Health Economics in Clinical Trials

Based on Drummond, Sculpher, Torrance, O'Brien & Stoddart, 2005.
Randomised controlled clinical trials have both strengths and weaknesses for economic evaluation. In general there are two ways to incorporate economic questions into a trial – either by including economic evaluation within an existing efficacy or safety trial, or through the design of a pragmatic trial specifically for the purpose of economic evaluation.

Given that randomised controlled trials are the accepted gold standard for assessing safety and efficacy, it is reasonable to consider collecting economic data within an existing clinical trial. This allows prospective collection of patient level cost and outcomes data, while utilising the existing trial infrastructure and therefore requiring minimal additional resources. However issues such as the comparison of new treatments to placebos, the use of surrogate endpoints and a lack of long term follow up can limit the ability of trials designed primarily for a clinical purpose to answer economic questions.

Alternatively, pragmatic trials can be designed specifically to answer economic evaluation questions. These trials minimise bias through random allocation, but aim to assess the effectiveness or cost-effectiveness of an intervention under ‘real world' conditions, by having fewer restrictions on patient recruitment and follow up. This allows greater generalisability of results. Difficulties in comparing more than two or three treatment options and a lack of long term follow up are still present in pragmatic trials and need to be addressed in the design phase.

Finally, economic modelling can be used when evidence from a range of sources is required for decision making. Data sources such as cohort studies and surveys can be used alongside clinical trials to obtain clinical, cost and quality of life information for use in decision analytic models. Coming soon to the CREST website will be a range of resources, including some providing an introduction to and further readings on health economics in clinical trials.

Conducting an Economic Evaluation: Pro-Forma Documents

Over the years, CREST has developed and provided a number of documents on the why and how of conducting an economic evaluation alongside a cancer clinical trial.  While some documents have provided background information to health economics concepts, others have been more applied in nature, focusing on collecting data for economic evaluation. 

Now, all those documents can be accessed through one portal on the CREST website

The aim of the portal is to allow groups intending to include health economics in a clinical trial to have (a) some initial starting guidance on how to conceptualise the problem, (b) an outline of how to collect the relevant data, and (c) access to example questions for data collection that can be included in study CRFs.

Conceptualising the Problem
The resources available in this section of the portal aim to provide background on the concepts of health economics and the conduct of an economic evaluation alongside a clinical trial.  This not only includes a step-by-step guide to conducting an economic evaluation, but the resources required to conduct such an evaluation.  The documents available include:

a.    Economic evaluations in cancer clinical trials (FactSheet).  This provides general guidance on understanding the purpose of conducting an economic evaluation.  It will be part of a section on “Conceptualising the Problem”.  Information presented to the TROG ASM (2016) on how to form the economic question will also be added to this section.
b.    Step by step guide to economic evaluations in clinical trials (FactSheet).  This provides an overview of the steps involved in undertaking an evaluation.
c.    How much does it cost to include an economic evaluation in a clinical trial (FactSheet).  This is to provide the developers of trials with a practical understanding of what is required when considering the budget for an economic evaluation.

Collecting the Data
Over the years, CREST has provided advice on the choice and development of questions to collect resource use (cost) and outcomes data from clinical trials.  The advice provided has been summarised, in a de-identified manner, into a proforma document (in Excel) that contains example questions on collecting information within clinical trials on the following:
i.    Treatment use and follow-up services (including medical and diagnostic services use)
ii.    Hospital services use
iii.    Treatment of adverse events
iv.    Patient reported outcomes:  quality of life and other PROs
v.    Demographic information

The Excel workbook can be downloaded here.  In general, the questions contained in this workbook can be customised (unless stated otherwise in the workbook); please ask one of the team at CREST if you need further advice.

Alongside that workbook, CREST has also provided information on collecting administrative data (data on patients’ health care use that are collected as part of the routine process of paying for health care services.  The Medicare Australia Data (FactSheet) provides general guidance on the rationale and process for accessing Medicare (MBS and PBS) data.  Accessing data from Medicare requires not only patient consent, but an application to the Department of Human Services.  If you would like assistance in completing that application, or to see an example application form, please contact CREST.

How to Construct QoL Scores for Health Economics Analysis
A critical step in using health related quality of life data within economic evaluations is being able to convert patient-rated health status into society-rated preference values, or utility weights.  The instructions and code required for the conversion of EQ-5D quality of life data into preference weights are available as a CREST FactSheet.  In due course, additional algorithms will be added for the conversion of the SF-6D and the EORTC QLQ-C30 into utility values.