in adherence emerged between patients participating in clinical trials and other patients at month 6 (P = 0.001) and remained at months 12 and 18 (P = 0.04 and P = 0.003, respectively). It was difficult to identify overall temporal trends, owing to the different follow-up periods among the three patient subpopulations related to a 33-month inclusion period, leading to few observations at certain months of treatment.

Patients' financial participation

Patients enrolled in the two clinical trials were treated free of charge, whereas the other patients paid monthly between 0 and 198 000 FCFA (264 US$) towards the cost of their treatment.

Mean adherence among patients receiving D4T/ddI/IDV outside the clinical trial setting decreased as their financial participation increased. This trend was noted during the first year of the study and, albeit less so, during the second year (Table 1). During year 1, most patients made monthly a substantial minimum payment of about 21 000 FCFA (28 US$), while the legal minimum wage in Senegal at that time was 36 250 FCFA (48.3 US$) a month. Thus a large proportion of the patients encountered financial difficulties. Following the antiretroviral drug price reduction that occurred early in year 2, the mean contribution made by patients already on treatment was cut four-fold, and the minimum participation was cancelled.

Table 1
Table 1:
Mean adherence according to patient's monthly financial contribution among patients receiving a three-drugs regimen including stavudine/didanosine/indinavir.

This sharp decrease in the sum the patients who were not participating in clinical trials had to pay towards their treatment probably contributed to the improvement in adherence during year 2 (83% in year 1 versus 90% in year 2). This is supported by the statements made by these patients: financial difficulties were reported as the leading cause of treatment interruption during the first year, and as only the fifth cause during the second year.

Treatment regimen

IDV was the most widely prescribed PI in our cohort, being the only low-cost PI available in Africa; EFZ was the most widely prescribed NNRTI. Among the patients who were receiving their treatment free of charge, mean adherence was 89% with IDV and 97% with EFZ during the 24-month study period. The differences at months 6, 12 and 18 of treatment were close to statistical significance (P = 0.09, P = 0.17 and P = 0.05, respectively).

Relationship between adherence and virologic efficacy

At months 6, 12, 18 and 24, adherence during the previous month was compared to contemporary viral load values (logarithmic scale).

In the group of patients who were not participating in clinical trials and who were receiving a PI-containing, three-drug regimen, we compared mean viral load values between those with stated adherence of 90% or more and those with poorer adherence. As expected, viral load was higher in the less adherent patients (mean differences of 1.7 and 1.8 log10 copies/ml at months 18 and 24, respectively; P < 0.05).


Methodological considerations

The method and the unit period (30 days) chosen to assess adherence in this study require discussion.

There is no reference method to quantify adherence [10]. The context of a resource-limited setting adds limitations in the choice of adherence measurements' methods: for example, plasma drug assays and electronic pill boxes are unavailable and illiteracy rules out the use of self-rating questionnaires. In this African setting, it seems reasonable to estimate adherence on the basis of stated drug intakes and/or unused tablet counts, as these parameters are simple, inexpensive to assess and recognized as valid estimates. It is known that patients tend to overestimate their adherence but counts of unused tablets, an objective measure, helps to refine this subjective estimate [11,12]. The unavoidable approximations inherent in all procedures used to assess adherence, given the lack of a reference tool, have been widely described in the literature [10,13,14]. It is generally agreed that the correlation between adherence and the virologic response observed by some authors (and confirmed in this study) tends to validate the subjective assessment by patients receiving their first antiretroviral treatment [1,2,10,14–16]. The high stated level of adherence is also in keeping with the good immunovirological efficacy observed in this cohort [17–20]. Moreover, adherence data gathered by the prescribers were very consistent with those collected by the dispensing pharmacist [20].

The unit period of recall chosen here is relatively long (30 days). Stated adherence during the last 3 days of each 1-month period in this survey tended to be slightly poorer than the corresponding 30-day estimate (89 versus 91% during the 24-month study period). However, we felt that the use of a 30-day unit assessment period was more likely to reflect the distribution of causes for non-adherence, notably illness and travel.

