Abstract
Background:
The authors studied the modalities of nonadherence to highly active antiretroviral therapy (HAART) and its sociodemographic associated factors and those in relation to caregiving perception in Ouagadougou.
Methods:
A cross-sectional study was performed from December 2013 to February 2014 in 2 health centers. Adults receiving HAART for at least 3 months were included. Adherence was studied according to the quantitative, qualitative, and global criteria. Factors associated with nonadherence were analyzed with chi-square and Fisher tests. A logistic regression model was applied for multivariate analysis.
Results:
The authors studied 152 patients: mean age 40.7 ± 7.8 years and sex ratio 0.34. Frequencies were 7.2% for self-reported quantitative, 20.4% for calculated quantitative, 31.6% for qualitative, and 38.2% for global nonadherence. Married status (P = .02), patient’s dissatisfaction regarding clinical monitoring (P = .01), and therapeutic education (P = .03) were associated with nonadherence. In multivariate analysis, married status remains associated (odds ratio = 7.00, 95% confidence interval = 1.89-25.8, P = .0004).
Conclusion:
Nonadherence to HAART needs to be correctly managed during HIV/AIDS monitoring.
Introduction
Highly active antiretroviral therapy (HAART) has improved the impact of HIV/AIDS programs: an equivalent of 14 million years of life has been saved in low- and middle-income countries, including 9 million in sub-Saharan Africa since 1996. 1 This is nevertheless threatened by treatment failures, mainly related to nonadherence. 2 Therapeutic education and other strategies (simplification of HAART in virological success condition, prescription of fixed-doses combinations of antiretroviral (ARV) drugs, and removal of molecules with an high toxicity) have been performed to reduce this deleterious behavior. Despite this, nonadherence remains a health challenge that in the absence of an appropriate reaction could worsen, given a greater access to therapy. If the United Nations AIDS program goals are achieved, more than 15 million patients to date should receive HAART versus 9.7 million in 2012. 1 Because of a greater number of treated patients and the increase of disease duration, it is important to have reliable data concerning factors of nonadherence in order to preserve the effectiveness of ARV drugs. In Africa, the issue of nonadherence is recent, due to the late arrival of ARV drugs. The problem should incite solutions taking into account socioeconomic and cultural specificities. In Burkina Faso, studies related to nonadherence have been previously performed in public health centers. 3,4 For a global comprehension, we reported, using another methodology, the modalities of nonadherence, its sociodemographic associated factors, and those in relation to caregiving perception by patients in 2 private health centers in Ouagadougou.
Materials and Methods
Study’s Design
We performed a cross-sectional cohort’s study from December 2013 to February 2014 at Saint Camille Medical Centre (CMSC) and Pietro Annigoni Biomolecular Research Centre (CERBA), 2 private health structures in convention with the government in Ouagadougou. The 2 study sites are administratively connected and have the same operating mode. Each center has an HIV outpatients care department that offers screening, treatment and monitoring, psychosocial, and nutritional support. Therapeutic education is a routine activity systematically delivered to patients initiating HAART. Afterward, patients benefit monthly from adherence support sessions in the first quarter, then quarterly, and semiannually. If necessary, additional sessions are possible. These individual discussions are facilitated by a staff of nurses expert in therapeutic education, and expert patients and various fields of HIV/AIDS including adherence to HAART are discussed. If possible, social, psychological, and economic supports are provided.
Population and Sampling
The study’s population consisted of HIV-infected adults (age ≥18 years) who were receiving HAART in CMSC and CERBA. An accidental sample of patients treated for at least 3 months without cognitive dysfunction and who consented to the study was included. Patients who changed therapeutic line or regimen for less than 3 months, those with incomplete clinical records, or not in possession of the remaining tablets of ARV drugs were not included. The sample’s size to estimate nonadherence was calculated using Schwartz’s formula: n = z2p (1 − p)/e2 (n = sample size, p = proportion of nonadherence, and e = error margin): p = 50% in reference to 40% to 50% of nonadherence in the area, 3 –5 z = 1.96 (confidence level 95%) and e = 10%. Then, n = 95 patients.
Data Collection
Data were collected with a questionnaire. An individual interview was supplemented by a clinical file review. Interviews and counting of tablets were realized in the consultation hall. A pretest was performed, and patients who took part were not included.
