Prevalence and associated factors of physical inactivity among adult diabetes mellitus patients in Felege Hiwot Referral Hospital, Bahir Dar, Northwest Ethiopia

Physical inactivity has been viewed as an emerging public health problem in developing countries including Ethiopia. Diabetes mellitus (DM) is a group metabolic disorder characterized by chronic hyperglycemia resulting from defects in insulin secretion, function, or both. Its prevalence increases with changing lifestyles including physical inactivity across the globe. However, there is limited research, and not yet received attention in Ethiopia. This study aimed to assess the prevalence and associated factors of physical inactivity among adult diabetic patients in Felege Hiwot Referral Hospital, Bahir Dar, Northwest Ethiopia. An institutional-based cross-sectional study design was conducted among 308 participants from February to June 2018 at Felege Hiwot Referral Hospital. A face-to-face interview was conducted using a structured questionnaire by trained data collectors. Participants were selected through a systematic random sampling technique. Physical inactivity was assessed by the international physical activity questionnaire (IPAQ). Collected data were entered in Epi info version 7 and transferred to SPSS version 20 for analysis. A summary of descriptive statistics and multiple binary logistic regression analyses were computed to identify associated factors of physical inactivity among adult diabetic patients. P < 0.05 with 95% CI was considered statistically significant. The overall prevalence of physical inactivity among diabetic patients was 30.5% ( 95% CI: 22.8–33.5%). Gender (AOR = 1.5, 95% CI: 1.1, 3.62), Old age (AOR = 18.17, 95% CI: 22.7, 61.9) Residence (AOR = 4.24, 95% CI: 1,12,16.028), Low self-efficacy (AOR = 20.59, 95% CI: 10.598, 41.608), Poor attitude (AOR = 2.75, 95%CI: 1.44,5.28), and Lack of social support (AOR = 4.22, 95% CI: 1.28,4.07) were found significantly predictor factors of physical inactivity. The prevalence of physical inactivity in this study was high. Being female, old age, dwelling in an urban, having low efficacy, poor attitude, and lack of social support was greater risk factors for being physically inactive. Diabetic education should focus on engagement in physical activity by overcoming barriers to performing physical activity. Government and health professionals should emphasize that evidence-based physical activity important to change their attitudes and require reaching a consensus on social support by their families.

The study was conducted at Felege Hiwot Referral Hospital, which is found in Bahir Dar city administration Northwest, Ethiopia. Bahir Dar is the capital city of the Amhara region and is located 550 km from Addis Ababa, the capital city of the country. Felege Hiwot Referral Hospital is a tertiary health care level hospital serving around 12 million population of Bahir Dar town and remote areas of Northwest, Ethiopia. Currently, it is the only governmental Hospital in the city and the regional referral hospital serving the population in the region as a referral center. It provides promotive, curative, and rehabilitative services to those populations. The reason to select this study area was that DM patient flow is high and the principal investigator noticed a lack of adequate recommended physical activity education for DM patients by healthcare professionals.
Sample size and sampling procedure. The sample size was calculated by assuming a 5% margin of error, 95% confidence interval (alpha = 0.05), and 50% proportion of physical inactivity among DM patients since no study has been done in the study area with the same population. Based on this assumption, the sample size was calculated by using the single population proportion formula, and by considering a 10% non-response rate, the final sample size reached 308.
A systematic random sampling technique was used to select the study participants. The diabetic clinic provides its service two days per week (i.e. Monday and Tuesday); on average 170 DM patients were treated per week. Based on this, 1020 DM patients may be expected to follow up within six weeks data collection period. The sample interval was determined by dividing the expected number of DM patients (1020) during the course of the data collection period divided by the sample size (308) of the study which gave a sample interval of 4. Thus, every four patients were interviewed until the total sample size was reached. The first respondents were selected randomly with a lottery method. www.nature.com/scientificreports/ Data collection tools and procedures. Data was collected through face-to-face interviews, recorded from patient files, and measuring anthropometric characteristics. Body weight was measured by a weight scale without shoes. Height, waist, and hip circumferences were measured by using a non-elastic measuring tape. Waist circumference was taken at umbilicus level directly over the skin whereas hip circumference over the buttock area with clothing in a standing position 31 . As per WHO guidelines the cut-off point for waist and hip circumference ratio to the normal range for both men (< 0.95 cm) and women (< 0.85 cm) 21 .
Four data collectors (two physiotherapists and two nurses) were trained by the questionnaire investigator for two days before data collection. The questionnaire was pretested on 5% 15 of the sample among DM patients who were not included in the main study area, in Gondar university hospital to detect potential problems, unanticipated and interpretations as well as to amend cultural issues to each question. The questionnaire was modified and corrected based on the pretest result. Furthermore, on each data collection day, the collected data were checked and reviewed to completeness by the main investigator.
Operational definition. Physical inactivity. Is those respondents who will not achieve WHO recommendations of total physical activity (level less than 600METs-minute/week) or reported practicing no physical activities will be classified as inactive. Whereas those who meet this criterion(> 600METs-minutes/week) will be classified as physically active 21 .
Cigarette smoking. Those who smoke any tobacco products daily will be considered tobacco users 33 . Alcohol drinking. Those who consume any type of alcohol 2 or more bottles daily will be considered alcohol users 34 . Good attitude. Those respondents who scored equal to or above the mean value of attitude towards physical activity score 35 . Poor attitude. Those respondents who scored below the mean value of attitude towards physical activity score 35 .
High self-efficacy. Those respondents who scored equal to or above the mean value of self-efficacy towards physical activity score 36 .
Low self-efficacy. Those respondents who scored below the mean value of self-efficacy towards physical activity score 36 . Lack of social support. Those participants who have no family, friends, or peer support to do physical activity 37 .
Lack of adequate facilities. Facilities not available to perform physical activity 37 .
Waist-hip-ratio (WHR). Person's waist circumference divided by hip circumference. Based on WHO 2008, convened participants who have in the normal range for both men (< 0.95 cm) and women (< 0.85 cm) 38 .

