Effect of a walking skill training program in patients who have undergone total hip arthroplasty: Followup one year after surgery

Authors


  • ClinicalTrials.gov identifier: NCT00808483.

Abstract

Objective

To investigate the effect of a 12-session walking skill training program of weight-bearing activities on physical functioning and self-efficacy initiated in patients 3 months after total hip arthroplasty (THA).

Methods

Sixty-eight patients with THA, 35 women and 33 men, with a mean age of 66 years (95% confidence interval [95% CI] 64, 67 years), were randomized to a training group (n = 35) or a control group without physiotherapy (n = 33). Assessments were performed before the intervention at 3 months (pretest), at 5 months (posttest 1), and at 12 months (posttest 2) after surgery. The primary outcome was the 6-minute walk test (6MWT). The secondary outcomes were the stair climbing test (ST); figure-of-eight test; Index of Muscle Function (IMF); active hip range of motion (ROM) in flexion, extension, and abduction; Harris Hip Score (HHS); self-efficacy; and Hip Dysfunction and Osteoarthritis Outcome Score.

Results

The training group had larger improvements than the control group at posttest 1 on the 6MWT with an adjusted mean difference of 52 meters (95% CI 29, 74 meters; P < 0.001) and on the ST of −1 second (95% CI −2, 0 seconds; P = 0.01).There were also improvements on the figure-of-eight test (P = 0.02), IMF (P = 0.001), active hip ROM in extension (P = 0.02), HHS (P = 0.05), and self-efficacy (P = 0.04). The difference between the groups persisted at posttest 2 on the 6MWT of 52 meters (95% CI 24, 80 meters; P < 0.001) and on the ST of −1 second (95% CI −3, 0 seconds; P = 0.05).

Conclusion

The walking skill training program was effective, especially in improving walking both immediately after the intervention and 1 year after THA surgery.

INTRODUCTION

Total hip arthroplasty (THA) is a common surgical procedure that is used when other methods, such as medical treatment and physiotherapy, have failed to modify pain or physical limitations caused by osteoarthritis (OA) of the hip. Patients report that their highest priorities after THA are to obtain pain relief, return to normal daily functioning, and maintain an active lifestyle (1, 2). Walking is related to living an active and independent life and is therefore important for achieving these goals (3). Previous research has shown that 3 months after THA, patients have less pain and improved walking than before surgery (3–5). However, studies show that several months and years after THA, patients' walking is impaired compared to that of healthy peers (6–13), and hip flexibility and muscle strength are poorer than in their unaffected hip (14–16). As prostheses are developed to tolerate more vigorous activity, it seems increasingly likely that these patients should be able to approximate their walking to that of healthy peers.

Physiotherapy, especially related to exercises, is a major part of patient rehabilitation following THA. A systematic review of effect studies on exercises showed variations in both program content and the time period after surgery when the program was administered (17). For example, the program focus varied from flexibility exercises, strengthening exercises, and cardiorespiratory fitness training in non–weight-bearing positions (18–21) to postural stability and aerobic dance in weight-bearing positions (22, 23). The start of the intervention varied from shortly after surgery to several months or even years afterward (19–21). The authors of the review stated that it was not possible to establish the efficacy of the various exercise programs in improving physical functioning after THA (17).

It seems likely that patients with THA have a longstanding history of hip pain and dysfunction before surgery that may have resulted in behavioral adaptations and altered movement patterns. Therefore, if the goal of physiotherapy is to restore some degree of previous walking (24), an adequate improvement for the patients would be to relearn the movement patterns they had prior to their illness, i.e., to improve their walking skill. According to one theory of motor control, skills are learned by practicing the particular activity in different ways (25, 26); in the present case, this would mean practicing walking tasks and activities. Guidance and feedback when practicing the task may promote the relearning of motion. By practicing these tasks and activities, it is also thought that joint flexibility, muscle strength, balance, and muscle endurance are trained. We therefore designed a program in which the patients relearned motor skills by practicing walking on uneven surfaces, stepping over different obstacles, and climbing stairs. To avoid limiting factors such as pain and mobility restrictions to prevent hip dislocation, the present walking skill training program was implemented 3 months postoperatively. A better movement pattern, balance, and muscle endurance were assumed to improve walking distance, and an assessment of walking distance during 6 minutes was chosen as the primary outcome variable.

The aims of the study were to examine the immediate effects of a walking skill training program on walking, stair climbing, balance, self-reported physical functioning, pain, and self-efficacy compared to a control group without supervised physiotherapy, and second, to examine whether the effects persisted 12 months after surgery.

