Drug Evaluation Committee Statistical Methods for Estimating Treatment Effects Using Non-randomized Controlled Trial Data
February 2020
In April 2019, a cost-effectiveness evaluation system was institutionalized to complement the existing drug pricing system.Under this system, companies conduct their own analyses following pre-analysis consultations regarding analytical methods and other matters. After a public analysis team verifies (reviews) the scientific validity of the companies’ analyses, a specialized cost-effectiveness evaluation organization performs a comprehensive evaluation. Based on the incremental cost-effectiveness ratio, adjustments are then made to the utility-based surcharge portion of the drug price (and, in the case of the cost-based pricing method, to operating income as well).Under this system, it is necessary to estimate the therapeutic efficacy of a new drug treatment relative to a comparator treatment during the cost-effectiveness analysis process.When estimating treatment effectiveness, if randomized controlled trial data are unavailable, it may be useful to utilize data from non-randomized comparative trials, including observational studies. However, since various biases must be taken into account when estimating treatment effectiveness in non-randomized comparative trials, appropriate analytical methods must be used depending on the trial design.In FY2019 DS Subcommittee Ongoing Project T5, based on Technical Support Document 17: The use of observational data to inform estimates of treatment effectiveness in technology appraisal: methods for comparative individual patient data (2015), the DS Subcommittee prepared a report providing an explanation of the analytical methods used in the analysis of data from non-randomized controlled trials (such as propensity score methods, instrumental variable methods, and regression discontinuity designs).This report focuses on analytical methods that can be applied using standard statistical software, and we hope it will serve as a useful reference for those who will be conducting cost-effectiveness analyses in the future. Please note that any references to or citations of TSD 17 in this report represent the interpretations of this task force, and NICE bears no responsibility whatsoever.Furthermore, this report does not cover all aspects of TSD 17; rather, it presents selected excerpts. In presenting these excerpts, the Task Force has added its own interpretations and supplementary explanations. These are the Task Force’s own conclusions and do not represent the views of NICE.