Measurement.
A structured questionnaire was prepared to collect sociodemographic information, behavioral, psychological, and environmental factors. Comorbid factors and DM-related complications were recorded from the patient's files. Physical inactivity was assessed by the international physical activity questionnaire (IPAQ). It assesses the frequency (days) and duration (hours or minutes) of vigorous-intensity, moderate-intensity, walking, and sitting activities during the last seven days 39 .WHO assigned a value of 8METs, 4METs, and 3.3 METs for the time spent in vigorous, moderate, and walking activities respectively. As per WHO guidelines, physical activity was calculated as the total MET-minute/week.
A combined total physical activity MET-min/week could be computed as the sum of Walking + Moderate + Vigorous MET-min/week scores. Then participants were classified as inactive if the participants were scored less than 600 MET whereas the participants classified in to physically active if the participants were scored greater than 600 MET per week 40 . The questionnaire was translated from the English version to the Amharic version by native Amharic speakers and translated back to English to insure its contingency.
Data processing and analysis. Data  www.nature.com/scientificreports/ as an odds ratio (OR) and 95% confidence interval (CI). All predictor variables that have an association in bivariable analysis with a p-value ≤ 0.25 41 were entered into a multivariable logistic regression model to identify independent factors associated with dependent variables by controlling the effect of confounders. The model was built by entering method and its fitness was assessed by the likely hood ratio test and Hosmer-Lemeshow goodness-of-fit statistic test. The crude and adjusted odds ratios together with their corresponding 95% confidence intervals with P-value < 0.05 were computed and interpreted accordingly.
Ethics approval and consent to participate. The study was conducted as per the Helsinki Declaration for biomedical research. Ethical approval and clearance were obtained from the University of Gondar, College of Medicine and Health Science, Institutional Review Board wth a reference number of 211/2018. An ethical clearance letter was submitted to the Felege-Hiwot Referral Hospital and a permission letter was obtained. Informed written consent was taken from each Adult.

