Predictors of dietary diversity among adults on antiretroviral therapy in Debre Tabor Hospital, northwest Ethiopia: A cross-sectional study

Background: Diversied diet to improve better tolerates human immunodeciency virus drugs, enhance antiretroviral therapy adherence and maintain a healthy weight to reduce mortality and morbidity. Therefore, the aim of this study was to assess dietary diversity and associated factors among adult human immunodeciency virus positive patients who were on antiretroviral therapy at Debre Tabor Hospital northwest Ethiopia Methods: Institutional based cross-sectional study was conducted at Debre Tabor Hospital and participants selected by the systematic random sampling techniques. Data collection was done by using a structured interview questionnaire. Dietary diversity was computed from all food and drink list consumed in the 24 hours preceding the survey. A variable having p- value < 0.2 in the bivariate logistic regression was entered into the multivariable logistic regression, and independent variables having p-value < 0.05 was considered as signicantly associated. Results: Of the total 341 adults on antiretroviral therapy 336 participants gave a complete response with a response rate of 98.50%. Among the participants 336 [32.4% (95% CI: 27.4, 37.1)] had adequate dietary diversity and 22.9% were BMI <18.5kg/m 2 . Government employment [AOR= 2.5; 95%CI: 1.28, 4.98) and diploma and above educational status [AOR=2.3; 95% CI: 1.01, 5.31) were factors for adequate dietary diversity. Conclusions: In this study the magnitude of adequate dietary diversity was low. Employment and educational status were signicantly associated with dietary diversity among people living with HIV/AIDS. Hence, improving education status as well as income generation activities are highly recommended strategy to improve the dietary diversity of patients on antiretroviral therapy.


Introduction
Globally, 36.7 million people are living with HIV/AIDS (PLWHA) (1). Poor nutrition and HIV/AIDS are interrelated and aggravate each other in a vicious cycle by attacking the immune system (2). The nutritional status of an individual is affected by HIV/AIDS through increasing energy requirements, reducing food intake, and decreasing nutrient absorption and metabolism (3,4). Having good nutritional status and eating diversi ed diets have an important role in enhancing antiretroviral therapy (ART), encouraging good adherence to ART drugs, and maintaining healthy weight (5).
HIV/AIDS attacks the immune system and make the body susceptible to opportunistic infections like fever, diarrhea, tuberculosis, pneumonia. Hence, adequate dietary intake alongside ART helps the immune system to be strong, and enable it to ght diseases better (6). On the other hand, undiversi ed diet can intake possibly contribute to micronutrient de ciencies that lead to HIV/AIDS progression and the depletion of the cluster of differentiation of the CD4 count which increases the risk of opportunistic infections in addition to oxidative stress (7,8).
Dietary diversity is the assessing qualitative utilization of food which re ects an individual's right use to different foods; it is also an indicator of a diet's micronutrient adequacy, an important dimension of its quality (9,10). It even serves as a proxy indicator of diet quality and indicates its association with the nutrient adequacy of an individual's diet (11,12).
Dietary diversity is a challenge to communities the cause their diets are mainly starchy staples with inadequate animal products, fruits, and vegetables (13). Utilizing diversi ed diet among PLHIV is still poor among resource limited nations like Africa. For instance, 58.8 and 62.3% of the PLHIV in Nigeria and Uganda (14,15) received inadequate dietary diversity respectively while, 28.7-58.8% of adults obtained inadequate dietary diversity in Ethiopia (13,16). Previous studies reported residence (17), wealth index (13,17), employment status (13), duration of ART treatment (16), owning mobile cell phones (18), media exposure status (18) and nutritional counseling (18) were factors affecting dietary diversity among adults living with HIV/AIDS. Ethiopia has made a notable effort to address the impact of HIV/AIDS on nutrition by preparing national guidelines for HIV/AIDS and by taking actions to provide quality care and support to people living with the disease (19). In addition, it provided livelihood support and food assistance and strengthened community based nutrition care and support activities for PLHIV through health extension workers, agriculture extension workers, and health development armies, in addition to HIV/AIDS treatment, care, and support (20,21). But, diversi ed diet is a serious unmoving problem among 29.5% of adults living with HIV/AIDS (18). Yet, there is little information regarding the effects of undiversi ed diet among PLHIV in Ethiopia, including the study setting (16, 18). Therefore, this study aimed to assess the prevalence and Sample size and sampling procedures Sample size was determined by the simple population proportion formula by considering the following assumptions; a 95% con dence interval, marginal error of 5%, and 58.8% (13) proportion of inadequate dietary diversity which yielded a sample of 372. However, as the source population was less than 10,000, we considered a population reduction formula plus a non-response rate of 10%. The nal sample size was then 341. Participants were selected by the systematic random sampling technique. According to the report from the hospital administration, the monthly average number of adult HIV/AIDS patients on ART was 990. The sampling interval (k th ) value was determined by dividing the total monthly adult patients by the sample size (990/341 = 3). The rst respondent was selected by the lottery method out of the rst three clients. The procedure continued until the required sample size was obtained.
Data collection methods and quality control Data were collected by a face to face interview using a structured questionnaire. The tool included socio demographic characteristics, clinical conditions, nutritional status, and dietary characteristics. Dietary diversity score was computed by asking the participants if they consumed all foods and drinks on the list in the preceding 24 hours of the survey according to the Food and Agriculture Organization (FAO) 2011 (22). Likewise, the nutritional status of the participants was assessed by measuring weight and height and calculating Body Mass Index (BMI). BMI less than 18.5 kg/m 2 was considered as undernutrtion, while a BMI scale of 18.5-24.9 and BMI >30 kg/m 2 were taken as normal and obesity, respectively (23).
To maintain data quality, training was provided to data collectors and supervisors for two days by the principal investigator. The training was related to how to approach participants ethically while interviewing them. A pre-test was conducted on 5% of the participants at Debre Tabor health center. The completeness, accuracy and consistency of the collected data were checked daily by an assigned responsible supervisor and the principal investigator.

