Menú Cerrar

Development and validation of the risk engine for an Australian Health Economics Model of Osteoporosis

L Si 1 2 3J A Eisman 4 5 6T Winzenberg 7K M Sanders 8 9J R Center 4 6T V Nguyen 4 6 10T Tran 4A J Palmer 11 12

Osteoporos Int. 2021 Apr 15.doi: 10.1007/s00198-021-05955-x. 

Abstract

The Australian Health Economics Model of Osteoporosis (AusHEMO) has shown good face, internal and cross validities, and can be used to assist healthcare decision-making in Australia.

Purpose: This study aimed to document and validate the risk engine of the Australian Health Economics Model of Osteoporosis (AusHEMO).

Methods: AusHEMO is a state-transition microsimulation model. The fracture risks were simulated using fracture incidence rates from the Dubbo Osteoporosis Epidemiology Study. The AusHEMO was validated regarding its face, internal and cross validities. Goodness-of-fit analysis was conducted and Lin’s coefficient of agreement and mean absolute difference with 95% limits of agreement were reported.

Results: The development of AusHEMO followed general and osteoporosis-specific health economics guidelines. AusHEMO showed good face validity regarding the model’s structure, evidence, problem formulation and results. In addition, the model has been proven good internal and cross validities in goodness-of-fit test. Lin’s coefficient was 0.99, 1 and 0.94 for validation against the fracture incidence rates, Australian life expectancies and residual lifetime fracture risks, respectively.

Conclusions: In summary, the development of the risk engine of AusHEMO followed the best practice for osteoporosis disease modelling and the model has been shown to have good face, internal and cross validities. The AusHEMO can be confidently used to predict long-term fracture-related outcomes and health economic evaluations when costs data are included. Health policy-makers in Australia can use the AusHEMO to select which osteoporosis interventions such as medications and public health interventions represent good value for money.