Study design, setting and period
This institution-based cross-sectional study was conducted at Dessie town Hospitals (both government and private Hospitals) from 15 February to 15 March 2020. Dessie town is located 400 km away from Addis Ababa, the capital city of Ethiopia. There were four private and one governmental referral Hospitals in Dessie town. According to annual summative report of all diabetic clinics in Dessie town hospitals, an estimated number of 14,000 people living with T2DM were served. In each Hospital, there was diabetic follow-up clinic for people living with DM (10210 clients for Dessie referral Hospital, 1440 clients for Selam general Hospital, 1250 clients for Ethio general hospital and 1100 clients for Bati general Hospital). The follow-up date in each hospital was from Monday to Friday.
Study population and eligibility criteria
All people living with T2DM visiting diabetic clinic of each Hospital during data collection period were participants and those T2DM who had follow-up and newly diagnosed patients at the same period were included in the study. People living with T2DM who were seriously ill, pregnant women and those who had HIV/AIDS, trachoma, acute eye infections and trauma to the eye and history of head injury, history of stroke were excluded from the study.
Sample size determination and sampling technique
Sample size was determined using a single population proportion formula with assumptions; p=0.289 (30) (proportion of T2DM with visual disturbance) , 95% CI, and 5% margin of error (d=0.05). The minimum sample size was 316 and the final sample size was 332 after adding 5% oversampling to account for unpredictable events during data collection.
Samples were recruited using simple random sampling technique (using lottery method) and allocated proportionally to the four hospitals based on the number of T2DM clients at each hospital.
Operational definitions
Visual impairment: Any loss or abnormality in an anatomical structure or a physiological or psychological function (5). It is presenting visual acuity of worse than 20/40 to no light perception (NLP) in either or both eyes that includes both low vision and blindness (31).
Low vision: Even with corrective lenses, it is inability to clearly see at a distance of 6 meters (20 feet) that individuals with normal vision can clearly see at a distance of 12 meters (40 feet) or visual acuity between 6/12 and 3/60. Low vision includes mild visual impairment (VA between 6/12 and 6/18), ‘moderate visual impairment (VA between 6/18 and 6/60)’ and ‘severe visual impairment (VA between 6/60 and 3/60)’ from all causes (7).
Blindness: Inability to read the largest letter on a vision chart at a distance of 3 meters (10 feet) or visual acuity was worse than 3/60 (32).
Visual acuity/VA: Simple, non-invasive measure of the visual system’s ability to discriminate two high contrast points in space. It is usually taken at a distance of 6m or 20 feet (16). It is expressed in terms of A/B, where A is the distance between observer and letters and B is the expected distance that the normal eye can see.
Obesity: Body Mass Index (BMI) of ≥30 kg/m2. Subtypes include class 1 obesity (BMI: 30–34.9 kg/m2 ) class 2 obesity (BMI:35–39.9 kg/m2) class 3 extreme obesity or morbid obesity ( BMI: ≥40 kg/m2) (33).
Exercise: A person who experienced regular exercise below 150 minutes (3-5days) per week was considered as having poor regular exercise otherwise it was considered as having good regular exercise (34).
Regular follow-up: Clients could visit the diabetic clinic of each hospital every 3 months.
Data collection instrument, procedure and quality control
Pre-tested, semi-structured-interviewer-administered questionnaire containing Socio-demographic variables, behavioral variables (diet, exercise and regular follow-up), clinical related variables (glycemic control, plasma glucose level, duration of DM, treatment option for DM, and presence of comorbidities such as hypertension and/or obesity) were used to collect data. Study participants were scheduled for eye examination at the eye clinic of the same institution on the same day. The participants had gotten counseling, care and referral depending on the ocular findings. We used tape meter, and weight balance for measuring height and weight, respectively. Each participant had visual acuity assessment with illuminated Snellen’s chart for each eye at 6 meter. Training was given to data collectors and the supervisor by principal investigator about the objectives of the study, data collection techniques and ethical issues. Data collectors were four (2 BSc ophthalmic Nurses and 2 BSc Nurses) and one supervisor (BSc Public health). Strict supervision was taken during data collection process.
Reliability and validity
The Cronbach’s alphas for this study was below 0.5 which is due to the questionnaires were self-developed and unstandardized. Content validity was ensured by pre-testing the data collection tool on 17 students out of the study area. The tool was modified based on the observed findings from the pre-test result. Some questions having ambiguous meaning were rewritten for better understanding of study participants.
Data processing and statistical analysis
Data were entered into Epi-Data 3.1 and exported into SPSS 22 for statistical analysis. Continues data were described by median and interquartile range (IQR) whereas frequency with percent was used to describe categorical variables. Bi-variable binary logistic regression analysis was performed to select potential candidate variables for the final model with cut-off point of p-value ≤ 0.25 (35). Multivariable binary logistic regression analysis was executed to identify predictors of visual impairment. The measure of association was described by crude odds ratio (COR) and adjusted odds ratio (AOR) with their 95% CI. In the final model, variables with a p-value ≤0.05 were considered as statistically significant. Model fitness was checked by Hosmer and Lemeshow goodness of fit test (at p>0.05) and multicollinearity was tested by VIF.