Study design, setting and study population
Institution -based cross-sectional study was conducted at Debre Markos Referral Hospital, Debre Markos, Northwest Ethiopia from April 1-30/ 2019. Debre Markos is a town in Amhara region, and found 300 km away from Addis Ababa, the capital city of Ethiopia. All diabetic patients fulfilling the inclusion criteria during the study period were included in this study.
Sample size determination
The required sample size for the study was computed by Epinfo using single population proportion formula with the following assumptions: taking 95% confidence interval (CI), 5% margin of error, and 41.4% proportion (P) of DR which was taken from a study conducted in Jimma University Specialized Hospital (17) . In addition, correction formula was applied since the source populations were less than 10,000 and 10% non-response rate was considered. Accordingly, a total of 302 study participants were included in this study.
Sampling technique and procedures
Systematic random sampling method was used to select study participants. Less than three years diabetic follow up for type I diabetes mellitus were excluded. Debre Markos Referral hospital diabetic follow-up clinic provides service from Monday up to Friday every week. There are a total of 2018 DM patients currently attending follow up at Debre Markos Referral Hospital. Every 3rd patients were systematically selected from the sampling frame to get a sample size of 302. Patient’s registry book was used as a sampling frame.
Operational definition
Diabetic retinopathy: On the retinal camera examination the presence of microaneurysm, hemorrhage, exudate, cotton wool spot, intra-retinal microvascular abnormalities, vein beading and/or new blood vessels at least one of the eyes.
Pre-proliferative DR: - On the retinal camera examination the presence of microaneurysm, cotton wool spot, hemorrhage, vein beading, exudate and/or intra-retinal microvascular abnormalities
Poor glycemic control: - Average of six consequents follow up fast blood sugar greater than 130 mg/dl.
Data collection technique and procedure
Data were collected using semi-structured questionnaire, standard stadiometer, digital weight scale, android digital sphygmomanometer, Snellen chart, slit lamp bio microscopy and binocular indirect ophthalmoscope. Blood pressure was measured by using android digital sphygmomanometer through keeping the respondent in a seating position. If a systolic blood pressure is 140 mm Hg or more and/or a diastolic blood pressure is 90 mm Hg or more two consecutive measurements apart from four hours and ongoing treatment with antihypertensive drugs taken as hypertensive respondent. Height was measured with a moveable headboard (stadiometer) and recorded to the nearest 0.1 centimeter. Weight was measured by digital weight scale and recorded to the nearest 0.1 kg (18).
Visual acuity test was tested separately at 6 meters in well-illuminated area by using Snellen chart. The participant’s vision was too poor to read any letters on the chart at 6 meters, then counting finger, hand movement and light perception consequently was done (19).Slit lamp biomicroscope (HAAG Streit Bern 90032742, Swiss Made) and binocular indirect ophthalmoscope (HEINE EN53®, Germany) involved for assessment of retina. Dilation of the participants pupil with 1% tropicamide were done before retinal examination. Fine and short beam size with small pinpoint of light was used to good illumination. Anterior segment examination was carried out by using a slit lamp and posterior segment examination was carried out by using a binocular indirect ophthalmoscope (20,21).
Data quality management
The questionnaire was translated into local language (Amharic) from its English version then back to English with the guidance of ophthalmologist. Training was provided for the data collectors for one day on how the data should be obtained and recorded. Pre-test was done on 16 DM patients at Fenoteslam General Hospital before the actual data collection time in order to see the validity of the instrument, to estimate the time needed to collect data, and to modify the questionnaire accordingly. Data collectors were supervised while collecting the data by the principal investigator and technical support was provided accordingly. Data were checked daily for completeness and consistency throughout the data collection period.
Data processing and analysis
Completeness and consistency of data collection forms were examined during data management, storage, cleaning and analysis. After data were entered to epi Data 3.1version, it was exported into Statistical Package of Social Science (SPSS) version 20 for analysis. Data were checked, cleaned, coded, merged, and categorized before analyzed. Continuous variables were expressed in terms of mean ± standard deviation while categorical variables were expressed in proportion. Descriptive statistics and cross tabulation were done for each independent variable. Binary logistic regression with 95% confidence interval was used to assess the association of variables with diabetic retinopathy. A variable had p-value <0.25 in the bivariable logistic regression was entered to multivariate logistic regression. Backward stepwise logistic regression was used for multivariable binary logistic analysis. In the multivariable logistic regression p-value <0.05 was considered as a significant associated variable with diabetic retinopathy. Finally, the results were summarized and presented by text, tables, charts, and graphs.