Using mHealth (Mobile Health) to Optimize Glycemic Control in Adults With Type 2 Diabetes: Proof of Concept Study

A sample size of 60 a-priori (n=30 intervention, n=30 control arm) was determined to assess feasibility of text messaging in optimizing glycemic control. It is anticipated that this proof-of-concept study will obtain parameters that will be used in determining a robust power calculation for a fully powered efficacy trial. Purposive sampling will be used to select two hospitals, one in a rural setting and another in an urban setting in Nakuru County, Kenya. Multistage random sampling using computer-generated list of random numbers will then be applied to randomly assign 15 participants to either the intervention arm or control arm. The hospitals' diabetes registers will be used to select patients on regular diabetes care from the hospital patient records. The intervention group will receive a total of 43 text messages developed using the Behavior Change Wheel (BCW) approach. The text messages that are linked to specific behavior change techniques will provide practical information or guidance to influence selection and eating of healthy diet for type 2 diabetes. Additionally, two-way text messages, will be sent to study participants to rate their ability on selection of food and eating of healthy diet based on the messages received after every 4 weeks. The combination of the one-way and two-way messages in this in the intervention arm are designed to increase engagement of the participants. The control group will continue receiving standard care in the hospitals. This group shall also receive clinic appointment reminder text messages one day prior to the actual date. Quantitative data will be analyzed by the intention to treat principle. Various patient characteristics and primary outcome measures, including anthropometric and biochemical measurements (HbA1c, Fasting Blood Glucose (FBG), lipid profiles) at baseline will be summarized using descriptive statistics. The descriptive statistics will include means and standard deviations for continuous variables and proportions for categorical variables. HbA1c levels will be compared at 12 weeks between the two arms using an ANCOVA which will control for baseline HbA1c levels. The intervention and the control groups will be followed-up for six months to assess sustainability and mHealth satisfaction of the intervention. One-way and two-way text messaging data will be analyzed descriptively. Subgroup analysis per hospital and participant socio-demographics including age, sex, education and income levels will be conducted. All statistical analyses will be performed using R version 4.0.3.