Study design
A retrospective review was conducted among patients included in the GT intervention conducted between February 2019 to August 2019 at a health-center in Pind Begwal, Pakistan. The center is owned by Saving 9, a registered non-profit organization. Information regarding patient demographics (age, gender), number of follow-up visits, examination findings (weight, waist circumference, blood pressure, height) and lab results (HbA1c, FBG, random blood glucose (RBG), creatinine, eGFR) were collected. Ethical exemption to review unidentified information was obtained from Ethical Review Committee at the Aga Khan University. (Review reference number: 2020-5221-13953).
Application development
The application was designed to ensure compatibility with consumer literacy, app security and user privacy. To meet these criteria, design and development was conducted by a multidisciplinary research team including, experts in software development, data privacy and security, a graphical user interface designer and language editor, and a health professional. The app has three main features: (I) Entry and record of patient data (study identification number, age, gender, weight, height, waist circumference, blood pressure) and lab investigations (FBG, RBG, HbA1c, creatinine, eGFR); (II) review of data by a remote physician (III) entry of treatment plans (IV) privacy (no medical data leaves the device) and restricted access to patient information for users of the application. It was available on the Google Play Store for android mobile devices.
Community involvement
Lady health workers (LHW) from Pind Begwal (PB) who were literate in written English and owned a smart phone compatible with the application were selected. Three sessions were conducted to train them to measure FBG, RBG, height, weight, waist circumference and blood pressure, operate the GT application and relay the treatment plan to patients. Equipment required for measuring the aforementioned variables (glucometer, weighing scale and automatic blood pressure monitor) was provided to each LHW.
Patient recruitment
An advertisement campaign was conducted for the intervention through distribution of pamphlets at the health-center and through home visits by LHWs. Patients were requested to visit the healthcare center for study enrollment where convenience sampling and was used and consent obtained to recruit patients with an established diagnosis of diabetes defined as, HbA1c > 6.5 and FBG > 126mg/dL or RBG > 200 mg/dL. Patients with end-stage diabetes complications including end-stage renal disease, chronic heart failure, peripheral vascular disease leading to foot ulceration/ amputation, proliferative retinopathy, gestational diabetes, non-ambulatory patients and patients with advanced medical comorbidities requiring frequent hospitalization such as active cancer, advanced liver disease etc. were excluded.
The intervention
HbA1c, FBG, blood pressure, waist circumference, weight, height, creatinine, eGFR, comorbid conditions and the patient’s current medications were recorded at the time of recruitment and entered into the application. A remote endocrinologist, licensed to practice in Pakistan, was selected to review patient information and prescribe appropriate treatment according to the guidelines.[9] he treatment plan was entered into the application was consisted of medication dose and frequency, diet and exercise recommendations and the next recommended follow-up appointment timing. The LHWs conveyed the treatment plan to the patient through a home visit or telephone call. The patients were then followed up via home visits for a period of six months. The frequency of visits for each patient was according to the discretion of the endocrinologist. The patient was also examined at each follow-up visit where fasting blood sugar, blood pressure, waist circumference, and weight were recorded and the updated measurements were recorded in the application.
Outcomes
The primary outcome consisted of glycemic control parameters including HbA1c and fasting blood glucose. Secondary outcomes included ancillary health parameters including BMI, weight, waist circumference and blood pressure.
Statistical analysis
All analysis was conducted using SPSS software version 22. Missing data from a visit was input with readings from the previous follow-up visit provided it occurred within one month. Mean, standard deviation and range of HbA1c, FBG, BMI, weight blood pressure and waist circumference was calculated for baseline and final readings. Due to high fluctuation in FBG readings, mean and standard deviation for values in the final two months of the patient’s participation were calculated for each patient and used as final values. Distribution of variables was assessed using the Shapiro Wilk test. Mean and standard deviation was calculated for normally distributed variable and a paired t-test was used to compare final and baseline variables. Median and interquartile range was calculated for variable without normal distribution. Wilcoxon signed rank test was used for comparison. P value below 0.0.5 was considered significant.