Internal Validation of Type 2 Diabetes and Cardiovascular Events Incidences in the Core Obesity and Diabetes Model
Author(s)
Simone Parisotto, PhD1, Anamaria V. Olivieri, MSc2, Francesca Fiorentino, PhD3, Guillem Laborda, BSc4, Anke van Engen, MSc5.
1IQVIA, Milan, Italy, 2IQVIA, Basel, Switzerland, 3IQVIA, Milano, Italy, 4IQVIA, Barcelona, Spain, 5IQVIA, Amsterdam, Netherlands.
1IQVIA, Milan, Italy, 2IQVIA, Basel, Switzerland, 3IQVIA, Milano, Italy, 4IQVIA, Barcelona, Spain, 5IQVIA, Amsterdam, Netherlands.
OBJECTIVES: Internally validate the Core Obesity and Diabetes Model (CODM), a new model integrating obesity into the IQVIA Core Diabetes Model (CDM).
METHODS: The CODM incorporates several risk prediction models to estimate type 2 diabetes (T2D) and cardiovascular (CV) events incidences: QDiabetes2018 for T2D; QRisk3, UKPDS82, UKPDS90 and SMART2 for CV outcomes. CV endpoints included first and recurrent myocardial infarction and stroke, as well as first unstable angina, and transient ischaemic attack. Simulations were conducted for different patient profiles (male [M]/female [F], smoker/non-smoker) with normal glucose tolerance (NGT, HbA1c 5%), prediabetes (HbA1c 6%), and T2D (HbA1c 7%), all free from other complications at baseline. For each profile, starting age was 46 years and BMI 38 kg/m². Clinical parameters progressed naturally over time. Incidences were compared with source studies using the slope of the ordinary least-squares linear regression line (OLS-LRL). Competing risks included age- and gender-specific BMI-adjusted mortality, and event-related deaths.
RESULTS: The 10-years cumulative incidences of T2D for non-smoker patients with obesity (PwO) and NGT were M: 4.33%; F: 2.13% (CODM), M: 4.68%; F: 2.30% (QDiabetes), those for PwO and prediabetes were M: 31.47%; F: 21.60% (CODM), M: 32.51%; F: 22.35% (QDiabetes). The 10-year cumulative incidences of CV events for non-smokers in PwO and NGT were M: 4.96%; F: 2.70% (CODM), M: 4.71%; F: 2.70% (QRisk3); in PwO and prediabetes M: 5.71%; F: 3.18% (CODM), M: 5.50%; F: 3.14% (QRisk3); in PwO and T2D, M: 10.25%; F: 7.12% (CODM), M: 10.28%; F: 7.14% (QRisk3). Predictions with the CODM showed a high degree of linear correlation with source studies (R2 = 0.999) and marginal degree of overestimation (OLS-LRL = 1.029).
CONCLUSIONS: The CODM accurately predicts T2D and CV events against studies used to develop the model, demonstrating high internal validity.
METHODS: The CODM incorporates several risk prediction models to estimate type 2 diabetes (T2D) and cardiovascular (CV) events incidences: QDiabetes2018 for T2D; QRisk3, UKPDS82, UKPDS90 and SMART2 for CV outcomes. CV endpoints included first and recurrent myocardial infarction and stroke, as well as first unstable angina, and transient ischaemic attack. Simulations were conducted for different patient profiles (male [M]/female [F], smoker/non-smoker) with normal glucose tolerance (NGT, HbA1c 5%), prediabetes (HbA1c 6%), and T2D (HbA1c 7%), all free from other complications at baseline. For each profile, starting age was 46 years and BMI 38 kg/m². Clinical parameters progressed naturally over time. Incidences were compared with source studies using the slope of the ordinary least-squares linear regression line (OLS-LRL). Competing risks included age- and gender-specific BMI-adjusted mortality, and event-related deaths.
RESULTS: The 10-years cumulative incidences of T2D for non-smoker patients with obesity (PwO) and NGT were M: 4.33%; F: 2.13% (CODM), M: 4.68%; F: 2.30% (QDiabetes), those for PwO and prediabetes were M: 31.47%; F: 21.60% (CODM), M: 32.51%; F: 22.35% (QDiabetes). The 10-year cumulative incidences of CV events for non-smokers in PwO and NGT were M: 4.96%; F: 2.70% (CODM), M: 4.71%; F: 2.70% (QRisk3); in PwO and prediabetes M: 5.71%; F: 3.18% (CODM), M: 5.50%; F: 3.14% (QRisk3); in PwO and T2D, M: 10.25%; F: 7.12% (CODM), M: 10.28%; F: 7.14% (QRisk3). Predictions with the CODM showed a high degree of linear correlation with source studies (R2 = 0.999) and marginal degree of overestimation (OLS-LRL = 1.029).
CONCLUSIONS: The CODM accurately predicts T2D and CV events against studies used to develop the model, demonstrating high internal validity.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
SA57
Topic
Economic Evaluation, Health Policy & Regulatory, Study Approaches
Topic Subcategory
Decision Modeling & Simulation
Disease
Cardiovascular Disorders (including MI, Stroke, Circulatory), Diabetes/Endocrine/Metabolic Disorders (including obesity)