Back to workshop overview

W2.3

Title: Methodological considerations to limit bias in low back pain prognosis studies - An international expert consensus.

Authors and affiliation: Jill Hayden 1,2, Pierre Côté 1,2,3 and other international experts (TBA).

1 Institute for Work & Health, Toronto, Canada
2 Department of Health Policy, Management & Evaluation, University of Toronto, Canada
3 Department of Public Health Sciences, University of Toronto, Canada

Email address: jhayden@iwh.on.ca

Introduction: Studies of prognostic factors improve our understanding of disease processes and help clinicians make decisions; they inform the design of intervention studies by helping define risk groups and predicting disease outcomes. In low back pain (LBP) research, the ability to accurately predict the clinical course early may lead to more effective management strategies and would facilitate future investigation. There is an abundance of studies in the literature attempting to predict those LBP patients at risk of chronicity. However there remains uncertainty, partly due to studies of varying quality, with often conflicting results. Prognostic factors are rarely amenable to randomization and, therefore, must be studied with observational study designs. Because these designs are more liable to residual confounding, they warrant careful planning in their conduct and careful critical appraisal in their evaluation. A recent review of methodological quality assessment of prognosis studies included in prognosis systematic reviews (Hayden et al.; to be published in a peer-reviewed journal March 2006) recommends that adequate quality assessment include judgments regarding six potential study biases: 1. study participation, 2. study attrition, 3. prognostic factor measurement, 4. confounding measurement and account, 5. outcome measurement, and 6. analysis. We propose to operationalize this study’s recommendations, with international expert consensus, to address specific issues involved in conducting and evaluating studies of LBP.

Objectives: To develop expert-consensus guidelines for limiting bias in conducting and evaluating LBP prognosis studies.

Methods/content: This project will involve a formal consensus development process, including 1. two or more email rounds with a small group of invited participants, 2. a structured consensus meeting (Forum workshop) and 3. follow-up ratification of a guideline document. In the first step, 10 -12 international experts, selected for their methodological and clinical knowledge, will be invited to participate. Prior to the Workshop, (in March, 2006; shortly after publication of Hayden et al.), background information and a survey will be sent to selected researchers. The survey will outline potential prognostic study biases and specific items that should be addressed for each bias in low back pain research. Experts will be asked to rate each item, and provide comments to justify their ratings. Furthermore, they will be asked to list additional items of importance not represented. Results will be tabulated anonymously and presented to the group in round 2. The workshop will include short presentations by the authors introducing the objectives, 1-2 of the invited participants presenting their views, and a presentation summarizing the results from the initial rounds. Issues that were not resolved in the initial rounds will be highlighted and individually debated.

Outcome: A peer-reviewed publication is planned as an output of this workshop. This publication will include a discussion of methodological challenges in LBP prognosis studies and a summary of our expert-consensus guidelines to limit bias. The discussion and guidelines will be illustrated with examples from the LBP prognosis literature.