Recommendations from the ISPO lower-limb COMPASS: Patient-reported and performance-based outcome measures.

The systematic review behind the process only looked at English-language studies, so non-English measures were missed and use of the COMPASS is restricted where validated translations do not exist. The expert panel that decided which measures were even worth voting on was made up entirely of people based in high-income countries. The Patient-Specific Function Scale was recommended despite being judged to have poor reliability and measurement error. No normative data and no responsiveness statistics such as a minimal important difference were produced, so a clinician cannot yet say whether a given change in a patient's score is meaningful. The EQ-5D-5L has not been tested for measurement properties in people with lower-limb absence. Scoring of the underlying studies is open to individual interpretation, and the standards applied were strict, so many measures failed to qualify. The panel acknowledges that where clinical time is already the main constraint, spending it on outcome measurement could further limit service provision. Of the 46 participants only 7 were physiotherapists and 2 were occupational therapists, against 16 prosthetists and orthotists; 6 were prosthesis users.

The work was coordinated and supervised by the International Society of Prosthetics and Orthotics (ISPO) and funded by the United States Agency for International Development in support of AtScale, managed through the United Nations Office for Project Services. The project team worked with input from ISPO's industry advisory group, and industry representatives took part in the consensus. One author obtained the funding on behalf of ISPO and helped run the consensus process, and states he was not paid for his contribution to the project. The COMPASS user guide is distributed through the ISPO website.

Do not read this as evidence that using these outcome measures improves rehabilitation. The paper reports which measures a consensus group agreed to recommend; no patients were treated and no outcomes were compared, so whether routine measurement changes anything for a patient is untested. It is also not a list of measures with uniformly strong evidence behind them: the Patient-Specific Function Scale was included for its clinical usefulness despite poor reliability, and the EQ-5D-5L has not been tested in this population.