Effectiveness of digital health interventions on blood pressure control, lifestyle behaviours and adherence to medication in patients with hypertension in low-income and middle-income countries: a systematic review and meta-analysis of randomised controlled trials

Abstract

Background: Digital health interventions can be effective for blood pressure (BP) control, but a comparison of the effectiveness and application of these types of interventions has not yet been systematically evaluated in low- and middle-income countries (LMICs). This study aimed to compare the effectiveness of digital health interventions according to the World Health Organisation (WHO) classifications of patients in terms of BP control, lifestyle behaviour changes, and adherence to medication in patients with hypertension in LMICs.

Methods: In this systematic review and meta-analysis, we searched the PubMed, Scopus, Web of Science, Embase, CINAHL, and Cochrane Library databases for randomised controlled trials (RCTs) published in English, comprised of adults (≥18 years old) with hypertension and the intervention consisted of digital health interventions according to WHO's classifications for patients in LMICs between January 1, 2009, and July 17, 2023. We excluded RCTs that considered patients with hypertension comorbidities such as diabetes and hypertension-mediated target organ damage (HMTOD). The references were downloaded into Mendeley Desktop and imported into the Rayyan web tool for deduplication and screening. The risk of bias was assessed using Cochrane Risk of Bias 2. Data extraction was done according to Cochrane's guidelines. The main outcome measures were mean systolic blood pressure (SBP) and BP control which were assessed using the random-effect DerSimonian-Laird and Mantel-Haenszel models. We presented the BP outcomes, lifestyle behaviour changes and medication adherence in forest plots as well as summarized them in tables. This study is registered with PROSPERO, CRD42023424227.

Findings: We identified 9322 articles, of which 22 RCTs from 12 countries (n = 12,892 respondents) were included in the systematic review. The quality of the 22 studies was graded as high risk (n = 7), had some concerns (n = 3) and low risk of bias (n = 12). A total of 19 RCTs (n = 12,418 respondents) were included in the meta-analysis. Overall, digital health intervention had significant reductions in SBP [mean difference (MD) = -4.43 mmHg (95% CI -6.19 to -2.67), I² = 92%] and BP control [odds ratio (OR) = 2.20 (95% CI 1.64-2.94), I² = 78%], respectively, compared with usual care. A subgroup analysis revealed that short message service (SMS) interventions had the greatest statistically significant reduction of SBP [MD = -5.75 mm Hg (95% Cl -7.77 to -3.73), I² = 86%] compared to mobile phone calls [MD = 3.08 mm Hg (-6.16 to 12.32), I² = 87%] or smartphone apps interventions [MD = -4.06 mm Hg (-6.56 to -1.55), I² = 79%], but the difference between groups was not statistically significant (p = 0.14). The meta-analysis showed that the interventions had a significant effect in supporting changes in lifestyle behaviours related to a low salt diet [standardised mean difference (SMD) = 1.25; (95% CI 0.64-1.87), I² = 89%], physical activity [SMD = 1.30; (95% CI 0.23-2.37), I² = 94%] and smoking reduction [risk difference (RR) = 0.03; (95% CI 0.01-0.05), I² = 0%] compared to the control group. In addition, improvement in medication adherence was statistically significant and higher in the intervention group than in the control group [SMD = 1.59; (95% CI 0.51-2.67), I² = 97%].

Interpretation: Our findings suggest that digital health interventions may be effective for BP control, changes in lifestyle behaviours, and improvements in medication adherence in LMICs. However, we observed high heterogeneity between included studies, and only two studies from Africa were included. The combination of digital health interventions with clinical management is crucial to achieving optimal clinical effectiveness in BP control, changes in lifestyle behaviours and improvements in medication adherence.

Funding: None.

Keywords: Adherence to medication; Blood pressure control; Digital health interventions; Lifestyle behaviours; Patients with hypertension.

Fig. 1

PRISMA flow chart of study selection. PRISMA flow diagram specifying the considerations to exclude and include the articles.