Study Variables
Two types of variables were studied: the dependent variable (nonadherence) and the independent sociodemographic and lifestyle-related variables (age, gender, marital status, instruction status, monthly income, residence, alcohol, and tobacco intake); and those in relation to the caregiving perception by patients (medical consultation, clinical monitoring, therapeutic education, and social support).
Operational Definition
Nonadherence considered as the nonrespect of prescription was evaluated by patients’ self-reports on the management of treatment by the last 4 day’s recall and the counting of tablets. Quantitative adherence was determined by dividing the number of tablets that the patient said to have taken and the number of tablets that have been prescribed. The ratio defined “nonadherence” in case of taking less than 95% tablets: self-reported quantitative nonadherence (patient’s statements) and calculated quantitative nonadherence (after counting tablets). Qualitative adherence was evaluated according to self-reports on compliance with schedules: gaps of more or less 2 hours were considered “qualitative nonadherence.” 6 The “global nonadherence” was when patient had taken less than 95% tablets and/or took tablets without compliance with schedules. Patients were considered “married” when they had contracted a civil, religious, and/or traditional marriage.
Ethical Considerations
The ethical committee approved the study. The confidentiality was monitored during the study: data collection was subjugated to professional secret, no nominative data were collected, and selection was conditioned by patient informed consent.
Data Analysis
Data were analyzed using Epi-Info 3.5.4 and IBM SPSS statistics 18. The sample was described in accordance with descriptive statistics. For qualitative variables, proportions were used. For quantitative variables, mean and standard deviation were used. To analyze association between nonadherence and potential explanatory variables, chi-square and Fisher tests were used. To study factors independently associated with nonadherence, we used a logistic regression model. A P value was significant for less than .05.
Results
Among 200 patients, 152 (76%) participated: 94 (61.8%) in CERBA and 58 (38.2%) in CMSC.
General Characteristics of the Participants
The sample consisted of 113 (74.3%) females and 39 (25.7%) males, that is, sex ratio 0.34. Patient’s mean age was 40.73 ± 7.88 years. One hundred thirteen patients (74.3%) were parents. Almost half (45.4%) of patients reported less than 3 meals per day. The mean CD4 count at the initiation and at the time of the study was, respectively, 167.20 ± 82.82 and 604.94 ± 289.68cells/mm3. The mean duration of HAART was 88.36 ± 30.53 months, and zidovudine (ZDV)/lamivudine (3TC)/nevirapine (NVP) was the common regimen (41.6%). The serotype of HIV and the modalities of HAART are reported in Table 1.
Serotype of HIV and Modalities of Highly Active Antiretroviral Therapy in HIV-Infected Patients in CERBA and CMSC.a
Abbreviations: CERBA, Pietro Annigoni Biomolecular Research Centre; CMSC, Saint Camille Medical Centre; NNRTI, nonnucleotide reverse transcriptase inhibitor; NRTI, nucleotide reverse transcriptase inhibitor; PI, protease inhibitor.
aN = 152.
Frequency and Modalities of Nonadherence
Eleven (7.2%) patients self-reported quantitative nonadherence. After counting tablets, this concerned 31 (20.4%) patients. Forty-eight (31.6%) patients self-reported qualitative nonadherence. Globally, 58 (38.2%) patients were nonadherents.
Lifestyle and Demographic Factors Associated with Nonadherence
Nonadherence was observed in 20 (41.7%) patients who used to take alcohol and 38 (36.5%) who didn’t, P = .54. It was observed in 2 (40%) patients who used to smoke and 56 (38.1%) patients who didn’t, P = .63. Table 2 shows the relation between global nonadherence and the sociodemographic characteristics of patients. In multivariate analysis, married status was associated with nonadherence (odds ratio = 7.00, 95% confidence interval = 1.89-25.8, P = .0004).
Relation between Global Nonadherence to Highly Active Antiretroviral Therapy and the Sociodemographic Characteristics of Patients in CERBA and CMSC.a
Abbreviations: CERBA, Pietro Annigoni Biomolecular Research Centre; CMSC, Saint Camille Medical Centre.
aN = 152.
Caregiving Perception by Patients and Nonadherence
Nonadherence was observed in 20 (48.8%) patients reporting difficulties to access to health center, P = .10. Table 3 shows the relation between the caregiving perception by patients and nonadherence.
