We read the paper of Jordan and colleagues  on the usefulness of physiotherapy care before total knee replacement (TKR) with great interest. The authors narratively synthesized eleven studies on preoperative exercise therapy for TKR that they extracted from six different electronic databases. By use of the CONSORT checklist they evaluated the studies’ methodological quality and concluded that the quality varied considerably. After finishing the manuscript we had mixed feelings regarding some conclusions of this review. At first glance, all looks well; perhaps the data synthesis was forgone for inappropriate reasons . However, a closer look reveals two more pressing issues:
We wonder whether the authors actually answered their research question. Jordan et al.  state that “approximately 20% of patients are unsatisfied following TKR in terms of pain relief and function”. Yet, they did not specify their eligibility criteria towards identifying studies that actually targeted persons at risk for unsatisfactory outcomes following TKR. On the contrary, nearly all studies in the review excluded high-risk individuals – those with comorbidities  – thus including the patients with low risk for unsatisfactory outcomes. If the authors had specifically included studies on preoperative exercise in high risk patient group, perhaps their conclusions had differed.
It seems that we are calling everything physiotherapy nowadays. Back in 2005, Herbert and Bø stated that “[…] physiotherapy, can be administered well or badly. Variation in the quality of administration of interventions may explain […] the variability in […] effects between trials in systematic reviews. We […]should assess the quality of interventions” . Interestingly we hardly see people take the variation in quality of the therapeutic intervention into account [, ]. One could argue that the majority of the interventions of the studies included in the review of Jordan et al. (2014) can be classified as being poorly administered physiotherapy (i.e., not reflecting, high quality, day-to-day clinical practice). Interventions that lack intensity, do not monitor progression, and/or are generally poorly thought out, thus have little potential for effectiveness and should perhaps be weighted differently in syntheses of the literature .
The two points addressed above are not exclusive for the study of Jordan et al. (2014), yet reflect common practice in physiotherapy research, where we put emphasis on methodological validity and forgo therapeutic validity . To improve upon this fallacy we really need to rethink the way we set up exercise programs in randomized clinical trials. Much is either “borrowed” material that is used under the premise that “it worked for [insert a random patient group]” or is just poorly thought out. In some cases, it even seems that research funds have mainly been spent on the logistical aspects of the trial, rather than on the therapeutic aspects of the intervention . Next, we need to start paying attention to the validity of the studied physiotherapeutic interventions in systematic reviews. If we develop new methods or tools that allow us to weed out studies that have no or very low potential to be effective before synthesizing interventions, we might draw substantially different conclusions in the 5000+ systematic reviews on the effectiveness of physiotherapeutic care. Lastly, reviewers for and editors of scientific journals in the field of physiotherapy might need to set a higher standard for studies that claim to study “physiotherapy”.
Can we just call everything physiotherapy?