Level of non-adherence and its associated factors among adults on first-line antiretroviral therapy in Amhara Regional State, Ethiopia

Background In Ethiopia, nearly one-third of people living with human immunodeficiency viruses do not adhere to antiretroviral therapy. Moreover, information regarding non-adherence and its associated factors among adults on first-line antiretroviral therapy in Northeast Ethiopia is limited. Therefore, this study aimed to assess the level of non-adherence and its associated factors among adults on first-line antiretroviral therapy in North Shewa Zone, Amhara Regional State, Ethiopia. Methods A facility-based cross-sectional study was conducted on 326 participants selected by systematic random sampling technique from the five randomly selected public health facilities. Data were collected using the questionnaire adapted from the studies conducted previously and the collected data were entered into Epi data version 3.1 and exported to Stata version 14 for further analysis. Multivariable logistic regression analysis was done and an adjusted odds ratio with its corresponding 95% confidence interval was used to declare a statistical significance. Results The overall prevalence of non-adherence was 17.4% [95% CI: (12.8%, 21.2%)]. Patients with no formal education [AOR (95% CI) = 5.57 (1.97, 15.88)], those who did not use memory aids to take their medications [AOR (95% CI) = 3.01 (1.27, 7.11)], travel more than 10 kilometers to visit the nearby antiretroviral therapy clinics [AOR (95% CI) = 2.42 (1.22, 25.86)], those who used substance [AOR (95% CI) = 3.57 (1.86, 28.69)], and patients whose medication time interfered with their daily routine activities [AOR (95% CI) = 15.46 (4.41, 54.28) had higher odds of having non-adherence to first-line antiretroviral therapy compared to their counter groups. Conclusion The level of non-adherence to first-line antiretroviral therapy was 17.4%, higher compared to WHO’s recommendation. Hence, patients counseling focused on avoiding substance use, use memory aids, and adjusting working time with medication schedule are very crucial. Furthermore, the ministry of health and the regional health bureau with other stakeholders should expand antiretroviral therapy service delivery at health facilities that are close to the community to address distance barriers.


Conclusion
The level of non-adherence to first-line antiretroviral therapy was 17.4%, higher compared to WHO's recommendation. Hence, patients counseling focused on avoiding substance use, use memory aids, and adjusting working time with medication schedule are very crucial. Furthermore, the ministry of health and the regional health bureau with other stakeholders should expand antiretroviral therapy service delivery at health facilities that are close to the community to address distance barriers.

Background
According to the 2019 global estimate, about 36.2 million adults were living with the human immunodeficiency virus (HIV), of these, 20.6 million were found in Sub-Saharan countries [1][2][3]. In Ethiopia, more than half a million (722,248) people are living with HIV, of these, one-third (30%) were from Amhara Regional State [4,5].
The adult HIV prevalence in Ethiopia is 0.9% [6], with 2.9% in urban and 0.4% in rural settings [4,7,8]. The incidence of a new infection is also at a steady state [7,9], below threefourth of People Living With HIV(PLWHIV) ever enrolled on antiretroviral therapy (ART) in the country [3,8,10].
Recent data showed that above 95% adherence to the ART regimen is required for HIV infected patients to reach full viral suppression [11][12][13], but sustaining its adherence level requires accurate and consistent monitoring activities, and this is a major challenge for sub-Saharan African countries like Ethiopia [14] In Ethiopia, 5% to 35% of PLWHIV did not adhere to the prescribed ART regimens [15][16][17][18][19][20] which is far away from the national target of HIV prevention programs [4]. Studies conducted in different settings identified socio-demographic, behavioral, disease characteristics, medication, and health systems-related factors as the common factors affecting adherence to ART regimens [15,[21][22][23][24][25][26].
In Ethiopia, different strategies like a transformation of a fee-based ART program in 2003 to a free ART program in 2005, decentralization of services to lower-level public health facilities and private hospitals, and capacity building for a service provider on counseling aspect have been tried to improve adherence to ART [27,28]. However, still maintaining adherence to ART regimens has been one of the challenges of ART programs in the country [17,29,30].
Studies conducted previously were focused on clinical determinants of non-adherence rather than looking at the clients' behavioral aspects. Besides, little is known about the level of non-adherence and its associated factors in the study area. Therefore, this study aimed to determine the level of non-adherence and its associated factors among adults on first-line ART in North Shewa Zone, Amhara Regional State, Ethiopia.

Study setting, design, period, and population
A facility-based cross-sectional study was conducted from 15 th March to 15 th April 2019 in North Shewa Zone, Amhara Regional State, Ethiopia, which is located 120 kilometers away from Addis Ababa, the capital of Ethiopia. In this Zone, there are thirteen health centers and five public hospitals. In all of the health facilities, the patients have a regular visit to the ART clinic on a monthly follow-up basis.
In this study, out of the ART providing public health facilities in the study area, one referral hospital (Debre Birhan Referral Hospital) and four health centers (Chacha, Debresina, Shewa-Robit, and Debre Brihan Health Centers) were randomly selected. All of the five health facilities are located in urban areas and are providing ART services for people coming from both urban and rural dwellers.
All adults on ART who were at least 18 years of age, able to hear, mentally fit, and being on follow-up for at least 3 months before the data collection period were eligible for this study. While those who were seriously ill during the data collection period were excluded.

