Prevalence and determinants of adherence to antiretroviral treatment among HIV patients on first-line regimen: a cross-sectional study in Dakar, Senegal

Introduction Successful and long-term positive impact of antiretroviral treatment requires high rates of adherence (> 90%). In Senegal, there is a lack of data regarding adherence to antiretroviral treatment and only a few studies have looked at the determinants. The aim of this study is to assess the prevalence and determinants contributing to antiretroviral (ARV) adherence among Human Immunodeficiency Virus (HIV) infected outpatients receiving care at four public hospitals in Dakar, Senegal. Methods A cross-sectional based study was carried out among HIV-positive ART adults in Dakar, Senegal. Patients were systematically sampled during either their clinical visits or visit to collect ARV drugs from six public hospitals and data collected with a questionnaire. The study outcome was adherence to antiretroviral treatment assessed by a multiple approach method which combined three self-reported adherence tools: self-reporting, Visual Analog Scale (VAS), and the Simplified Medication Adherence Questionnaire (SMAQ). Data were entered with an Excel spreadsheet and transferred to STATA for descriptive, bivariate and multivariate analysis. All the statistical tests were done at the threshold level of 0.05. Results A total of 150 HIV-positive patients on first line ART regimen at six public health facilities were enrolled into the study. The mean age of patients was 43.1 years with a sex ratio of 0.3. Most of the patients were prescribed Tenofovir-based regimen. Of these patients, 26.67% were found to be highly adherent. After adjusting for health-related variables, demographic and socio-economic variables, better adherence was associated with participating actively within an association of persons living with HIV (AoR=2.89; 95% CI: 1.04 - 7.99; p value 0.041) while being widowed patient was associated with lower adherence (AoR=0.17; 95% CI: 0.03 - 0.94; p value 0.043). Conclusion Our study findings imply that adherence should be routinely assessed during medical visits. Ongoing strategies to improve adherence such as out-of-clinic group-based models or psychological support should be directed toward outpatients' clinics to assist in improving adherence and long term virologic suppression in Senegal.


Introduction
Successful and long-term positive impact of antiretroviral therapy requires high rates of adherence. There is a clear relationship between adherence to ARV treatment, viral load suppression, acquired drug resistance and treatment failure [1,2]. In the long term, non-adherent patients on tri-therapy are more likely to die than adherent patients on the same therapy [3,4]. In developing countries where older first-line therapies are being used, the development and transmission of drug-resistant strains of HIV will cause the switch to costly second line antiretroviral treatment and limit the treatment options available. Successful antiretroviral treatment requires sustaining high rates of adherence and is a complex behavior which is influenced by a wide range of factors. These factors have been previously categorized into socio-demographic, condition-related, treatment-related, patient-related, and interpersonal factors [5][6][7][8].
However, there is only few studies on the prevalence and the main determinants of adherence to antiretroviral treatments in the context of routine clinical care at the public health facilities. Since 1998, Senegal has launched an ambitious antiretroviral treatment programme with a free access to antiretroviral treatment and care for all HIV infected patients [9]; in 2010, there was more than 15,000 patients on antiretroviral treatment in more than 100 health facilities across the country and currently, the coverage rate for ARV among HIV adult patients in Senegal is estimated to be at 75% [10]. Since 2012, the country has adopted the WHO/UNAIDS "90-90-90" treatment target [11]. Thus, Senegal's National AIDS program will need to initiate nearly 36,450 HIV positive patients on ART in order to attain the objective of 90-90-90 during the next following years [10] because of this increase in the number of persons initiated on ART in the upcoming years, monitoring adherence will be the key strategy for the country to attain the third goal of the "90-90-90" approach and globally for the success of the ART program across the country. In Senegal, there is a lack of study investigating adherence to antiretroviral treatment in the context of routine public health care system. Most of the studies that have assessed adherence to ARVs among infants and adult patients have been implemented alongside clinical trials and have mainly recruited patients from highly specialized research centers. Thus, few of them have assessed adherence to antiretroviral treatment in the context of routine clinical care [12][13][14][15]. The aim of this study is to assess the factors contributing to ARV adherence after initiating ARVs at six public hospitals in Dakar, Senegal.  [16]. Currently, more than 15,000 patients are enrolled in care in these six (6) sites with more than 2,000 having had initiated ART. The treatment programme provides patients with access to counselling, antiretroviral treatment viral load monitoring and psychosocial support.

