Cost-effectiveness of a telehealth intervention to reduce CVD risk: Lifetime simulation modelling using QRISK2

The problem

Cardiovascular disease (CVD) is an increasingly prevalent long-term condition associated with healthcare burden, morbidity, and mortality. A challenge in estimating the cost-effectiveness of technologies intended to reduce CVD risk is that RCTs may have follow-up periods that are shorter than the period over which the intervention may affect patient health. These issues were confronted in the economic analysis of the Healthlines CVD risk RCT. The trial randomised 641 people to receive either usual care or a theory-based telehealth intervention. The primary clinical outcome of the CVD trial was 10-year CVD risk, estimated using the QRISK2 risk prediction algorithm. A 10-year period is substantially longer than the follow-up period of the trial, but may be shorter than the actual duration of the effect of the intervention on patient health. The approach : Trial participants were followed-up for 12 months, and quality of life and resource use data was collected from questionnaires and primary care records. To estimate the expected net benefits of the intervention beyond the end of trial follow-up, a state transition cohort simulation Markov model was developed to assess the consequences of the intervention from an NHS perspective. Patients were modelled as transitioning through discrete health states defined by specific CVD events (such as stroke or myocardial infarction). The probability of moving between different states was determined by the interaction of trial participant characteristics and transition matrices. Information on differences between trial arms in QRISK2 scores and data from epidemiological and other sources was used to parameterise the model. Lifetime cost-effectiveness was measured by incremental Quality Adjusted Life years (QALYs) and incremental cost. Probabilistic sensitivity analysis was used to characterise the uncertainty around estimates of net benefit. Findings : The simulation model showed that the probability of the intervention being cost-effective was sensitive to (a) the duration for which the effects of the Healthlines intervention were modelled as persisting and (b) the time period over which effects were modelled. For longer time periods, the number of CVD events that were avoided by patients in the intervention arm tended to make the Healthlines telehealth intervention more likely to be cost-effective compared to usual care for the trial patient group. Consequences NICE guidance requires technology assessments to adopt a time horizon for analysis that is sufficient to reflect costs and effects that may differ between intervention and control groups. The evidence from the economic evaluation of the Healthlines trial shows how the output of a commonly used algorithm (QRISK2) can be used to inform extrapolations of costs and outcomes beyond the follow-up period of an RCT.

Credits

• Padraig Dixon, University of Sheffield, Sheffield, UK
• Sandra Hollinghurst, University of Sheffield, Sheffield, UK
• Roberta Ara
• Chris Salisbury, University of Sheffield, Sheffield, UK