To our knowledge this is the first randomized clinical trial comparing the effect of a patient education plus supervised exercise program to education alone on the time to total hip replacement in patients with an indication for hip replacement surgery.
Such a trial is warranted because exercise as non-surgical treatment may help to postpone or eliminate the need for TJR without deterioration and also improve health-related quality of life through enhanced function and decreased pain. Hence, studying patients with this severe disease condition is clinically important. Further, the primary outcome measure ‘time in days without surgery’ was chosen because TJR should only be recommended once improvements cannot be achieved by other treatment strategies. Moreover, postponing or avoiding TJR is the most relevant outcome when it comes to: reducing health care costs providing the program is cost-effective, minimizing complications, and improving healthcare pathways. Finally, such a trial is also warranted because the recommendations for lower extremity exercises are based on studies of knee OA patients. Although exercise trials on patients with hip OA have been conducted, most of these studies have also included patients with coexisting knee OA, and results were not joint-specifically reported, or too few patients with hip OA were included, preventing valid conclusions to be drawn [6, 9, 12, 15]. Consequently, evidence supporting exercise for patients with hip OA is needed .
The secondary outcome measures used in this trial are those recommended for use in clinical studies of OA, and include patient-reported outcome measures (PROs) for pain, physical function, physical activity, health-related quality of life, and global response to treatment . The patients will be asked to rate their current hip condition at baseline, and at each follow up (HOOS questionnaire), but also to rate possible change in their condition since baseline (GPE questionnaire). The 3-month follow-up was chosen to identify possible improvements at the end of the intervention period. Three months is also a reasonably long timeframe for clinical improvements to occur in patients with long-standing hip pain, yet short enough to assume that patients would be able to recall their baseline condition. The 12-month follow-up was chosen to match the clinically relevant period for postponing hip replacement surgery.
As for the other measures, both muscle mechanical and physical performance-based outcome measures will be part of the evaluation for primarily three reasons. First, muscle mechanic outcomes provide information about the effectiveness of the exercise intervention in terms of improving muscle strength, and both muscle weakness  and muscle asymmetry [46, 47] have been found to be present in hip OA patients. Second, physical performance-based outcomes provide information about the effectiveness of the exercise intervention in terms of improved function in tasks resembling daily activities. Third, muscle atrophy has been associated with functional limitations in aging individuals , and this trial is adequately powered to further explore such associations. Furthermore, measures of exercise-related pain, adherence, and joint-specific adverse events are essential for evaluating the feasibility of the exercise intervention program. Finally, the cost-effectiveness of non-surgical interventions for hip OA is needed . Total costs during the first year will be compared with gains on the HOOS and the patient’s perceived effect (GPE), and the societal costs of the intervention per patient during the first year and the incremental qualify of life in years/days, as obtained from the EuroQol (QALY during the first year). The gain in health and the costs associated with the heath gain could possibly be used as a decision-making aid for optimal resource allocation.
A number of considerations went into designing the supervised exercise intervention in this trial. One consideration was how to prescribe an exercise program, where the level of progression gradually increases, without provoking persistent pain. The use of self-reported pain (VAS) immediately after each exercise session has previously proven to be a feasible way of adjusting the level of progression . Hence, patients who experience recurrent high risk pain after exercise or whose pain either doesn’t return to ‘pain as usual’ or increases during the day following exercise will be asked to reduce the level of exercise to the previous level of progression.
Another consideration was selecting the type of exercises to be used in this trial. An exercise program based on neuromuscular principles has been shown to be patient- and goal-oriented and feasible for elderly patients scheduled for hip and knee replacement . All exercises are individualized with progression guided by the patient’s neuromuscular function, and for each exercise, three levels of progression are available, meaning the program can be tailored to the individual’s pain level and function. It utilizes functional exercises and emphasizes neutral alignment of the hip, knee and foot to resemble conditions in daily life and more strenuous activities, without compromising the quality of the performance. Another exercise program using progressive high load (intensive) resistance training principles has been shown to be effective in elderly patients in the early postoperative phase  and elderly hospitalized patients . Also, evidence of the efficacy of general pre-operative strengthening exercises exists , and strengthening exercises alone show moderate effect on pain and function in patients with OA . Rather than using general strengthening exercises, the use of high load (70-80% of 1RM) in the current trial is based on evidence from systematic reviews, where the intensity of the training was the strongest factor affecting strength and functional outcome [16, 17]. To be effective, exercises must target isolated muscle groups, be progressive [26, 27], and possibly be combined with a more complete exercise program . Both exercise programs in this trial are supervised, based on individualized goal-based patient treatment that allows the program to target the presenting musculoskeletal impairments that contribute to the patient’s symptoms and functional limitations. This approach aligns with the guidelines from EULAR, which recommend tailored treatment .
A final consideration was how to most effectively distribute the two types of exercises throughout the intervention period, and across each week. We decided to mimic performance enhancement in sports. As a result, neuromuscular and intensive resistance exercise sessions will be administered on separate days within the same week, so that separate session goals can then be pursued.
Optimal outcomes in population health require both efficacious treatments and adherence to those treatments. Adherence to health interventions is a complex issue, especially for individuals with chronic conditions . The interaction between patients and providers in terms of verbal communication, physical interaction and empathy could affect the patient’s enthusiasm, adherence, and participation. Nevertheless, it can be enhanced by the use of supervised, class-format exercise sessions in the initial exercise period  combined with later ‘refresher or booster’ sessions [52, 53]. Also, to minimize the provider effect, three physiotherapists will randomly interact and execute the training program at each center in this trial.
The internal validity of the trial may be affected by a negative performance bias in the group who receive patient education only, as they may feel neglected and missing out on active treatment. The exclusion criteria always limit the generalizability to other populations. For example, the results from this trial may not be generalized to an individual patient with hip OA who is not fit enough to take part in moderate-to-high intensity exercise. Conversely, the external validity is improved by the multi-center design, and the delivery of the interventions by multiple practicing community physiotherapists. Also, the intervention will be implemented for both genders, a wide age range, and in patients with severe hip OA. Hence, it is reasonable to assume that the program could benefit the general hip OA population. Lack of adherence is a potential confounding factor for the outcomes in this trial. Consequently, the number of sessions completed will be recorded, and patients will be asked specifically at follow-up points if they have initiated other treatments for their hip condition since the last examination .