Methods
Eligibility criteria: we included HIV-positive patients who were on ART following the Senegalese antiretroviral treatment guidelines.
Eligible subjects were patients ≥18 years of age who initiated standard government first-line ART regimens of Tenofovir (TDF) or Zidovudine (AZT) with lamivudine (3TC) and either Efavirenz (EFV) or Nevirapine (NVP). We excluded patients who were referred from another facility or were hospitalized at the time of the study. Thus, our study only included out-patients who initiated on ART and who came for their scheduled antiretroviral (ARV) pickup or medical visits during the time of the study. For each patient, eligibility criteria were ascertained by a clinical pharmacist at each site.
Sample size determination: based on a literature review of previous studies, we hypothesized that the true proportion of HIVpositive patients highly adherent to antiretroviral treatment (Po) was at most 55%. Our study wanted to identify correctly with a power of 90% (zβ = 1.28) and for two-sided test and 5% significance (zα = 1.6449) a difference of at least 10% (P1: 60%), given that we used more stringent criteria which combine three self-reported adherence methods into on single multiple adherence measurement tool. Hence, we applied the Woodward formula [17]. Thus, the total patients that should be sampled was 173 (n = 173). Self-report questionnaire: in the self-report questionnaire, there are four questions on which the patient responded with either "yes" or "no". A patient who answers "no" to all four questions was recorded as highly adherent, but the one whose answer is "yes" to one of the items is recorded as moderately adherent. When a patient responds "yes" to two (2) or more questions, he or she was rated as poorly Visual analog acale (VAS): each patient was asked to mark on a scale of measurement from 0 to 100%, his or her adherence to the medication over the past 4 weeks. The results were converted to an adherence level expressed as a proportion (%) and classified into three (3) categories of adherence. Patients with result above 95% were classified as adherent and those with a result equal to or below 95% were classified as poorly adherent to antiretroviral treatment Simplified medication adherence questionnaire (SMAQ): the SMAQ was used to collect information on adherence over the previous 3 months period. The SMAQ score ranged from 0 to 7 with 0 corresponding to 100% adherence. A patient was considered as positive or non-adherent when a positive response was given to one of the questions, or the patient did not take any medicine over the past weekend, or had missed taking the medicine for more than 2 days over the past 3 months [19].

Multi-method approach: World Health Organization (WHO)
recommends a multi-approach method when measuring adherence to antiretroviral treatment [19][20][21]. The multi-method approach tool included self-reports combined with VAS and the SMAQ. Overall adherence assessment with the multi-method approach was rated into two (2) categories: high and low. A high level of adherence corresponds to a patient who reported "no" to all questions with selfreporting, had a VAS score "yes" 95% and who is adherent following the SMAQ method. A patient who did not meet the above-mentioned criteria was classified as poorly adherent with the multi-method approach.
Study variables: the outcome of interest was poor adherence to ARV treatment. We considered a multi-method approach by further categorizing overall adherence as adherent or non-adherent. A patient was categorized as adherent if they answered "no" to the all the self-report questions, reported 90% VAS or more and knew the dose, time and instructions. Where responses to self-report, VAS or pill identification were less than optimal (e.g. moderately or poorly adherent), overall adherence was categorized as non-adherent. We  (Table 1).
Social support and socio-economic characteristics: patients were predominantly unemployed (62.67%) ( Table 1) (Table 2). Results from the self-reporting adherence questionnaire, found that only 11.33% were adherent and with the SMAQ the proportion of patients classified as being highly adherent increased again to 30.6%. In the multi-method approach which combines visual analogue scale, self-reporting and SMAQ, of a total of 150 patients, 26.67% were found to be highly adherent.
Determinants of ART adherence: in bivariate analysis, factors associated with poor adherence were tobacco use and being involved into the activities of associations of persons living with HIV (Table 3).
Results from the multivariate regression analysis (Table 4) indicates that there was a decrease in the odds of having high adherence level depending on patient's demographic and psychosocial support.
Patients classified as widowed (Aor: 0.49 95% CI: 0.03-0.94. p value: 0.043) were less likely to be highly adherent as compared with patients who reported they were single at the time of the study.
Additionally, patients who declared that they participate in the activities within an association of person living with VIH (PLWHIV) were 2.34 times more likely to be adherent as compared to patient who did not belong to any association of persons living with HIV (Aor:2.89 95% CI:1.04-7.99 p value: 0.041) ( Table 4).