Results
Socio-demographic characteristics. Three hundred two study participants were involved with a response rate of 98%. Those six diabetic patients who were selected for a sample could not participate due to refusing to take the interview. Of the total respondents, the majority of them 174 (57.6%) were male, 235 (77.8%) were orthodox Christians and 88 (29.9%) were government employees. In terms of marital status, more than two third of respondents 210 (69.5%) with Diabetic Mellitus were married and 38 (2.6%) were widowed. Occupational status of participants with DM, among the total respondents 88 (29.1%) were government employees, and 72(23.8%) were housewives. Regarding age distribution, the mean and standard deviation of participants was 49.4 ± 10.5. As to the type of Diabetic Mellitus, 221 (73.2%) were type two and 176 (58.3%) of participants were less than the seven-year duration of DM. Likewise, the median monthly income of respondents was 3540 ETB (≥ 3000 ETB) and 101 (33.4%) of the participant attended college/university level (Table 1).
Anthropometric and behavioral characteristics. Thirty-six participants (11.9%) declared that they were smoking cigarettes daily and 22 (7.3%) of them were also alcohol consumers. Most of the participants 145 www.nature.com/scientificreports/ (48.0%) were in the normal weight category on BMI whereas 120 (39.7%) and 37 (12.3%) were overweight and obese respectively. Regarding their central obesity, more than half of the participants 167 (55.3%) were in a normal category on WHR (Table 2).
Psychological and environmental factor characteristics. Of  Prevalence of physical inactivity among DM patients. The overall prevalence of physical inactivity among participants with adult DM was 30.5% (95% CI: 22.8-33.5%) when a total metabolic equivalent which is a physical activity measurement cut of point 600 MET-minute/week was used. In contrast, 210 (69.5%) participants fulfilled the criteria of the physical activity category which was 600 MET-minute/week and above. Among the active group, the majority of the respondent 153 (50.7%) were performing moderate physical activity while only 57 (18.9%) were performing vigorous physical activity (Fig. 1).

Factors associated with physical inactivity among DM patients.
After controlling potential confounders variables such as gender, old age, place of residency, self-efficacy, attitude towards physical activity, lack of social support, and retinopathy were significantly associated with physical inactivity among diabetic patients (p < 0.05). Diabetic patients who were female participants were 1.5 times more likely to be physically inactive (AOR = 1.5, 95% CI: 0.64, 3.62) than males. As well as those who are residing in urban areas were 4.24 more likely to be physically inactive (AOR = 4.24, 95% CI: 1.12, 16.02) compared with participants residing in rural areas. Regarding the age of participants, DM patients whose age is 60 years and above were 18.169 times more likely to be physically inactive (AOR = 18.16, 95% CI: 22.74, 61.95) compared with the younger age group.
Regarding self-efficacy and attitudes of participants toward physical activity, those respondents with low selfefficacy were 20.59 times more likely to be physically inactive (AOR = 20.59, 95% CI: 10.59, 41.60) than the high self-efficacy group. In addition, those who had poor attitudes also are 2.753 times more likely to be physically inactive (AOR = 2.75, 95%CI: 1.44, 5.28) than those who had a good attitude toward physical activity.
Individuals who lack social support were 4.22 times more likely to be physically inactive (AOR = 4.22, 95% CI: 1.28, 4.07) compared with those who have good social support (Table 6).