Sample size determination and sampling techniques
For the first specific objective (i.e. to determine the level of non-adherence to first-line antiretroviral therapy among Adults in North Shewa Zone, Amhara Regional State, Ethiopia), the sample size was estimated using a single population proportion formula by considering a 95% confidence level, 5% margin of error, and non-adherence level of 25.8%, taken from the study conducted in Southwest Ethiopia [31].
Where, n = required sample size, Zα/2 = critical value for normal distribution at 95% confidence level (1.96), p = expected level of non-adherence, and d = margin of error). Based on the above assumptions, the final sample size for the study was 296.
For the second specific objective (i.e. to identify the factors associated with non-adherence to first-line antiretroviral therapy among adults in North Shewa Zone, Amhara Regional State, Ethiopia), the sample size (118) was estimated using Epi-info version 7.2.1 by considering different variables that have a statistically significant association with non-adherence [20], the ratio of exposed to unexposed to be 1:1, 95% confidence level, and power of 80%. Finally, from the sample sizes estimated for both specific objectives, the minimum largest sample size was 296 (i.e. the sample size of the first objective). Then, with a 10% non-response rate, the final sample size was 326 adults on first-line ART.
Finally, the sample size was proportionally allocated to each health facility based on their preceding one-year adult ART case flow. Then, the sampling fraction was determined by dividing the total number of the eligible population getting services in each facility by the allocated sample size (N/n), which is approximately equal to six for all facilities. Accordingly, every 6 th participant was selected using a systematic random sampling technique from the patients' registration book.

Study variables
Dependent variable. Adherence status to first-line ART that was dichotomized into "non-adhered" and "adhered" was measured by using the patient's self-reported percentage of missed doses during the last month. Thus, participants who missed more than 5% of the prescribed ART doses based on self-report were considered to have non-adhered to the treatment and otherwise considered to have adhered to the treatment [32,33] Independent variables. Sociodemographic characteristics (educational level, monthly income, marital status, residence), use of memory aids to take ART drugs, get social support, type of substance used, medication time interfere with daily activities, and distance from ART clinics.

Definition of terms
Social support. A participant was considered as getting social support if he/she gets economic, psychological, and social support from his relative's/family's members.
Emotional life. The emotional life of the participants was assessed using five (yes/no) questions. Participants who responded correctly to three or more questions were categorized as having a stable emotional life and those who provided a correct response for less than three questions were categorized as having unstable emotional life.
Substance use. Participants who reported using at least one substance (alcohol, cigarette, and khat) 7 days before the data collection period were considered as "using a substance" and otherwise considered as "not using a substance".

Data collection tools and procedure
The questionnaire was adapted from studies conducted in developing countries [6,[34][35][36][37]. The questionnaire consists of sociodemographic, social support, behavioral factors, level of adherence, use of memory aids to take the ART drugs, and history of other medical illnesses. It was initially developed in the English language and translated to the local language (Amharic) and translated back to English to confirm its consistency. The tool was pretested on 5% (16 participants) of sample size other than selected health facilities and some amendments were made based on the pretest findings. Data were collected by trained health professionals who were not working in the ART clinic in the selected health facilities.

Data management and statistical analysis
The data were cleaned, coded, and entered into Epi data version 3.1 and exported to Stata version 14 for analysis. To assess the association between the dependent variable with each independent variable, bivariable logistic regression analysis was done and those independent variables with a p-value less than 0.25 were considered in the final multivariable logistic regression model. The eligible covariates were socio-demographic characteristics, use of memory aids, distance to ART clinics, social supports, substance use (alcohol, chat chewing, or cigarette smoking), and time of medication taking. The correlation between independent variables was checked using variance inflation factor (VIF) and its value for all variables was less than 5. The model fitness was assessed using the Hosmer-Lemeshow model fitness test (p-value of 0.32). Adjusted odds ratio with its corresponding 95% confidence interval was used to declare a statistical significance.

Ethical consideration and consent to participate
Ethical clearance was obtained from Research Ethics Review Committee, Dream Science and Technology College (DSTC) with an ethical letter of DSTC/0123/2019. Confidentiality of the information was secured by excluding respondents' identifiers, such as names, from the data collection format. Finally, written and verbal informed consent was obtained from those who were willing to participate in the study.

Participants' socio-demographic characteristics
In this study, 310 participants were included with a response rate of 95.1%. The mean age of the respondents was 40.3±1.32 years. The median duration on ART was 39 [interquartile range (24-72)] months. Of the total respondents, 193 (62.3%) were residing in rural areas, 60 (19.4%) did not attend formal education. Regarding occupation and religious status, 137 (44.2%) were merchants, and nearly half (48.1%) were orthodox religious followers ( Table 1).

