The rising cost of oncology drugs has been drawing increasing questions from payers, providers and regulators about the relative value of individual therapies and how to quantify that value in light of pricing trends. Launch prices of oncology drugs in the US have risen dramatically, from roughly $100 a month in the 1970s, to more than $10,000 a month today, and some argue that these prices are unjustified.

In response, several healthcare industry organizations have been developing value frameworks in attempt to assign a quantitative value to these drugs and to determine whether that value corresponds to their market cost. One of the most prominent tools to emerge from this trend is the American Society of Clinical Oncology’s (ASCO) Value Framework.

At the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) annual meeting in May, our team presented a poster detailing research we conducted into the sensitivity of clinical endpoints in the ASCO Value Framework and its implications for clinical trial design. Our ultimate aim with this study was to identify a combination of endpoints that optimizes the value outcome in the context of clinical development.

A holistic view of value

ASCO’s Value Framework is a simplified tool designed to guide patients and oncologists in joint decision-making about treatment options by comparing new oncology treatments to the current standard of care (SoC) or placebo. The framework computes relative value of an oncology therapy based on the weighted combination of six underlying clinical endpoint scores. Points are earned by comparing certain outcome metrics to the SoC or placebo to generate a single net health benefit (NHB). The endpoints include:

  • Median Overall Survival (mOS) and confidence interval
  • Median Progression-free Survival (mPFS) and confidence interval
  • Response Rate (RR) and confidence interval
  • Toxicity (T), defined as count of grade 3-5 toxicities
  • Palliation (P)
  • Treatment-free Interval (TFI)

To assess the range and probability distribution of the NHB scores, our team conducted Monte Carlo simulations on three marketed oncology products in distinct therapeutic areas.  For each product, a Monte Carlo simulation of 10,000 iterations produced a probability distribution of NHB scores for twelve possible endpoint combinations composed of: one efficacy endpoint (mOS, mPFS or RR),  toxicity (included in all scores), and none, one or two of the bonus endpoints (palliation or TFI). For each scenario, it was noted which endpoints yielded the highest NHB. Then clinical trial cost multipliers were developed based on assumption for each of the twelve endpoint combinations to illustrate cost associated with obtaining NHB score.

Lessons for industry

The results of the study uncovered several noteworthy insights:

  • While overall survival is theoretically the lever most capable of driving higher NHB scores, we found other clinical benefit endpoints can produce a greater NHB under certain scenarios.
  • While progression-free-survival maximized the NHB score for two of the three assessed products in a majority of the scenarios, response rate maximized the score in the third product.
  • Palliation and treatment-free intervals generally yielded a marginal five point increase in median NHB.
  • While an improvement in patient quality of life, or treatment-free interval can drive higher NHB scores, the increment is tempered compared to efficacy.
  • Most importantly, median NHB for the optimal endpoint was typically 15-20 points higher than the next best alternative.

These results offer useful lessons for industry on how to think about the value of key endpoints in the drug development process, and how to choose the endpoints that will maximize the value of a treatment while managing research costs. Notwithstanding the limitations of an analysis based on a sample of three oncology drugs, our exploratory analysis appears to corroborate the clinical approach of pursuing median progression free survival or response rate in pivotal trials at relatively lower trial costs, and subsequently obtaining median overall survival data post-market approval through real-world outcomes studies.

We submit that that this evaluation be viewed as a reproducible method for assessing clinical endpoint strategy to optimize NHB scores in the context of clinical development, and can be adapted to other value frameworks, such as those from ICER, DrugAbacus, among others.

Though it is also important to note that these frameworks are still evolving, and stakeholders should keep a close eye on their development in order to align their development goals with industry standards of value. Unless the industry harmonizes behind a single value framework it’s also advisable to keep track of all relevant frameworks and how various endpoints might impact the commercial value of new products. Tracking this trend will help biopharma companies make better decisions about which clinical endpoints to pursue, and give them a stronger position to make a case for the value of their treatments.


This post was co-authored by John Doyle.

Topics in this blog post: Biopharma, CER, Healthcare Cost, Oncology, Value