Physical activity level and associated factors among adult HIV patients in Ethiopia

This study was conducted using a cross-sectional design in three selected hospitals, one comprehensive specialized hospital and two general hospitals from May to June 2019. Hawassa University Comprehensive Specialized Hospital (HUCSH), a tertiary level hospital, delivers specialized and referral services for general hospitals. The two general hospitals, Adare and Yirgalem, deliver secondary level health care, providing preventive and curative services that require diagnostic facility and therapeutic intervention [16]. HUCSH and Adare general hospitals are found in Hawassa town, the capital of the Sidama regional state, and the South Nation Nationality Peoples Region (SNNPR) of Ethiopia, which is located, 275 km south of Addis Ababa, the capital of Ethiopia. HUCSH, Adare, and Yirgalem general hospitals at the beginning of this study gave ART service for 2553, 1821, and 1476 HIV patients respectively.

The study sample was selected from adult HIV patients (18 years plus), enrolled in ART care and visiting the three selected hospitals during the study period. Pregnant and lactating women were excluded from the study. A single population proportion formula was used to calculate the sample size. So far, there was no similar study conducted in the area or elsewhere having the same status as in Ethiopia. Therefore, the estimated proportion of 50% was taken to have maximum sample size. A sample of 384 was obtained and by considering the 10% non-response rate, the final sample size became 422. Proportional allocation was used to determine the number of study units to be sampled from each facility. Based on this, 184, 131 and 107 study participants were selected from HUCSH, Adare, and Yirgalem general hospitals, respectively. Individual study participants were selected by random arrival at the ART clinic.

Data were collected through interview administered questionnaires, conducted at routine consultation by nine trained nurses. A structured questionnaire was used to collect information on socioeconomic and ART-related characteristics of the study participants. To collect data on the level of physical activity of the participants, we used the the short form IPAQ [17]. Since several examples of physical activity were not regular activities in Ethiopia, we replaced these by other physical activities with approximately the same Metabolic Equivalent of Task (MET) [18]. In the category of vigorous physical activity, fast bicycling was replaced by rope jumping. In the category of moderate physical activity, bicycling at a regular pace and double tennis was replaced by cleaning and gardening. After the cultural adaptation, the questionnaire was translated to Amharic and retranslated to the original version to check for consistency. We have presented the data on physical exercise as a continuous score using MET-min per week (MET level × minutes of activity × events per week) or as a categorical variable in three categories: low, moderate, and high. Participants who did not fulfill the criteria of moderate and high were considered as low active or inactive.

Participants who fulfilled one of the following criteria were categorized in the moderate group.

Participants were categorized into the high activity group, if they fulfilled one of the following criteria:

Dietary diversity and household food insecurity data were collected using the Food and Nutrition Technical Assistance (FANTA) indicator guide for Household Dietary Diversity Score (HDDS) [19] and Household Food Insecurity Access Scale (HFIAS) [20] respectively. Anthropometric measurements, such as height, was measured using a stadiometer (Seca Germany) by positioning the patient at the Frankfert plane recorded to the nearest 0.1 cm. Weight was measured using a pretested and calibrated digital Seca^® scale and recorded to the nearest 0.1 kg.

Blood pressure (BP) was measured with the standard mercury sphygmomanometer BP cuff with the appropriate cuff size by measuring the left arm consistently, three times at a 5-min interval. The average of the two last readings was taken, and the diagnosis of high BP (hypertension) was made according to the WHO criteria as systolic BP ≥ 140 mmHg or diastolic BP ≥ 90 mmHg [21]. Random blood glucose levels were determined by using Fia Biomed Blood Glucose Meter (Glucometer) Salut by finger puncture. According to the American Diabetes Association’s guideline, fasting plasma glucose levels ≥ 126 mg/dL, 2 h plasma glucose ≥ 200 mg/dL during oral glucose tolerance test, hemoglobin A1C ≥ 6.5% and in a patient with classic symptoms of hyperglycemia or hyperglycemic crisis a random plasma glucose ≥ 200 mg/dL are defind as diabetes [22]. Participants who were unaware of the fact that they had diabetes and/or hypertension were linked to the respective hospital for further diagnosis and management of their conditions.

Data were analyzed using SPSS for windows version 20.0 (IBM, USA). Descriptive statistics were presented in the form of frequency, percentage, mean and standard deviation. Chi square test was used to assess the association between categorical variables. Variables having P-value < 0.25 in the bivariate logistic regression analyses were considered as potential candidates in the final multivariable logistic regression analysis. P-value < 0.05 was used to declare statistical significance in the multivariable model. The overall goodness of fit of the model was checked by using Hosmer–Lemeshow. Finally, the adjusted odds ratio (AOR) with its 95% confidence interval (CI) was used to determine statistical significance.

We included a sample of 422 adult HIV patients receiving ART service from three selected public hospitals. Most participants (64%) were women and from an urban area (93%), with a mean age of 38.7 ± 9.13 years. Nearly half of the participants (48%) were married and had a monthly household income of ≤ 1500 Ethiopian Birr (52%). The majority of participants (37%) had completed secondary education and about 28% were privately employed (Table 1).

More than half of the participants (58%) had a recent CD4 count that was ≥ 500 cells/mm³ and most patients had normal BMI (59%). Majority of the participants (92.2%) were in clinical stage I, had been on treatment for ≥ 2 years (70%), had no chronic comorbidity (11%), and household food insecurity (75%) (Table 2).

We observed significant relationship between gender and PA level (χ² = 6.08, P ≤ 0.048), where men were, in general, more physically active than women. Participants living in rural areas were found to have more overall PA (P = 0.007) compared to other urban living study participants (Table 3).

More than half of the participants, 288 (68%) had low PA levels, followed by moderate, 75 (18%) and high 59 (14%) respectively (Fig. 1). More, 194 (72%) women were physically inactive compared to men 94 (61%) (Fig. 2).

In bivariate logistic regression analysis; gender, place of residence, duration of treatment and HHFIS were factors associated with physical activity level. In multiple logistic regression analysis; gender, place of residence and duration of treatment remained significantly associated with physical activity level of participants. Women participants were two times [(AOR = 1.64, 95% CI (1.07, 2.53)] more likely to be physically inactive than men. Participants from urban residence were three times [(AOR = 2.57, 95% CI (1.16, 5.72)] more likely to be physically inactive than rural patients. Participants, who had been on ART for ≥ 24 months were nearly two times [(AOR = 1.88, 95% CI (1.15, 3.08)] more likely to be physically inactive than those who were on ART for < 24 months (Table 4).