Does the Diabetes Alliance Program Improve Longer-term Health Outcomes of Patients with Type 2 Diabetes Attending Primary Care Practices? A Target Trial Protocol for Emulating a Cluster Trial Using Linked General practice and Tertiary Health Data

BackgroundEffective management of type 2 diabetes is an increasing challenge for health services globally. Integrated care, targeted at increasing capacity in primary care for earlier intervention in type 2 diabetes, may reduce adverse health outcomes, with benefits accrued over time. This protocol aims to describe a causal approach for the evaluation of a specialist-led integrated model of care delivered in Australian general practices, using linked administrative health data. MethodsThis protocol outlines an observational cohort study using the Lumos data asset (New South Wales Health), linking general practice and hospital data. It follows the target trial framework, emulating a parallel cluster-randomised trial, and applies the estimands framework proposed by the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use. It describes the data structure and statistical analysis plan required for causal inference. The primary causal estimand is the effect of the Diabetes Alliance Program Plus, an integrated care intervention, on all-cause hospitalisation rates for adults actively attending general practice with current or newly diagnosed type 2 diabetes over a 5-year follow-up period. DiscussionThis protocol applies the target trial framework to emulate a parallel cluster-randomized trial of the Diabetes Alliance Program Plus using linked administrative data. This approach addresses confounding and selection bias inherent in observational evaluations of practice-level interventions and will generate robust real-world evidence to inform policy regarding type 2 diabetes management in primary care settings. Trial registrationACTRN12622001438741; 10th November 2022, retrospectively registered: https://www.anzctr.org.au/ACTRN12622001438741.aspx.