HIV/AIDS and ART service-related information
Two hundred twenty-six participants (72.9%) heard about HIV/AIDS before their first contact with health care providers, and 199 (64.2%) participants got information about ART from health care providers. Nearly two-thirds (69.7%) of the participants were aware of the benefits of ART before they started treatment and 226 (72.9%) reported as they were benefited from their ART ( Table 2).

Behavioral, psycho-social, health service, and ART drug-related information
The main reasons for missed ART doses were living far away from the health facilities (11.9%), forgetting (7.4%), drug side effects (6.8%), and being busy (6.1%). Moreover, 61.0% of the participants reported as they had ART-related side effects predominantly vomiting (24.8%) and diarrhea (18.7%). The study revealed that 60.3% of respondents had a history of substance use ( Table 3).

Factors associated with non-adherence level
After Adjusting for selected covariates, patients with no formal education had more than five times higher odds of non-adherence compared to those who attended at least secondary education [AOR (95% CI) = 5

PLOS ONE
Level of non-adherence and its associated factors among adults on first-line antiretroviral therapy

Discussion
Non-adherence to ART led to HIV-related morbidity and mortality [38,39]. Our study revealed that ART non-adherence level based on self-report at the time of follow up was 17.4% (95% CI: 12.8%, 21.2%), higher than the findings of the studies in Southwest Ethiopia [17], northwest Ethiopia [20,40], Eastern Ethiopia [22], Tanzania [41], South Africa [42], and Uganda [33]. The discrepancy could be due to differences in the study setting and the existence of socio-cultural differences of the participants. However, the level of non-adherence in this study is lower than the findings of the studies in Southwest Ethiopia [19], Northeast Ethiopia [43], Southern Ethiopia [31], Papua New Guinea [44], and Nigeria [35]. The discrepancy might be explained by differences in the access to ART service, interventions taken [22], and awareness about the benefits of ART in the study areas.
In this study, the odds of non-adherence among adults who had no formal education was more than five times higher compared to those who attended secondary education and above. This finding is consistent with the result of the studies in Togo [45] and Guangzhou-China  [24]. Higher education level has been one of the independent predictors to have adherence to ART [21,22,24,42]. This study revealed that the odds of non-adherence among adults who used substances were more than three times higher compared to counterparts. This finding is supported by the studies conducted in Uganda [33], Vietnam [46], Togo [45], United States [47], Atlanta [48,49], Amana Hospital, Tanzania [37], and Yaounde, Cameroon [50]. This will be can be explained by the fact that the use of substances like alcohol associated with memory impairment, subsequently failing patients to take the drugs as per schedule [34,46,49,[51][52][53]. It has also the potential to disrupt self-organizing skills and sleeping time patterns.
The odds of non-adherence among adults who did not use memory aids to take their ART drugs were three times higher compared to those who were using memory aids. This finding is in line with the study conducted in Jimma, Ethiopia [25], Vietnam [54], and Chicago [15]. The use of memory aid methods like peer counselors, mobile phone text messages, reminder devices, cognitive-behavioral therapy, and behavioral skills or medication adherence training [1,5,25,55] have a great role in improving adherence. Thus, people using different reminders had a better level of adherence to the ART treatment [3,56].
In this study, adults whose medication-taking time interfered with their daily routines had more than fifteen times higher odds of non-adherence compared to those whose medication schedules did not interfere with their daily activities. This finding is consistent with the findings of other studies [35,56]. Hence, the people whose medication-taking time interfere with their working time are more likely to miss their ARV drugs which results in non-adherence [23,31,41].
In this study, the odds of non-adherence among adults living at a distance greater than 10 km and above from the ART clinic were more than two times higher compared to those residing nearer to the clinic. This finding is similar to the study conducted in Harari, Ethiopia [15], southwest Ethiopia [18], Indonesia [57], and the Democratic Republic [58]. This might be related to the availability of transportation and its costs. Thus, patients who live far away from the ART-providing clinics may not afford the transportation cost which leads to defaulting from the follow-up treatment.
As a limitation of the study, there might be recall bias since adherence was assessed based on data taken one month before the study period. Besides, the findings may not be generalized to patients who get ART services in private health facilities. However, to minimize recall bias we used triangulation of information from a medical record for some of the variables. In addition, due attention was given to all the study procedures throughout the field activities.

Conclusions
In this study, the level of non-adherence to first-line ART was higher than the WHO standard. patients with no formal education, those who did not use memory aids to take their medications, travel more than 10 kilometers to visit the nearby ART clinics, those who used substance and patients whose medication time interfered with their daily routine activities had higher odds of having non-adherence to first-line ART compared to their counter groups.
Therefore, patients counseling focused on avoiding substance use, use memory aids, and adjusting working time with medication schedule are very crucial. Furthermore, the ministry of health and the regional health bureau with other stakeholders should expand ART service delivery at health facilities that are close to the community to address distance barriers.