Including Decision Uncertainty About the Clinically Minimal Important Difference: Shifting From Risk Aversion to Risk Awareness in Drug Reimbursement Processes
Author(s)
JL (Hans) Severens, MSc, PhD1, Maarten Jacobus Postma, PhD2, Mark Sculpher, PhD3.
1Owner/advisor, Severens HTA Consultancy, Venray, Netherlands, 2University of Groningen, Groningen, Netherlands, 3University of York, York, United Kingdom.
1Owner/advisor, Severens HTA Consultancy, Venray, Netherlands, 2University of Groningen, Groningen, Netherlands, 3University of York, York, United Kingdom.
OBJECTIVES: Evidence regarding the effectiveness of novel, more expensive technologies is essential for decisions on implementation and/or reimbursement. Incremental clinical effectiveness is often benchmarked against the minimal important difference (MID). For instance regarding oncology medicines, for ZIN, the Dutch Healthcare Institute, the hazard ratio (HR) for overall survival of a new treatment versus the current standard is a key metric. This study explores the value of decision analysis and risk awareness in drug reimbursement decision-making, using the MID for HR as example.
METHODS: We compared alternative approaches to addressing the statistical uncertainty surrounding the MID: A) The current approach, ignoring the probability that a reimbursement decision based on the MID may be incorrect; B) Explicitly incorporating the probability that a decision based on the HR's statistical uncertainty relative to the MID may be incorrect. Approach A uses a stepwise method: Step 1: Determine whether there is certainty about the HR, defined as the 95% confidence interval (CI) of the HR not including the MID. Step 2: If certainty is present and the HR is unfavorable, the reimbursement decision is negative, and cost-effectiveness (e.g., cost per QALY) is not assessed. Approach B accounts for the probabilities of both false-negative and false-positive reimbursement decisions. False negatives result in QALYs forgone, while false positives lead to increased healthcare costs without patient benefit.
RESULTS: A decision-analytic model was used to compare both approaches. The example demonstrates that Approach A (stepwise, resulting in a negative net monetary benefit (NMB)) leads to suboptimal decisions when compared to Approach B (integrated, resulting in a positive NMB).
CONCLUSIONS: This study shows that a stepwise approach to assessing clinical and cost-effectiveness may result in suboptimal reimbursement decisions. An integrated decision-analytic approach leads to more informed and balanced outcomes. Decision-makers should shift from risk aversion to risk awareness.
METHODS: We compared alternative approaches to addressing the statistical uncertainty surrounding the MID: A) The current approach, ignoring the probability that a reimbursement decision based on the MID may be incorrect; B) Explicitly incorporating the probability that a decision based on the HR's statistical uncertainty relative to the MID may be incorrect. Approach A uses a stepwise method: Step 1: Determine whether there is certainty about the HR, defined as the 95% confidence interval (CI) of the HR not including the MID. Step 2: If certainty is present and the HR is unfavorable, the reimbursement decision is negative, and cost-effectiveness (e.g., cost per QALY) is not assessed. Approach B accounts for the probabilities of both false-negative and false-positive reimbursement decisions. False negatives result in QALYs forgone, while false positives lead to increased healthcare costs without patient benefit.
RESULTS: A decision-analytic model was used to compare both approaches. The example demonstrates that Approach A (stepwise, resulting in a negative net monetary benefit (NMB)) leads to suboptimal decisions when compared to Approach B (integrated, resulting in a positive NMB).
CONCLUSIONS: This study shows that a stepwise approach to assessing clinical and cost-effectiveness may result in suboptimal reimbursement decisions. An integrated decision-analytic approach leads to more informed and balanced outcomes. Decision-makers should shift from risk aversion to risk awareness.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
HPR122
Topic
Health Policy & Regulatory, Health Technology Assessment
Topic Subcategory
Reimbursement & Access Policy
Disease
Oncology