The Cochrane collaboration
Donald L. Patrick (Co-Convenor)1, Gordon Guyatt (Co-Convenor)2, Tatiana Gauchon (Coordinator)3
1 University of Washington, Seattle, WA, USA; 2 McMaster University, Hamilton, Ontario, Canada; 3 Mapi Research Trust, Lyon, France
The PRO methods group (PROMG), an entity of the Cochrane Collaboration, focuses on patient-centered outcomes, self-reported by participants in treatment trials. This year, the PROMG has four initiatives.
First is an update of the Cochrane Handbook Chapter on Patient-Reported Outcomes with a focus on methods for pooling continuous data from outcome measures in meta- analyses and for using the minimally important difference (MID) in interpretation. This updated chapter will be prepared by Bradley Johnson, Gordon Guyatt, and Donald Patrick and published in Version 6 of the Cochrane Handbook.
Second, Dr. Johnson has led two articles expanding on our work on using continuous data from outcome measures in meta-analyses. The first paper (ref. 1 below) (i) describes PROs and why they are important for healthcare decision-making, (ii) illustrates the key risk of bias issues that systematic reviewers should consider and, (iii) addresses outcome characteristics of PROs and provides guidance for combining outcomes.
We suggest a step-by-step approach to addressing issues of PROs in meta-analyses. Systematic reviewers should begin by asking themselves if trials have addressed all the important effects of treatment on patients’ quality of life. If the trials have addressed PROs, have investigators chosen the appropriate instruments? In particular, does evidence suggest the PROs used are valid and responsive, and is the review free of outcome reporting bias? Systematic reviewers must then decide how to categorize PROs and when to pool results.
The second paper (ref. 2 below) summarizes approaches to enhancing the interpretability of pooled estimates of PROs in meta-analyses. When differences in PROs between groups are statistically significant, decision-makers must be able to interpret the magnitude of effect. This is challenging when, as is often the case, clinical trial investigators use different measurement instruments for the same construct. For such cases, in addition to pooling results as a standardized mean difference, we recommend that systematic-review authors use other methods to present results. These include relative (relative risk, odds ratio) or absolute (risk difference) as dichotomized treatment effects, complemented by presentation in either: natural units (e.g., overall depression reduced by 2.4 points when measured on a 50-point Hamilton Rating Scale for Depression); minimal important difference units (e.g., where 1.0 unit represents the smallest difference in depression that patients, on average, perceive as important the depression score was 0.38 units less than the control group); or a ratio of means (e.g., where the mean in the treatment group is divided by the mean in the control group, the ratio of means is 1.27, representing a 27% relative reduction in the mean depression score).
Third, we are embarking on a project to collect, analyze, and summarize data from PROs in systematic reviews and meta-analyses by identifying MIDs used in different PROs and evaluating their use. An application has been made to the Canadian Institute on Health Research to support this effort.
Finally, we plan to update the review of PROs used in Cochrane Reviews. This project will be led by Zbys Fedorowicz ([email protected]). We plan on surveying all Cochrane Reviews to summarize and evaluate how PROs are used, particularly in the Summary of Findings Tables, now an integral part of the Reviews. If you are interested in this project, please contact Zbys at the e-mail address above.
Another continuing activity is to provide evaluations of PROs in the actual Cochrane treatment review areas; you can locate them at this website. The next Cochrane Collaboration meeting will be held in Hyderabad, India. Gordon Guyatt and Donald Patrick will facilitate the workshop entitled “Making results of patient-reported outcomes interpretable.” The Patient-Reported Outcomes Methods Group Annual Meeting will also take place at this time.
We thank Mapi Research Trust for providing funds and superb administrative support for our activities, as well as their assistance in setting up our website.
1. Johnston BC, Patrick DL, Busse JW, Schünemann HJ, Agarwal A, Guyatt GH: Patient-reported outcomes in meta-analyses—Part 1: assessing risk of bias and combining outcomes. Health Qual Life Outcomes 2013, 11(1):109.
2. Johnston BC, Patrick DL, Thorlund K, Busse JW, da Costa BR, Schünemann HJ, Guyatt GH: Patient-reported outcomes in meta-analyses—Part 2: methods for improving interpretability for decision-makers. Health Qual Life Outcomes 2013, 11(1):211.