Discussion
Recent estimates state that 6 million HIV-positive people have initiated ART in sub-Saharan Africa and this figure is likely to increase in the next year following the implementation of the "test and treatment initiative" and the "90-90-90" approach [16][17][18][19][20][21][22][23][24]. Initiating all the HIV-positive patients onto care will be a remarkable success for the National AIDS programs. However, one of the main challenges will be keeping these patients on care and highly adherent to antiretroviral treatment first line regimen. In this context, a thorough understanding of factors associated with ART adherence allows for targeted interventions that can be implemented within the context of health care services in public hospitals areas in order to keep patients in care and adhering to treatment in the early stages of ART [25]. It is therefore necessary to continue to tackle the issue of adherence and identify patients at risk of poor clinical outcomes who need therapeutic education and support. In our study, approximately 30% of patients were found to be adherent to ARV treatment. This finding is consistent with previous reports from industrialized countries, documenting 25%-44% of adherent patients [22]. However, this finding is different from previous studies in Senegal [12][13][14][15]. In a study implemented in a cohort of patients who were on ART from 1998 to 2000, Laniece et al. [13] showed that long term adherence among HIV patients in Dakar was very high with nearly more than 95% being adherent This situation may be explained by the fact that in our study, we used a more stringent criteria to assess adherence with the combining of three self-reported adherence as recommended by WHO [26]. Additionally, these previous studies were conducted Data from industrialized countries showed that younger patients with poorer education and those with low socio-economic status [29,30] were less likely to being adherent, but there is little evidence from resource-limited settings. Surprisingly, we did not detect a relationship between adherences to antiretroviral therapies and socioeconomic factors. This difference may be explained by the smaller sample size in our pilot study but also by the fact that in these previous studies, adherence was assessed in a long-term manner with a prospective cohort study. Additionally, in our study there was no association between income or socio-economic levels and adherence to ART. This situation may be explained by the fact that ART is free for all HIV positive patients regardless of their socio-economic level in Senegal since 2002 [10]. In our study, there was a significant relationship between marital status and adherence to antiretroviral treatments. Widowed patients were less likely to be highly adherent as compared to single patients. Our findings confirm previous studies in Senegal, West Africa and in Uganda who have reported a similar pattern among HIV-positive patients [15][16][17][18][19][20][21][22][23][24][25][26][27][28][29][30][31] Fourth, we did not have access to ARVs medical records and informations on the availability of ARVs drugs at the health facilities.
Therefore, there was no possibility to assess the impact of drug stock out in adherence levels or to correlate the adherence levels with viral load which represents the gold standard when assessing adherence to antiretroviral treatment. Finally, since this was a cross-sectional survey, the outcome and exposure were collected at the same time and therefore no inference or causalities association can be concluded. However, even with these limitations, the analysis of data collected from hospitals provides important insights into the factors influencing adherence to antiretroviral treatment in Senegal and can be useful in guiding policy implementation strategies.

Conclusion
The adherence rate found in this study seems to be low. The use of three different adherence indicators was important for reducing bias through self-reporting and therefore strengthened the indicator. For