Adherence data were unavailable for 386 patient-months of follow-up (14%). In nearly half these cases the patients had received their drugs but the corresponding datasheet had not been filled out (pharmacist or patient unavailable, or dispensing of more than 1 month of treatment). In these cases we postulated that the treatment had not been taken differently than in the documented months. During the remaining months (7% of all months of follow-up), the patients had not received the drugs from the pharmacy but may have had sufficient personal stocks.

Further analysis taking into account the correlation structure of the data generated by the repetition of the observations within each patient as well as multivariate methods to assess more potential determinants of adherence will deepen these first results.


The main result of this study is the high stated level of adherence in each patient category. Indeed, the patients declared that they had taken, on average, 91% of their dose during each month of follow-up, and the entire dose during nearly 70% of months. These results are similar to those obtained in follow-up cohorts in industrialized countries; for example, on the basis of self-rating questionnaires, 73.3% of respondents in the Aproco cohort stated that they had taken the entire dose during the previous 4 days at month 4; and 67% of respondents in the Ciel Bleu trial stated that they had taken 100% of their doses) [16,21]. A recent study conducted in an industrialized country suggests a lower adherence level among African patients [22]. Several factors may explain the good adherence observed in this Senegalese study, such as the experimental nature of the programme, recent inclusions in a small cohort, and the fact that most of the patients had never previously received ARV and were highly motivated (a large proportion of patients were symptomatic) [23].

The size of the patients' contribution to their treatment costs had a major impact on adherence. Treatment interruptions because of financial problems had been reported in other African studies [5,24–26]. In our study, financial obstacles led to lengthy treatment discontinuations until a more appropriate pricing structure was adopted. The increase in adherence between October 2000 and April 2001 among patients not included in clinical trials was probably largely due to the ARV cost reduction.

Adherence tended to be better with EFZ- than with IDV-containing regimens, and experience in industrialized countries suggests that this difference will persist in the post-trial period [27].

Adherence support programmes have to be extended in low income countries while remaining affordable, feasible and acceptable. In our experience, pharmacist counseling was a major measure, complementary to prescribers' counseling. Discussion groups, support from other persons living with AIDS were under implementation during the study period. Patients often expressed a fear of lack of confidentiality regarding treatment delivery and uptake. Some of them wished to get treatment for 2 months or more in order to reduce the frequency of contacts with the medical infrastructures, sometimes far from their homes. The high mean level of adherence observed along with patients' fears and wishes suggest that a programme proposing DOTS (directly observed treatment strategy) may be inadequate and may not be useful for a majority of patients in this context [28].

In conclusion, adherence to three-drug regimens was consistently high in this 24-month follow-up cohort study conducted in Senegal. These results provide evidence of high adherence capacities in people living with HIV/AIDS in Africa.

Two main factors were found to influence adherence, namely the cost of treatment and the type of drug combination. As in industrialized countries, simplified treatments (especially with NNRTI) appear to be better managed and better accepted. The cost of treatment must be adapted to the individual patient's financial resources (which are often nil) if the continuity of treatment is to be assured. These two points, and the measures set up to support adherence, should receive particular attention in the design of future ARV access programmes in Africa.


The authors are grateful to the multidisciplinary team involved in the Senegalese ART Access Program. They also thank the social workers of the Infectious Diseases Unit of Fann Hospital for their active collaboration, and the members of the Welfare Committee (especially Dr Khoudia Sow and Dr Bernard Taverne) for their technical advice.


