JMIR Publications

ABSTRACT

Background:

Noncommunicable diseases (NCDs) are the leading cause of death globally, and much of this burden is driven by modifiable lifestyle factors, highlighting the need for effective prevention strategies. Composite digital health scores could support population-level risk stratification, but their measurement properties are rarely evaluated against routinely collected health records.

Objective:

This study assessed the validity and reliability of a proposed “Health Index” among 5,997 users, delivered through BruHealth, the national mobile health app of Brunei Darussalam.

Methods:

In this cross-sectional study, we analyzed Health Index data from BruHealth users who were 18 years old and above. Individuals who are pregnant are excluded. We assessed the concurrent-criterion validity by correlating the Health Index score with the number of outpatient visits during the period of one year preceding completion of the Health Index. For construct validity, we compared Health Index scores among three groups of users with different health statuses (clinical conditions). For the reliability, we analyzed test-retest intraclass correlation (ICC Type 2) from their repeated submission of Health Index.

Results:

The Health Index score was negatively correlated with number of visits related to NCDs during the recent one-year period (Spearman ρ=−0.23, P<.001). Mean number of visits was also significantly different across Health Index risk groups (healthy n=1,871, at risk n=4,102, poor health n=24; P<.001). The mean Health Index scores were strongly significantly different among the three groups by health status or clinical conditions (P<.001). For the repeated submission of Health Index, the test-retest ICC was 0.636 (95% CI: 0.59, 0.68) which suggest moderate reliability.

Conclusions:

Health Index, which has been integrated into the Brunei national mobile health application, showed moderate test-retest reliability and evidence of concurrent-criterion validity in measuring overall health status and stratifying people into different risk groups.