Relation between Nonadherence to Highly Active Antiretroviral Therapy and the Caregiving Perception by Patients in CERBA and CMSC.a
Abbreviations: CMSC, Saint Camille Medical Centre; CERBA, Pietro Annigoni Biomolecular Research Centre.
aN = 152
Discussion
Nonadherence was often qualitative, and its global frequency was 38.2%. The married status, patient’s dissatisfaction regarding clinical monitoring, and therapeutic education were the sociodemographic factors and those in relation to caregiving perception by patients associated with nonadherence in univariate analysis. In multivariate analysis, married status remained associated with nonadherence.
In sub-Saharan Africa, nonadherence is diagnosed in 33% to 58% patients receiving ARV drugs 5,7,8 ; In Tunisia, frequencies are 27% and 65%, 9,10 while a meta-analysis in North America reports 45%. 7 There is no “gold standard” to evaluate adherence so that comparisons of frequencies are sometimes difficult. However, observations are consistent with the alert of the World Health Organization on the extent of nonadherence during the management of chronic diseases. 11 Qualitative nonadherence occurred in 31.6% patients. Its evaluation, especially for schedules in drugs’ administration, is often overlooked; failure is nevertheless reported in 42% to 80% patients in Cameroon 12,13 and 55.6% in Morocco, 14 although tolerable gaps in schedules are not harmonized. In addition, in low-income countries, schedules are biased because most patients don’t have watches and/or are illiterates: sunrise and sunset are usually used, although not having a clock precision. Fixed-dose combinations of ARV drugs in a daily single administration could partially decrease this obstacle. Diet conditions haven’t been estimated in our study; limitations in sub-Saharan Africa could be a deleterious factor when specific or nonspecific diet is recommended in the prescription. In Cameroon, the lack of food is a reason of nonadherence for 7.6% patients. 13 Self-reported quantitative nonadherence concerned 7.2% patients, this was 5% and 6.9%, respectively, in Cameroon and Ethiopia. 15,16 Calculated quantitative nonadherence concerned 20.4% patients; frequencies vary from 20% to 22.2% elsewhere in Africa. 4,16,17 This confirms the underestimation of nonadherence based on patients’ self-reports. 13,18,19 However, adherence is commonly evaluated in accordance with this method whose subjectivity could have been increased in our study because of the administration of the questionnaire by a health professional: bias in declaration could be related to a need of “social conformity” 20 or a fear of invectives. A self-administered anonymous questionnaire could limit such bias, but this method is questioned because of the frequency of illiteracy in patients. The bias also affects the effectiveness in tablets administration and respect of schedules. Despite limits, our methodology, compared to what have been previously done in Burkina, is more complete, taking into account quantitative and qualitative components.
Married status was the sociodemographic associated factor to nonadherence. Folefack reports an association in patients living in couple. 21 In the opposite, studies report nonadherence more in patients living in loneliness situation. 17,22 Our result is paradoxical from that time almost patients in couple informed their partners and benefited from their support, anything that should have increase conditions for compliance. The partner’s HIV status and especially the structure of families and households have to be clarified in order to understand this paradox. In the family microcosm, the polygamy, the large size, and the composite structure of households and families in Africa don’t allow discretion and confidentiality, anything to be considered for therapeutic education in order to reduce a source of stigma. Indeed, studies report nonadherence in case of difficulty to stock discreetly medication in group housing conditions. 9,12
In univariate analysis, patient’s dissatisfaction regarding the clinical monitoring and the therapeutic education was associated to nonadherence. Adherence support is positively perceived by patients. 22,23 The strategy is usually implemented by community-based organizations that provide psychosocial support with medical care facilities. Then, patients with “high potential of nonadherence” are identified and benefited from adequate support. The quality of the doctor–patient relationship (listening, discretion, confidentiality, and support) also influence adherence. 22 –24 Beyond these relational associated factors, interruptions in care’s continuum and long expectations during clinical consultation contribute to a poor perception and a poor adherence to care.
Conclusion
Adherence to HAART is a major component in HIV/AIDS programs. Our findings report different obstacles related to sociodemographic factors and those in relation to caregiving perception by patients. The study provides data for optimizing pretreatment and adherence consultations. Further studies could look at adherence in couples affected by HIV in order to its better comprehension.
Footnotes
Acknowledgments
The authors thank the medical staff and patients of the 2 centers who have participated or supported the study. They also thank the partners of these centers: the Ministry of Health of Burkina Faso, UEMOA (PACER2) and CEI Italy.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.