1. Salmon-Céron D, Deleuze J, Coste J, Guerin C, Ginsburg C, Blancl P, et al.Enquête sur l'observance des traitements antirétroviraux comportant un inhibiteur de protéase chez les patients infectés par le virus de l'immunodéficience humaine.Ann Intern Med 2000; 151:297–302.
2. Paterson DL, Swindells S, Mohr J, Brester M, Vergis EN, Squier C, et al.Adherence to protease inhibitor therapy and outcomes in patients with HIV infection.Ann Intern Med 2000; 133:21–30.
3. Bangsberg DR, Perry S, Charlebois ED, Clark R, Zolopa AR, Moss AR. Adherence to HAART predicts progression to AIDS.Eighth Conference on Retroviruses and Opportunistic Infections, Chicago, 2001 [abstract 483].
4. Moatti JP, Spire B. Au-delà de l'observance. Les recherches socio-comportementales sur l'impact des multithérapies antirétrovirales.Transcriptase 2000; 83:21–25.
5. Weidle PJ, Malamba S, Mwebaze R, Sozi C, Rukundo G, Downing R, et al.Assessment of a pilot antiretroviral drug therapy programme in Uganda: patients' response, survival, and drug resistance.Lancet, 2002; 360:34–40.
6. Desclaux A, Sylla O, Laniece I, Mbodj F, Ciss M, Diop K. Le dispositif d'appui à l'observance de L'ISAARV. In: Desclaux A, Lanièce I, Ndoye I, Taverne B, editors. L'initiative sénégalaise d'accès aux médicaments antirétroviraux. Analyses économiques, sociales, comportementales et médicales. Paris: ANRS, Collection Sciences Sociales et Sida; 2002. pp. 119–128.
7. Desclaux A, Lanièce I, Ndoye I, Taverne B (eds). L'initiative sénégalaise d'accès aux médicaments antirétroviraux. Analyses économiques, sociales, comportementales et médicales. Paris: ANRS, Collection Sciences Sociales et Sida; 2002.
8. Sow-Ndiaye NK, Laniece I, Desclaux A, Taverne B, Faye-Niang MA, Badiane S, et al.Aspects of observance during antiretroviral therapy: preliminary study.XIIIth International Conference on AIDS, Durban, February 2000 [poster D4581].
9. Sow K, Desclaux A. L'adhésion au traitement antirétroviral. In: Desclaux A, Lanièce I, Ndoye I, Taverne B, editors. L'initiative sénégalaise d'accès aux médicaments antirétroviraux. Analyses économiques, sociales, comportementales et médicales. Paris: ANRS, Collection Sciences Sociales et Sida; 2002. pp. 129–139.
10. Costagliola D, Barberousse C. Comment mesurer l'observance? In: L'observance aux traitements contre le VIH/sida: mesure, déterminants, évolution, (edited by ANRS). Paris: ANRS, Collection Sciences Sociales et Sida; 2001. pp. 33–42.
11. Stone VE. Strategies for optimizing adherence to highly active antiretroviral therapy: lessons from research and clinical practise.Clin Infect Dis 2001; 33:865–872.
12. Moulin P. Enquêtes sur les pratiques des professionnels de santé en matière d'observance dans le domaine du VIH/sida en France: recherches, actions, évaluations, (edited by ANRS). Paris: ANRS, Collection Sciences Sociales et Sida; 2001. pp. 21–31.
13. Chesney MA, Morin M, Sherr L. Adherence to HIV combination therapy.Soc Sci Med 2000; 50:1599–1605.
14. Lui H, Golin CE, Miller LG, Hays RD, Beck CK, Sanadaji S, et al.A comparison study of multiple measures of adherence to HIV protease inhibitors.Ann Intern Med 2000; 134:968–977.
15. Gifford AL, Bormann JE, Shively MJ, Wright BC, Richman DD, Bozette SA. Predictors of self-reported adherence and plasma HIV concentrations in patients on multidrug antiretroviral regimens.J Acquir Immune Defic Syndr 2000; 23:386–395.
16. Spire B, Duran S, Souville M, Chêne G, Leport C, Raffi F, et al.L'observance des multithérapies des personnes infectées par le VIH: de l'approche prédictive à l'approche dynamique. In: L'observance aux traitements VIH/sida: mesure, déterminants, évolution, (edited by ANRS). Paris: ANRS, Collection Sciences Sociales et Sida; 2001. pp. 43–54.