Measurements of dietary diversity score
The dependant variable of this study is dietary diversity. Individual Dietary Diversity Score (IDDS) of the study participants was measured by adding the food groups consumed over the preceding 24 hours before the survey from starchy staples, pulse, nuts and seeds, dairy, meat, poultry and sh, eggs, dark green leafy vegetables, other vitamin A-rich fruits and vegetables, other vegetables and other fruits. If participants consumed ve or more food groups out of the ten listed, they were considered to have an adequate dietary diversity score (22).The independents variables in this study include socio demographic characteristics, clinical conditions and nutritional status of the participants.

Data analysis
The collected data were entered, coded, and cleaned using Epi INFO version 7.0 and data management and analyses were performed using SPSS version 20.0 software. The association of dependent and independent variables was assessed by using the binary logistic regression and in the bivariable analysis variables with p values <0.2 were entered into the multivariable logistic regression with 95% Con dence Intervals. The corresponding p value of < 0.05 was considered as statistically signi cant at a 95% con dence interval.

Result
A total of 336 HIV positive adults on ART participated in the study with a response rate of 98.5% and a mean age 39.08± 10.9 SD. Of the total respondents, almost all (99.4%) of respondents were Orthodox Christians and less than half (43.8%) were married. Most, (81.8%) of the respondents were urban dwellers; while 32.7% and 33% were college diploma graduates and government employees, respectively (Table 1).
More than half (58%) of the respondents were in WHO clinical stage one, and about half (53%) had CD4 count of greater than 500mg/mm, while nearly three-fourths (72.3%) were taking ART for more than three years ( Table 2).

