ADDRESSING DECISIONS ABOUT CLINICAL GUIDANCE AND IMPLEMENTATION STRATEGIES – A SEQUENTIAL OR AN INTEGRAL APPROACH? AN APPLICATION RELATING TO METASTATIC HORMONE-REFRACTORY PROSTATE CANCER

Background: In allocating resources to improve patient care, there are two decisions which must be addressed. The first is whether it is cost-effective to adopt a health care technology or clinical practice guideline. The use of such guidance by clinicians and/or patients, however, does not necessarily follow. Therefore, a second decision has to be made concerning whether it is worthwhile adopting an active implementation strategy (e.g., education and audit & feedback) for promoting guidance usage. Given that these two decisions are related, resource allocation can be based on either sequential (1-by-1) or integral analysis.

Purpose: To identify the preferred approach (sequential or integral) to the analysis for addressing decisions about the adoption and implementation of clinical guidance. Application: The application relates to the evaluation and improvement of care for patients with metastatic hormone-refractory prostate cancer (mHRPC) in the UK.

Method: An integrated Bayesian approach to decision modeling and evidence synthesis is adopted using MCMC simulation in WinBUGs. Evidence on the costs and quality-adjusted life-years (QALYs) of all plausible treatment regimens is combined with estimates of nationwide treatment usage and population size. Implementation costs and effects are assumed to vary between treatment options. Both sequential and integral analysis of decision making about resource allocation are performed.

Result: (preliminary) For threshold values between £25,000 and £32,000 per QALY, decisions about how best to improve care for mHRPC patients differ between the alternative analytic approaches. On the basis of sequential analysis, Mitoxantrone+Prednisone should be adopted and the active promotion of its clinical use is not deemed cost-effective. When implementation costs are considered an integral part of care improvement, however, actively implementing Docetaxel+Prednisone (3 weekly) appears to be the cost-effective option, yielding an expected additional 0.05 QALY per patient. By combining uncertain economic evidence on both treatment regimens (including baseline usage) and implementation strategies, allowance is made for all uncertainty associated with resource allocation in mHRPC.

Conclusion: For the analysis of the related decisions about the adoption of clinical guidance and that of strategies for implementing the guidance, an integral approach is preferred over a sequential one. As the application in mHRPC demonstrates, integral analysis provides better options for improving patient management, more comprehensive insight in decision uncertainty about guidance implementation and, consequently, an efficient allocation of resources.