Gocen 2004.

Methods	Randomised parallel group trial.
Participants	59 participants undergoing total hip replacement. Intervention group: 13 women, 16 men (mean age 47 years, SD 11); control group: 8 women, 22 men (mean age 56 years, SD 14). Inclusion criteria: osteoarthritis of the hip joint, no previous physiotherapy for hip osteoarthritis. Exclusion criteria: other chronic diseases or arthritis in other joints necessitating treatment. Location: Izmir, Turkey.
Interventions	Intervention group (n = 29) were given stretching and strengthening exercises 3 times daily for 8 weeks before surgery, and an education programme that included advice on movements to avoid, use of devices and activities of daily living. Control group (n = 30) received neither the preoperative exercises nor the education programme.
Outcomes	Function* (Harris Hip Score (maximum 100 points) 90‐100: good function and excellent results, 80‐90: good, 70‐80: fair, < 70: poor results); range of motion; pain (VAS; units of the scale not reported); days to mobility (walking and climbing stairs); length of hospital stay. Participants were assessed preoperatively, at discharge, and 3 months and 2 years postoperatively.
Notes	*Harris Hip Score data at 2 years included in Analysis 1.2. Both groups received the same postoperative and education programme beginning the day after surgery.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The patients were randomly divided into two groups using a table of random numbers of a computer programme (Excel 2000)."
Allocation concealment (selection bias)	Unclear risk	Quote: "Even numbers were allocated to the control group and odd numbers to the trial group." Comment: it is not clear who was involved in the allocating the participants or what steps were taken to conceal the random sequence.
Blinding (performance bias and detection bias) Self‐reported outcomes	High risk	Comment: no description of any attempt to blind participants and blinding is unlikely to have been possible. Knowledge of being in the intervention group may have influenced participants' experience of pain.
Blinding (performance bias and detection bias) Objective outcomes	Low risk	Quote: "Postoperative treatment was performed by a physical therapist who was blinded to the patients' groups. ... All measurements were performed by a staff physical therapist who was blinded to the trial." Comment: it is possible that participants could disclose which group they were in during treatment or measurement by the physiotherapist.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: 1 participant assigned to the trial group was not operated on because of cardiovascular problems and was excluded from the analysis. Data for the 59 participants that underwent surgery were available at discharge and 1‐year follow‐up. No mean or SD given for length of hospital stay.
Selective reporting (reporting bias)	High risk	Comment: trialists reported measuring pain with VAS in the methods section, but no pain data were reported in the results section. Also, without the trial protocol, it is unclear whether any other outcomes were measured but not reported based on the results.
Other bias	Low risk	Comment: no other sources of bias identified.