Factors associated with dietary diversity among HIV/AIDS patients on ART
In the multivariate logistic regression analysis, educational and occupational statuses were signi cantly associated with adequate dietary diversity. Accordingly, patients who had diploma and above educational status were 2.3 times more likely to consume adequate diversi ed diet compared with those who were unable to read and write (AOR=2.3 ,95% CI:1.01,5.31). Likewise, higher odds of diversi ed diet were noted among government employees than among the unemployed (AOR=2.6, 95%; CI: 1.28, 4.98) (

Discussion
This study assessed the magnitude and predictors of dietary diversity among adult patients on ART. The overall prevalence of inadequate dietary diversity intake of the current study was 67.6%. This nding was comparable with that of a study done in Hosanna town, southern Ethiopia, (67.9%)(24); whereas, it was higher than those of studies done in Dembia (11.3%) (25), Uganda (14.7%) (26), Hiwot Fana and Dilchora hospital, eastern Ethiopia (28.7%) (16), Metema (58.8%) (13), and Nigeria (62.3%) (14). This variation might be due to differences in study settings, seasons, socio-economic status of respondents and cut-off points used to ascertain the outcome variable. Our study also veri ed that only 3.9% of the participants ate eggs. This nding was supported by that of a study done in Nigeria (14). That is because eggs were too expensive compared with staple diets.
In this study, patients who had completed college and above were more likely to get adequate dietary diversity compared to the non-educated. This was in line with studies done in Jimma (27), Metema (13), Amatole and Nyandeni districts, South Africa (28). This is due to the fact that if patients level of education is low, their access to nance and their contribution to the total producing income will be low while good education creates employment opportunities to generate income to assure food security and improve purchasing capacity of diversi ed diet. The other possible reason might be that more educated participants get better service in terms of nutrition education and counseling to improve their nutritional information. Nutritional knowledge has a role to play in dietary practices and the number of meals based on their requirements (29,30). Uneducated respondents may not understand the consequences of inadequate diversi ed diets easily (30,31). Higher education is more likely to correlate with higher income to spend on a variety of foods (32). Furthermore, educated respondents have media exposure to improve their nutritional knowledge and allocate a larger proportion of their household food budget to food groups. This was supported by previous studies that participants who had exposure to the media were more likely to receive diversi ed diets (33,34).
Employment status was the other factor signi cantly associated with dietary diversity in this study. Government employee HIV patients were 2.5 times more likely to consume adequately diverse diets than non-government employee patients on ART. This was supported by studies done in Jimma (27), Metema (13), and north India (35). The most possible explanation might be that government employees had regular income which they used to vary their food, while their counterparts had less chance of doing that. In addition, joblessness might be the result of how educational status which could result in poor wealth status that affects the capacity to purchase adequate and appropriate food. Besides, unemployed participants experienced food insecurity which decreased their capacity of consuming diverse diets at individual and household levels (36, 37).

Conclusion
This study showed that the prevalence of adequate dietary diversity was low. Government employment and diploma and above educational status were signi cantly associated with dietary diversity among adult people living with HIV/AIDS. Therefore, special emphasis on improving education status as well as income generation activities are highly recommended strategy to improve the dietary diversity of patients on antiretroviral therapy.

Limitations Of The Study
Although, this study has attempted to show factors affecting the dietary diversity of PLWHA, it has some limitations which should be taken into consideration. We didn't measure food weight to estimate calorie quantities in foods, and during measuring dietary diversity, recall bias was one of the likely constraints of the study. We also did not address seasonality when administering the FANTA HDDS questionnaire. Declarations Ethics approval and consent to participate Ethical clearance was obtained from Institute of Ethical Review Board of University of Gondar. An o cial permission letter was obtained Debre Tabor hospital. Written informed consent was obtained from study participants in their local language after explaining the purpose of the study, potential risks and bene ts of the study, and the right to withdraw from the study at any time. The participants were also assured that the data was con dential.

Consent for publication
Not applicable.
Availability of data and materials' Data will be available upon request from the corresponding authors.

Funding
No fund was obtained for this study Authors' contribution KW conceived the study, developed the tool, coordinated the data collection activity, and carried out the statistical analysis. EA, MTH and GA participated in the design of the study and tool development, performed statistical analysis, and reviewed the manuscript. AK involved in the tool development, and performed statistical analysis. All authors read and approved the nal manuscript.