JMIR Publications

ABSTRACT

Background:

Preventable chronic conditions remain a leading cause of morbidity and mortality globally. Despite the importance of prevention, its delivery in primary care is constrained by competing clinical demands, limited resources and fragmented, reactive models of care. Risk assessment also relies on multiple disease-specific risk calculators often focused on short-term risk, limiting opportunities for risk identification and integrated prevention. Consumer-facing digital health platforms may help address these challenges by integrating behavioural and biomedical risk factors to support proactive risk identification, personalised prevention and consumer engagement beyond clinical consultations.

Objective:

This paper describes the development of THRIVE, a consumer-facing digital preventive health platform designed to support proactive, integrated preventive care, and reports preliminary construct benchmarking of its longer-term risk classifications against established clinical risk prediction tools.

Methods:

THRIVE was developed using an iterative user-centred design process informed by clinicians and consumers. The platform integrates 130 evidence-based behavioural, biomedical and demographic determinants to estimate long-term risk across 23 chronic conditions. Construct benchmarking was conducted using purposively constructed synthetic profiles (n=95 per condition domain), incorporating comparable demographic characteristics and condition specific clinical risk factors for cardiovascular disease, type 2 diabetes and obstructive sleep apnoea. THRIVE risk classifications were compared with established tools, the Australian Type 2 Diabetes Risk Assessment (AUSDRISK), Australian Cardiovascular Disease Risk Assessment (AusCVDRisk) and STOP-Bang, using quadratic weighted Cohen’s κ and Spearman correlation.

Results:

THRIVE demonstrated moderate to almost perfect agreement with established risk prediction tools across all domains, with quadratic weighted κ ranging from 0.51 (95% CI 0.39-0.64) for obstructive sleep apnoea to 0.86 (95% CI 0.78-0.94) for type 2 diabetes. Ordinal associations were moderate to strong across domains. Patterns of discordance were consistent with differences in risk horizon, predictor composition, and the underlying constructs assessed by THRIVE and the comparator instruments.

Conclusions:

These findings provide preliminary construct benchmarking evidence for THRIVE’s dynamic longer-term risk assessment framework and demonstrate concordance with established clinical risk prediction tools across multiple chronic condition domains. As a consumer-facing platform, THRIVE has the potential to complement primary care by supporting proactive risk identification and engagement with prevention. Further research using real-world patient data is required to evaluate predictive validity, feasibility and clinical utility in primary care. Clinical Trial: The ePREVENT-360 study is registered with the Australian New Zealand Clinical Trials registry 12624001174572.