Guidelines for reporting economic evaluations in health care are very welcome. A task force of the International Society for Pharmacoeconomics and Outcomes Research has organized a two-round, modified Delphi Panel process to compile a set of recommendations for reporting Consolidated Health Economic Evaluation Reporting Standards. The result is a 24-item checklist with the friendly acronym CHEERS (Reference Husereau, Drummond and Petrou1), published in this issue of the International Journal of Technology Assessment in Health Care.
This brief version of CHEERS is published simultaneously in ten medical journals, with British Medical Journal taking the lead. A more extensive Explanation and Elaboration Report is published in Value in Health (Reference Husereau, Drummond and Petrou2). The instructions will also be conveniently at hand on the Equator website (www.equator-network.org/resource-centre/library-of-health-research-reporting). Other key guidelines for reporting medical studies – such as CONSORT (Reference Schulz, Altman and Moher3), PRISMA (Reference Moher, Liberati, Tetzlaff and Altman4) and STARD (Reference Bossuyt, Reitsma, Bruns, Gatsonis, Glasziou, Irwig, Lijmer, Moher, Rennie and De Vet5) – are available there as well.
Instructions for writing different types of research designs have been available for years in textbooks and journal articles. The various lists are all based on principles of scientific research, especially clinical epidemiology. Although useful, they do not substitute for real understanding of why a certain detail about a study is essential for its credibility and must be reported.
The trend has been toward more elaborate lists of items, which sometimes means they become less practical. From single lists the development has continued toward separate sets for authors and for readers. Author guidelines specify what a report should contain and quality assessment tools help to evaluate what actually has been written. Some checklists identify the crucial few items that alone can ruin the validity of a study, such as a poor method of outcome measurement or a follow-up time that is too short for the relevant health results to appear.
Separating between the quality of research and the quality of reporting is not always straightforward. A good checklist helps authors and reviewers to spot the key items that must be known, and reported, for judging the internal validity of a study. The possibility of presenting a black box, by providing intractable or too consolidated cost items, should be excluded. A good economic evaluation is transparent and allows the reader to see where the assumptions are similar to those that would be used in another environment.
Reporting monetary units with the ISO-standard format is especially important for IJTAHC, and actually for any journal with a wide international audience. The Wikipedia lists more than fifty countries or regions that use a dollar and reports 22 different currencies called dollar, each with its own three-letter acronym. Although we get more submissions from Canada (CAD) or Australia (AUD) than from Singapore (SGD) or Namibia (NAD), the authors' address or study site do not necessarily help the reader in guessing which dollar is used in reporting. Even though the US dollar still dominates the global currency reserve with a 60% share, a dollar is not only USD.
For understanding the results of a cost-utility analysis and putting them in a wider perspective, the choices of health-related quality of life instrument and the associated valuation algorithm used for calculating quality-adjusted life years are very important but often inadequately specified. It is therefore very welcome that the Explanation and Elaboration Report recommends their clear reporting.
Unfortunately the brief CHEERS statement does not directly recommend reporting these quality of life measures or the use of ISO units for currencies; the full document must be consulted to find these messages. Reading the whole Report is also otherwise highly educational and useful for anyone dealing with economic evaluation.
The guideline distinguishes single study-based economic evaluations and modeling. Researchers reporting models would also benefit from getting acquainted with the PRISMA statement (Reference Moher, Liberati, Tetzlaff and Altman4) so they can provide enough information on how their searches for data were done.
Economist, be economical!
Early in the course of clinical studies, medical students are asked to examine patients and report their findings in complete detail; at the Helsinki University Hospital the nickname for this thorough report was “German status”. Spending two hours with an asthmatic patient, examining not just the lungs but the function of all cranial nerves too (and writing it down in the patient record) served learning. The purpose of CHEERS guidelines is to clarify the reporting, not to make reports excessively long. When a point is not relevant to your research, it does not have to be entered in the report. So, as the guideline recommends: “If applicable, report incremental cost-effectiveness ratios”.
Authors should attempt to identify information that is important for the entire journal readership and report this in the main text. Details that will be of interest to a small number of specialists can be specified in supplementary files. It's seldom useful to publish all figures that your clever software can produce. In health economic reports, for example, our typical reader will rarely be interested in a graph on expected value of perfect information. Instead, most users would like to see a comprehensive list of cost items.
The CHEERS recommendations are best suited for use in classic cost-effectiveness analyses comparing two clinical interventions. More complex study questions, for example regarding organizational or public health interventions, may not be fully covered here. Other guidelines should be used alongside when data are retrieved from both randomized and observational studies.
We congratulate the Task Force for providing this useful and compact advice on minimum reporting requirements for publication of economic evaluations in biomedical journals. The statement ends with an observation it should be reviewed for updating five years from now. We agree, and suggest that the review be based on user feedback, and optimally also on analyzing a sample of quality-assessed economic evaluations. An updated guideline for reviewers and users of economic analyses in health care, identifying key items that save or kill an analysis, would also be welcome.
Competing Interests
Harri Sintonen has received funding as a Board member or consultant from several drug companies (MSD, Merck, Eli-Lilly, Pfizer and Novartis). He is the developer of the 15D and one of the developers of the EQ-5D, and receives royalties from the electronic version of the 15D. The other author has no potential conflicts of interest.
