Table 6 Evidence-based recommendations for CMD policy modelling
From: Policy models for preventative interventions in cardiometabolic diseases: a systematic review
Area | Key recommendations |
|---|---|
Model selection | State-transition models (e.g., Markov models) are commonly used for CMD progression, but analysts should align the model choice with the policy question, available data, and computational feasibility |
Integration of CMDs | Given the shared risk factors of T2DM and CVD, incorporating them into the same model can improve accuracy and capture event-related risks |
Risk factors | Models can integrate modifiable risk factors (e.g., BMI, cholesterol, lifestyle changes) to ensure more realistic projections |
Data quality | High-quality patient-level and representative epidemiological data should be prioritised. Incorporating clinical biomarkers and capturing heterogeneous effects can improve generalisability |
Economic perspective | If data are available, considering a societal perspective can enhance health-economic modelling beyond direct healthcare costs |
Uncertainty analysis | Specifying uncertainty and conducting appropriate sensitivity analyses is essential for ensuring robust conclusions |
Validation | Reporting validation tests (internal, external, face validity) is recommended to improve model reliability and reproducibility |
Transparency & reporting | Clearly document model rationale, assumptions, and methodologies. Conceptual models should be well-documented to enhance credibility |
Equity & distributional analysis | Ensuring that models assess distributional impacts can support policies that reduce health inequalities |
Reproducibility & open Science | Adhering to best research practices and making policy models open source can improve transparency, accessibility, and reproducibility |