Article type
Year
Abstract
Background: BackInfo (www.rygginfo.no) is a collaboration between the Norwegian Back Pain Association and the Norwegian Health Services Research Centre. Its goal is to develop high quality, patient-centred, evidence-based information for people with back pain. The primary data sources for the evidence are the Cochrane Back Group s reviews.
Objective: To extract outcomes and comparisons from the Cochrane Back Group reviews that are statistically supported and relevant to people with back pain.
Methods: An international, multidisciplinary team assembled in Oslo, Norway. Cochrane Back Group reviews were used as the primary information sources to identify relevant interventions, comparisons, and outcomes (n=20, The Cochrane Library Issue 2, 2003). Three team members identified the comparisons and outcomes presented in each review. They divided them into (i) included team member agreed with the comparison and the studies included in the outcome, considered it relevant, full data were available and synthesized; (ii) excluded comparison or outcome not considered relevant, or comparison or outcome was relevant, yet insufficient data were provided to allow calculation; and (iii) created data were recalculated due to change in comparison and/or included studies. All decisions on included/excluded/created comparisons and outcomes were reviewed by the full team before a final decision was made.
Results: A total of 63 comparisons and 250 outcomes from the 20 reviews were accepted. Of 99 possible comparisons presented in the reviews, 34% (34/99) were included, while an additional 29 comparisons were created (46%, 29/63). Of 169 possible outcomes presented in the reviews, 52% (88/169) were included, while an additional 162 outcomes were created (65%, 162/250). The overall distribution of outcomes was: pain -- 32% (81/250), functional status -- 37% (87/250), return to work -- 10% (26/250), adverse events 4.8% (12/250), and other (e.g. patient satisfaction) 17% (43/250).
Of the outcomes presented in the reviews, 48% (81/169) were excluded. Exclusion criteria included: missing numerical data (means and standard deviations for continuous variables, event rates for dichotomous variables); or reporting of unspecified time points (e.g. short-term), surrogate measures (e.g. trunk flexion), or non-patient-rated measurements (e.g. surgeon or clinician ratings).
Conclusions: Patient-centred outcomes are important to facilitate clinical decision-making. Deyo et al.[1] advocated for standardization of outcome measures in back pain research. They recommended measuring a core set of outcomes: pain symptoms, back-related function or general well-being, disability and satisfaction with care. In a recent qualitative study of people with back pain, Glenton [2] discerned that there was an unmet need for practical, understandable information for patients.
Based on our experience with this project, it would be easier to extract and understand patient-centred outcomes if primary studies were well conducted and reported consistent outcomes, measured with validated tools, at more explicit time points; and systematic reviews followed a more consistent method of reporting and synthesizing data from primary studies. Acknowledgements: Thanks to Eva Schonstein for her contributions.
References: 1. Deyo, RA. Outcome measures for low back pain research. A proposal for standardized use. Spine. Sep 15 1998;23(18):2003-13. 2. Glenton C. Developing patient-centred information for back pain sufferers. Health Expect. Dec 2002;5(4):319-29.
Objective: To extract outcomes and comparisons from the Cochrane Back Group reviews that are statistically supported and relevant to people with back pain.
Methods: An international, multidisciplinary team assembled in Oslo, Norway. Cochrane Back Group reviews were used as the primary information sources to identify relevant interventions, comparisons, and outcomes (n=20, The Cochrane Library Issue 2, 2003). Three team members identified the comparisons and outcomes presented in each review. They divided them into (i) included team member agreed with the comparison and the studies included in the outcome, considered it relevant, full data were available and synthesized; (ii) excluded comparison or outcome not considered relevant, or comparison or outcome was relevant, yet insufficient data were provided to allow calculation; and (iii) created data were recalculated due to change in comparison and/or included studies. All decisions on included/excluded/created comparisons and outcomes were reviewed by the full team before a final decision was made.
Results: A total of 63 comparisons and 250 outcomes from the 20 reviews were accepted. Of 99 possible comparisons presented in the reviews, 34% (34/99) were included, while an additional 29 comparisons were created (46%, 29/63). Of 169 possible outcomes presented in the reviews, 52% (88/169) were included, while an additional 162 outcomes were created (65%, 162/250). The overall distribution of outcomes was: pain -- 32% (81/250), functional status -- 37% (87/250), return to work -- 10% (26/250), adverse events 4.8% (12/250), and other (e.g. patient satisfaction) 17% (43/250).
Of the outcomes presented in the reviews, 48% (81/169) were excluded. Exclusion criteria included: missing numerical data (means and standard deviations for continuous variables, event rates for dichotomous variables); or reporting of unspecified time points (e.g. short-term), surrogate measures (e.g. trunk flexion), or non-patient-rated measurements (e.g. surgeon or clinician ratings).
Conclusions: Patient-centred outcomes are important to facilitate clinical decision-making. Deyo et al.[1] advocated for standardization of outcome measures in back pain research. They recommended measuring a core set of outcomes: pain symptoms, back-related function or general well-being, disability and satisfaction with care. In a recent qualitative study of people with back pain, Glenton [2] discerned that there was an unmet need for practical, understandable information for patients.
Based on our experience with this project, it would be easier to extract and understand patient-centred outcomes if primary studies were well conducted and reported consistent outcomes, measured with validated tools, at more explicit time points; and systematic reviews followed a more consistent method of reporting and synthesizing data from primary studies. Acknowledgements: Thanks to Eva Schonstein for her contributions.
References: 1. Deyo, RA. Outcome measures for low back pain research. A proposal for standardized use. Spine. Sep 15 1998;23(18):2003-13. 2. Glenton C. Developing patient-centred information for back pain sufferers. Health Expect. Dec 2002;5(4):319-29.