Aim: Several recently developed frameworks aim to assess the value of cancer treatments, but the most appropriate metrics remain uncertain. Methods: We use data from the Patient Access to Cancer care Excellence Continuous Innovation Indicators to examine the relationship between hazard ratios (HRs) from clinical trials and dynamic therapeutic value accumulating over time. Results: Our analysis shows that HRs from initial clinical trials poorly predict the eventual therapeutic value of cancer treatments. Conclusion: Relying strongly on HRs from registration trials to predict the long-term success of treatments leaves a lot of the variance unexplained. The Continuous Innovation Indicators offer a complementing, dynamic method to track the therapeutic value of cancer treatments and continuously update value assessments as additional evidence accumulates. Concerns about increased healthcare costs and the ability to pay for meaningful health improvements have fueled discussions about value in healthcare. Various organizations have responded by developing value frameworks to assess the relative benefits of therapeutic interventions. These frameworks typically assess new treatments at the time of launch based on data from initial registration trials and other available information. Some of these frameworks rely mainly on hazard ratios (HRs) of primary end points from registration trials, because these measures often provide the only available estimate of efficacy. However, the extent to which HRs from initial trials predict future value of new therapies is unclear. Moreover, while current value frameworks provide a good initial basis for conceptualizing value, they are not designed to consider value across the lifecycle of a treatment, and they inadequately account for patient preferences.Several examples of published case studies attempt to quantify value dynamically. Among these is the application of Garrison and Veenstra's dynamic lifecycle model to trastuzumab to calculate annual incremental cost-effectiveness ratios (ICERs) for two indications: metastatic breast cancer and adjuvant treatment for early breast cancer [1]. That model demonstrated a decrease over time in ICERs for the combined indications. Similarly, Lu et al. conducted an analysis of the dynamic cost-effectiveness of docetaxel and paclitaxel as new indications arose and patents expired [2]. They concluded, "One direction for future research might be to create a predictive tool that estimates the predicted long-term dynamic ICER for a specific drug at product launch. … All these sources of information can be used together to predict the probability of different events over a drug's life cycle."
KEYWORDShazard ratio • health economics • value assessment