Discussion
Physical inactivity is continually rising in many developing countries like Ethiopia and has been considered the fourth leading cause of global mortality 16 . Despite its huge negative impact, little is known about physical inactivity in patients with DM in Ethiopia particularly study area. Therefore, this study sought to assess the prevalence and associated factors of physical inactivity among adult DM patients in Felege Hiwot Referral Hospital North West Ethiopia.
This study found that the prevalence of physical inactivity among adult DM patients was 30.5% when the total metabolic equivalent cut-off point score (< 600 MET-minute/week) as per WHO guidelines, the international physical activity questionnaire (IPAQ) measuring tool was used 42 . Outdoor walking is the most commonly reported physical activity among the respondents. It might be due to the cost of joining Gym and exercise facilities, fear of being injured by practicing high-intensity activities, and self-belief like being embarrassed to wear sportswear, especially in old age and women.
This study's prevalence was comparable with a similar study conducted in Nigeria found out relatively similar prevalence which was 31% 13 . It was also compared with the 31% prevalence reported from Manila, Philippines with a similar tool IPAQ used. This similarity might be due to similar methodology and measuring tools used 26 .
Prevalence of physical inactivity in this study was lower than in institutional based cross sectional studies conducted in New York, America (40%), Brazil (47.9%), Spain (35.7%), and Istanbul, Turkey (39.5%) [23][24][25] . The discrepancy between these studies may be due to methodological differences like sample size and study period. Another possible reason might be the socio-economic pattern means the majority of work-related activities changed to computerization and increased car ownership may reduce the chance to perform physical activity in www.nature.com/scientificreports/ those study participants while the majority of the respondents in our study expressed the long working hours represent the largest component of physical activity 43 .
In the study conducted in Saulo, Brazil, the prevalence of physical inactivity was 17.5% 20 . The possible reason for this difference might be the sample size and study participants. The Brazilian study includes only the urban population and unlike Ethiopia, the Brazilian federal government developed and disseminated national physical activity guidelines to on emphasize physical activity could be another reason 44 .
This study revealed that women participants were 1.5 times more likely to be inactive than male diabetic patients. This finding was supported by previous studies 14,21,45,46 . This is might be due to the women in developing countries like Ethiopia, commonly participating in light-intensity household chores activities such as housekeeping, carrying the babies, and food preparation activities which are routine daily activities. In urban areas these www.nature.com/scientificreports/ daily household activities commonly are carried out by domestic helpers whereas men are tending to participate vigorous activities like playing a sport, running, and cycling. Furthermore, in this study, aged 60 and above was more likely to be physically inactive compared to younger age. This finding is in line with the studies conducted in Lebanon, Srilanka and Barcelona 12,14,24 , there is a positive significant association between an increase in age and decreased physical activity level on DM patients. The possible reason of this might be participants of this age group retired person and may be associated with a sedentary life style after retirement. Furthermore, older people are reluctant to do physical activity due to their perception to that DM weakened their body and causing them to have some demotivation effect to perform regular physical activity and feeling tiredness.
Moreover, urban residency was more likely to be physically inactive compare to rural residence. This finding is in line with previous studies done in Nigeria, Brazil and Barcelona 13,20,24 showed that physical inactivity among those participants residing in urban (34.3%) area was higher than rural (14.0%) population. This might be due to participants who residing in urban area may distracted by television long time, computerization and mostly use passive mode of transport while majority of those rural participants perform high demanding work activity like farming and fishing as well as lack public transport access may a factor.
Regarding to the psychological attributes, this study agreed with other previous studies 2, 21,29 revealed that low efficacy and poor attitude towards physical activity were a major predictors of physical inactivity among adult DM patients. In this regard participants who had low efficacy towards physical activity were 20.59 times more likely physically inactive compared with those high efficacies. Furthermore, participants who had poor attitude  www.nature.com/scientificreports/ towards physical activity also 2.75 times more likely risk of physically inactive compared with those who had a good attitude towards physical activity. This may arise by wrong belies and poor knowledge about benefit of physical activity as well as inadequate educational intervention and adapts self-care behavior, particularly physical activity for better management of the diseases. Concerning to the environmental factors, our study revealed that lack of social support have statistical positive significant with physical inactivity on diabetic patients. Those participants who had lack of social support (family, peers, and friends) were 4.22 times more likely to be inactive compared with those who had social support. This study was supported by cross sectional studies conducted in Hamadan and UK 2,28 . This is may be due to that families or friends of participants who had no social support may have no money to pay in exercise facilities, recreational facilities which needs to pay money and may have also poor knowledge about the benefit of physical activity.
Unlike other studies so far 17,24,47 behavioral factors such as smoking cigarettes, alcohol consumption and anthropometric measurements like general obesity (BMI) and central obesity (WHR) were not significant associated in our study. This might be because of low prevalence in smoking and alcohol users in our study. As well as those who were overweight, obese and higher central obesity individuals being motivated to engage physical activity in order to have weight loss.
Limitation of the study. To the best of investigators knowledge, this is the first study that assessed the magnitude and determinant factors among DM patients in Ethiopia. Due to this, it was impossible to compare results with previous findings. This can be a limitation of this study. In addition, physical inactivity was assessed by using Questionnaires (subjective method approach) which inherited to recall bias. However, this instrument is internationally accepted reliable, and validated tool for cross-sectional study worldwide 42 .
Despite these limitations, it is a pioneer study in the country and adds to the limited body of evidence by providing useful data on the prevalence of physical inactivity among adult diabetic patients in Ethiopia, particularly in this study area.

Conclusion
Overall this study found that almost one-third (30.5%) of participants with DM did not practice recommended physical activity level in Felege Hiwot Referral Hospital Bahir Dar, Northwest Ethiopia. Outdoor walking was the most commonly reported practiced physical activity.
Gender, older age, place of residence, low self-efficacy, poor attitude towards physical activity, lack of social support, and diabetic related complication (retinopathy) were significantly associated with physical inactivity among adult diabetic patients. Diabetic education should focus on engagement in physical activity by overcoming barriers to performing physical activity. Government and health professionals should empasize evidence based physical activity important to change their attitude and require to reach consensus on social support by their family.

Data availability
The datasets used and/or analyzed during the current study are available from the corresponding author based on reasonable request.