Preoperative physiotherapy and one-year patient-reported outcomes after primary total knee arthroplasty: a registry-based cohort study of 1,688 patients.

The two groups were never measured before physiotherapy started, so there is no way to tell whether they were comparable at that point. The physiotherapy group had more women and more patients using a walking aid, and probably started with worse function. Nobody was assigned to physiotherapy or to no physiotherapy. Patients either attended or did not, and the groups were sorted out afterwards from the registry, so the comparison reflects who chooses physiotherapy. The paper gives no information on how long the preoperative physiotherapy lasted or what it involved, and none on whether patients had physiotherapy after surgery. All patients came from one hospital's standardised fast-track pathway, so the results may not carry over to services organised differently. The comparisons of joint score, quality of life and pain rely on confidence intervals that do not overlap rather than on a direct statistical test of difference between the groups. One results sentence says 1,094 patients had preoperative physiotherapy, while the table and the rest of the paper say 1,041.

The authors declare no conflicts of interest. The supplied text does not say who funded the study.

Do not read this as proof that preoperative physiotherapy is useless or harmful. The patients who had it were more likely to be women, more likely to be using a walking aid and had worse function to begin with, and neither group was measured before the physiotherapy started, so the slightly better scores in the group who skipped it most likely reflect who those patients were rather than what physiotherapy did.