Fig. 2

Forest plot of the mean difference in SBP between the intervention and control groups. Forest plot of mean difference in systolic blood pressure (SBP) (expressed as mm Hg) between the digital health intervention and the control groups, and subgroup analysis by mode of delivery of the intervention (Mobile phone call, Short message service (SMS) and Smartphone app). The size of the blue squares indicates the weight of the evidence from each of the studies. Studies with CI (horizontal line) crossing zero (vertical line) are inconclusive. Studies with more participants have narrower CIs. The red diamonds represent the summary effect sizes in each of the subgroups and the green the overall sample, with the width of the diamond indicating the 95% CI. A statistically significant greater reduction in SBP is seen in the intervention group, compared with the control group in the overall sample and with the two modes of delivery (SMS and smartphone app). SMS interventions displayed the greatest reduction, compared with smartphone apps and mobile phone calls, but the differences between the three modes were not significant. The data present high heterogeneity.

Fig. 3

Meta-analysis of dichotomous outcome measurements for BP control. Forest plot of odds ratio in blood pressure (BP) control between the digital health intervention and the control groups. The size of the blue squares indicates the weight of the evidence from each of the studies. Studies with CI (horizontal line) crossing zero (vertical line) are inconclusive. Studies with more participants have narrower CIs. The black diamonds represent the summary effect sizes in the overall sample, with the width of the diamond indicating the 95% CI. A statistically significant greater reduction in BP is seen in the intervention group, compared with the control group. The data present high heterogeneity.

Fig. 4

Meta-analysis of continuous outcome measurements for low-salt diets. Forest plot of standard mean difference for low-salt diets between the digital health intervention and the control groups. The size of the green squares indicates the weight of the evidence from each of the studies. Studies with CI (horizontal line) are inconclusive. The black diamonds represent the summary effect sizes in the overall sample, with the width of the diamond indicating the 95% CI. A statistically significant greater reduction for low-salt diets is seen in the intervention group, compared with the control group. The data present high heterogeneity.

Fig. 5

Meta-analysis of continuous outcome measurements for PA. Forest plot of standard mean difference for physical activity (PA) between the digital health intervention and the control groups. The size of the green squares indicates the weight of the evidence from each of the studies. Studies with CI (horizontal line) crossing zero (vertical line) are inconclusive. The black diamonds represent the summary effect sizes in the overall sample, with the width of the diamond indicating the 95% CI. A statistically significant greater reduction for PA is seen in the intervention group, compared with the control group. The data present high heterogeneity.

Fig. 6

Meta-analysis of dichotomous outcome measurements for smoking reduction. Forest plot of risk difference for smoking reduction between the digital health intervention and the control groups. The size of the blue squares indicates the weight of the evidence from each of the studies. Studies with CI (horizontal line) crossing zero (vertical line) are inconclusive. The black diamonds represent the summary effect sizes in the overall sample, with the width of the diamond indicating the 95% CI. A statistically significant greater reduction in smoking reduction is seen in the intervention group, compared with the control group. The data present no heterogeneity.

Fig. 7

Meta-analysis of continuous outcome measurements for adherence to medication. Forest plot of standard mean difference for continuous outcome measurements for adherence to medication between the digital health intervention and the control groups. The size of the green squares indicates the weight of the evidence from each of the studies. Studies with CI (horizontal line) crossing zero (vertical line) are inconclusive. The black diamonds represent the summary effect sizes in the overall sample, with the width of the diamond indicating the 95% CI. A statistically significant greater reduction in adherence to medication is seen in the intervention group, compared with the control group. The data present high heterogeneity.

Fig. 8

Meta-analysis of dichotomous outcome measurements for adherence to medication. Forest plot of standard mean difference for dichotomous outcome measurements for adherence to medication between the digital health intervention and the control groups. The size of the blue squares indicates the weight of the evidence from each of the studies. Studies with CI (horizontal line) are inconclusive. The black diamonds represent the summary effect sizes in the overall sample, with the width of the diamond indicating the 95% CI. A statistically significant greater reduction in adherence to medication is seen in the intervention group, compared with the control group. The data present moderate heterogeneity.