OBJECTIVES: Digital healthcare presents a growing opportunity to improve diabetes management. However, the impact of digital interventions on patient outcomes and healthcare costs is poorly understood, partly due to their heterogeneity. We present a framework for assessing the cost impact of digital interventions in people with type 2 diabetes (PWD2).

METHODS: Diabetes-relevant clinical measures impacted by digital interventions were identified by focused literature review. These were subsequently used as ‘input variables’ in a machine learning regression model based on a large RW cohort of ~300,000 adult PWD2 (Optum Market Clarity). For each PWD2, a central ‘index’ was selected and ≥100 covariates (during the year pre-index) and medical costs (in the year post-index) were determined to estimate the impact of clinical measure variation on total annual costs. Ranges for the impact of digital interventions on clinical measures, obtained from literature review, were then applied to the regression model to estimate the potential economic impact of the interventions.

RESULTS: We identified 30 studies assessing digital interventions and their effect on clinical measures. Over 100 clinical measures were identified; these were refined to five clinical measures in the final model. For a unit change in each clinical measure, the effect on total cost was estimated by the following coefficients: glycated hemoglobin, 5.02%; number of hypoglycemic events, 1.01%; blood glucose, 0.78%; weight, 0.27%; diastolic blood pressure, 0.04%. Clinical measure impact ranges for each digital intervention, identified from literature, could then be applied to the regression model to estimate ranges of cost impact.

CONCLUSIONS: Our model framework directly links clinical impacts of digital interventions to RW healthcare costs and suggested that selected interventions can drive meaningful change in patient costs. The model framework is modifiable and can be further developed for a more clinically robust and accurate cost assessment of digital intervention in diabetes patients.