17. Laurent C, Diakhaté N, Ngom Gueye NF, Toure MA, Sow PS, Faye MA, et al.The Senegalese government HAART initiative: an 18-month follow-up study of feasibility, effectiveness, adherence, toxicity and viral resistance.AIDS 2002; 16:1363–1370.
18. Landman R, Schieman R, Thiam S, Canestri A, Delaporte E, MboupS, et al.Once a day HAART regimen in treatment naïve HIV-1 infected adults in Senegal: ANRS 12-04/IMEA 011 study.XIIth international conference on AIDS and STDs in Africa, Ouagadougou, December 2001 [poster 10PT3-211].
19. Landman R, Canestri A, Thiam S, Delaporte E, Mboup S, Vray M, et al.First evaluation of D4T, DDI and EFZ in antiretroviral naïve patients in Senegal: ANRS 12-06 / IMEA 012 study.XIIth international conference on AIDS and STDs in Africa, Ouagadougou, December 2001 [poster 11PT3-234].
20. Laurent C, Ngom Gueye NF, Diakhaté N, Gueye PM, Diouf M, Lanièce I, Touré Kane NC, et al.Efficacité et tolérance du traitement antirétroviral dans le contexte de l'Initiative sénégalaise d'accès aux antirétroviraux. In: Desclaux A, Lanièce I, Ndoye I, Taverne B, editors. L'initiative sénégalaise d'accès aux médicaments antirétroviraux. Analyses économiques, sociales, comportementales et médicales. Paris: ANRS, Collection Sciences Sociales et Sida; 2002. pp. 143–155.
21. Bernard N, Goujard C, Peyramond D, Chwalow J, Cavailler P, Therme N, et al.Etude des déterminants de l'observance auprès de 324 patients sous traitement antirétroviral (protocole ciel bleu).Ier congrès national sur l'observance thérapeutique dans les maladies chroniques: vers de nouvelles stratégies, Nantes, Octobre 2000 [poster].
22. Frater AJ, Dunn DT, Beardall AJ, Ariyoshi K, Clarke JR, McClure MO, et al.Comparative response of African HIV-1-infected individuals to highly active antiretroviral therapy.AIDS 2002; 16:1139–1146.
23. Desclaux A. L'observance en Afrique: question de culture ou “vieux problème” de santé publique? In: L'observance aux traitements VIH/sida: mesure, déterminants, évolution, (edited by ANRS). Paris: ANRS, Collection Sciences Sociales et Sida; 2001. pp. 57–66.
24. Eholié SP, Bissagnene E, Koffi S, Konan-Kolo R. Evaluation de la trithérapie antirétrovirale à Abidjan (RCI): bilan de 24 mois de prescription.XIIth International Conference on AIDS and STDs in Africa, Ouagadougou, December 2001 [abstract book, p. 25; oral communication, 10DT3-2].
25. Macharia D. Reasons for modifying antiretroviral therapy in the private sector in Kenya.XIIth International Conference on AIDS and STDs in Africa, Ouagadougou, December 2001 [abstract book, p. 26; oral communication, 10DT3-4].
26. Drabo YJ, Minoungouu S, Ouedraogo M, Bambara M, Bonkoungou K, Diallo I. Traitements par les ARV: à propos de 45 patients suivis à Ouagadougou.XIIth International Conference on AIDS and STDs in Africa, Ouagadougou, December 2001 [abstract book, p. 168; poster, 11PT3-233].
27. Staszewski S, Morales-Ramirez J, Tashima KT, Rachlis A, Skiest D, Stanford J, et al.Efavirenz plus zidovudine and lamivudine, efavirenz plus indinavir, and indinavir plus zidovudine and lamivudine in the treatment of HIV-1 infection in adults.N Engl J Med 1999; 341:1865–1873.
28. Farmer P, Léandre F, Mukherjee JS, Claude MS, Nevil P, Smith-Fawzi MC, et al.Community-based approaches to HIV treatment in resource-poor settings.Lancet 2001; 358:404–409.

Adherence; Africa; antiretroviral; epidemiology; highly active antiretroviral therapy; HIV; Senegal

© 2003 Lippincott Williams & Wilkins, Inc.