# Effectiveness, Reach, Uptake, and Feasibility of Digital Health Interventions for Culturally and Linguistically Diverse Populations Living With Prediabetes Across the Lifespan: Systematic Review and Meta-Analysis

## TL;DR

Digital health interventions may help improve blood sugar control in culturally and linguistically diverse people with prediabetes, but results are mixed and require further study.

## Contribution

This is the first systematic review and meta-analysis to evaluate digital health interventions for prediabetes in culturally and linguistically diverse populations.

## Key findings

- Digital health interventions reduced hemoglobin A1c levels in CaLD populations with prediabetes.
- High uptake and engagement rates were observed among CaLD participants using digital health interventions.
- Heterogeneity in study outcomes suggests variability in intervention effectiveness based on factors like comorbidity and duration.

## Abstract

Culturally and linguistically diverse (CaLD) populations are at a higher risk of developing prediabetes; however, the effectiveness and implementation of digital health interventions for prediabetes management in this population are not well understood.

This review aims to evaluate the effectiveness and implementation of digital health interventions (DHIs) versus usual care for glycemic control in CaLD populations living with prediabetes.

This review aimed to include people of any age living with prediabetes who are from a CaLD background. Experimental and quasi-experimental studies that compare digital health interventions to usual care, waitlist, or active control were eligible. The primary outcome was glycemic control as measured by hemoglobin A1c. A comprehensive search was conducted in CINAHL, Cochrane Library, Embase, MEDLINE, 3 trial registers, and gray literature databases, along with reference lists for additional studies. Studies published in English and published since the inception of each database were included. Statistical analyses included meta-analysis, sensitivity analyses, subgroup analyses, meta-regression, and publication bias assessments. The methodological quality was assessed using the JBI critical appraisal tools, and the quality of evidence was evaluated using Grading of Recommendations, Assessment, Development, and Evaluation to create summary of findings tables. Random-effects models with restricted maximum likelihood estimation were employed.

A total of 14 studies involving 5714 adult participants were included. The meta-analysis showed that DHIs were associated with a reduction in hemoglobin A1c (P<.001), though evidence certainty was low (mean difference=−0.14, 95% CI −0.24 to ‐0.05). Effects on fasting blood glucose and body weight remain uncertain. Implementation outcomes demonstrated high uptake (>78.8%), engagement (>80%), and intention rates (89.1%) among CaLD populations with prediabetes. Significant heterogeneity was observed in both randomized controlled trials and pre-post studies. Subgroup analyses revealed significant effects at the 6-month follow-up point only for interventions (P<.001). Meta-regression identified comorbidity status as the only significant contributor to heterogeneity (P=.02). Sensitivity analyses demonstrated robust significant effects (P<.001). Publication bias assessment showed mixed results (Begg P=.23, Egger P=.02), but trim-and-fill analysis confirmed the robustness of the findings with no missing studies. Despite these positive findings, substantial heterogeneity across most outcomes and low-to-very low certainty evidence limit the reliability of these results, warranting cautious interpretation.

DHIs demonstrate potential for improving glycemic control in CaLD populations living with prediabetes. The observed heterogeneity could be attributed to intervention duration, control type, and participants’ comorbidity status. While the findings related to implementation were encouraging, the certainty of the evidence and substantial heterogeneity suggest that DHIs should be used as adjunctive tools with health care provider involvement rather than stand-alone solutions due to low certainty evidence and substantial heterogeneity. Further rigorous research considering contextual, individual, and cultural factors is needed.