New interventions to decrease the incidence and severity of cardiovascular disease is being developed continuously. To the extent interventions are effective they are not necessarily cost-effective. The aim of this project was, on the basis of recent data, to develop a general cardiovascular model to simulate cohorts from healthy (asymptomatic) through different types of cardiovascular morbidity to death.

Methods

We developed a Markov model that encompasses 6 adverse events (myocardial infarction (STEMI and non-STEMI), stroke, heart failure, angina and death) and 6 states (healthy, asymptomatic CVD, heart failure, stroke sequelae (moderate and severe) and death). The model was developed in Treeage.

The incidence of primary cardiovascular events was based on national registry data, while the the risk of subsequent events was based on various randomized controlled trials. The model allows adjustment of incidence based on risk factors such as blood pressure and cholesterol level.

Costs are attached to both adverse events and health states. Utilisation of health care was based on guidelines and expert judgement while unit costs were based on market prices, DRG charges, and various fee scedules. Quality of life weights were taken from the CEA registry.

The main model outputs are (quality adjusted) survival time and life time costs. We used various distributions (gamma, beta, lognormal and normal) to model parameter uncertainty to allow probabilistic sensitivity analyses.

Results

We run the model with and without a hypothetical intervention for 40-year-old men with average risk levels. The intervention reduced the risk of all adverse outcomes by 20%. The life expectancy was 39.01 years without and 39.99 years with intervention, while the life-time treatment costs were $67 000 and $75 000 respectively (all numbers undiscounted).

This intervention results in avoided cardiovascular events, and hence some treatment costs are averted in addition to the gain of 0.98 life-years. An intervention with these effects could cost up to €200 per year and still be cost-saving, and up to €1 700 and still be cost-effective.

Conclusions

The NorCaD model is flexible and allows economic evaluation of several types of CVD interventions, although the model was primarily developed to evaluate primary prevention strategies. While several earlier models were based on Framingham risk equations, this model is based on recent local data.