Evaluation of patient-specific mobile phone–generated reminders for hypertension care in primary care settings in Western Kenya

Abstract

Background: Kenya has a high prevalence of hypertension with rates of 20 – 50% in urban areas. Hypertension care is limited to few highly skilled healthcare facilities with physicians and better diagnostic and clinical care. The Academic Model Providing Access to Healthcare (AMPATH) – Chronic Disease program has instituted task-shifting of hypertension care to dispensaries and health centers with the use of smartphones for electronic data capture and clear hypertension care algorithms and clinical decision support systems (CDSS) rules implemented to ensure best quality of care. Objective: To evaluate the impact, adherence and clinician barriers to use of patient-specific mobile phone–generated care suggestions for hypertension care in primary care settings in western Kenya. Methods: This was a prospective comparative study that was conducted in Ministry of Health dispensaries and health centers offering hypertension care in two counties in western Kenya: Uasin Gishu (Turbo sub-county) and Nandi (Chesumei sub-county). Study participants were patients with hypertension. Participants were clustered into the intervention or control group Participants in the intervention group were seen by clinicians who had an mHealth application that generated and availed patient-specific care suggestions during the clinical encounter. In the control arm, care suggestions were triggered in the application but not availed to the providers. A log of the triggered care suggestions in both the control and intervention groups was maintained. The unit of analysis for all analyses was each generated hypertension care suggestion in both the control and intervention groups. Data analysis was done using STATA version 13 SE (College Station, 77845 Texas USA). Comparison of the median for continuous variables was done using two-sample Wilcoxon rank-sum test. Pearson’s Chi Square test was used to assess the association between the independent categorical variables and the study arms. Results: In the study, 378 patients with hypertension had care encounters in which care suggestions were generated (217 in intervention group and 161 in control group). Participants were similar by age, gender and blood pressure (P>0.05). Proportion of participants with SBP ≥ 140 mm Hg | DBP ≥ 90 mm Hg were similar in the two groups (67.4% vs. 61.0%, p = 0.198) respectively. A total of 440 of 481 generated care suggestions were fulfilled in the intervention group and 318 of 371 in the control group. There was a higher proportion of adherence to hypertension care suggestions in the intervention group due to the presence of patient specific care suggestions compared to the control group 91.1% vs. 85.7%, however this was not statistically significant with adjusted odds ratio (AOR): 1.78 (95% CI: 0.83, 3.80). Adherence to patient-specific care suggestions had no significant impact on blood pressure control (OR: 2.41, CI: 0.60 – 9.67). Conclusion: Presence of patient-specific care suggestions led to higher adherence to hypertension care guidelines. However, no impact of care suggestion was observed on blood pressure control. Recommendation: Use of patient-specific care suggestions should be considered as an additional strategy for improving adherence to hypertension care guidelines. We recommend further mixed methods studies to elucidate the cause of poor control of blood pressure other than adherence to care guidelines.