Significance & Innovations

  • The walking skill training program performed 3 to 5 months after total hip arthroplasty had immediate effects on walking distance and stair climbing that still persisted 1 year after surgery.

  • Training in weight-bearing activities was safe and well tolerated by the patients.

PATIENTS AND METHODS

Study design and sample size estimation.

A single-blind randomized controlled study design was used. The patients were randomized to either the training group or the control group receiving no physiotherapy by drawing an opaque envelope containing a note assigning them to one of the groups. Thirty-five envelopes were prepared for each group. The measures were administered before the intervention started at 3 months after surgery (pretest), on completion of the program at 5 months after surgery (posttest 1), and at 12 months after surgery (posttest 2). The assessments were performed by a single physiotherapist, who was blinded for group allocation. The primary outcome was the distance walked during 6 minutes. To detect a clinically relevant difference of 50 meters (27) with an SD of 70 meters (28), a statistical power of 80%, and a statistical significance of 0.05, the size of the sample was estimated to be 32 patients in each group (29). To take account of possible dropouts, we wanted to include 70 participants. The study was approved by the regional committee for medical research ethics and Norwegian Social Science Data Services, and registered online at ClinicalTrials.gov.

Participants.

Patients who were scheduled for primary unilateral THA at 2 hospitals were sent information about the ongoing study before hospitalization. They were asked to participate in a longitudinal study the day before surgery if they met the following criteria: diagnosis of OA of the hip joint (30) and residence close to the hospital so as to be able to attend training sessions, i.e., within a radius of approximately 30 km. They were excluded if they had OA in a knee or the contralateral hip that restricted their walking, a neurologic disease, dementia, heart disease, drug abuse, and inadequate ability to read and understand Norwegian. Three months after surgery, those willing to participate preoperatively were asked again to participate in the present study. The patients were enrolled consecutively from October 2008 to March 2010.

Surgery and physiotherapy during the hospital stay.

Exeter or Spectron hip prostheses were used, and for the operation a posterolateral approach was chosen with a curved incision of approximately 13 cm posterior to the gluteus medius muscle and the greater trochanter. The short external rotators were detached, and reattached during closure (31). The acetabular and femoral components were cemented and the posterior capsule was sutured. While in the hospital, all of the patients followed the same anesthetic procedure during and after surgery, which included intraarticular local injection analgesia. After surgery there were no restrictions on weight bearing. To prevent dislocation, the patients were told to avoid hip flexion beyond 90°, hip adduction, and hip internal rotation beyond the neutral position in the operated hip for the first 3 months (32). All of the patients got daily routine physiotherapy care for approximately 30 minutes, which consisted of self-care instructions, joint mobility and muscle-strengthening exercises in a bed or on a bench, and learning to walk with a supporting device.

Physiotherapy and training after discharge from the hospital.

The patients got requisition for physiotherapy when they were discharged from the hospital. From training logs we know that 73% of the patients exercised under supervision of a physiotherapist, mainly comprising flexibility and strengthening exercises on a bench or in an apparatus. All but 1 patient reported additionally that they had done home exercises and walks for more than twice a week.

Walking skill training program.

The program was performed in groups of 2 to 8 patients, and the group was led by a physiotherapist. Each patient participated in 12 sessions, which were held twice a week. Each session lasted for 70 minutes. Before the training started the patients were asked to identify some activities they wished to become better at, and 79% reported that they wished to improve their walking ability and 21% to improve their balance. This was taken into consideration when adjusting the training program to the individual patient.

The program was based on 2 main principles: to train neuromuscular functioning by doing several repetitions of different ambulatory tasks and activities, and to relearn more adequate movement patterns from guidance and feedback of the physiotherapist (Table 1). During the sessions the difficulty and number of repetitions of the exercises were continuously adjusted by the physiotherapist to each individual's level of physical functioning, personal goals of improvement, and progress over time. When the patient managed to do one activity, they had to practice the activity in a more demanding way, for example, by increasing the speed of the movements and height of the walking obstacles, as well as making the ground more uneven. To avoid cardiovascular risks, the patients should be able to talk while exercising.

Table 1. Components of the walking skill training program
Weight-bearing tasks with supervision, guidance, and feedbackTargetDescription
Warm-up with music Ten minutes of standing with weight transfers, sidesteps with arm swing, walking in a circle at different speeds, and step length
Sit to standStrength and flexibilityFive minutes of rising from and lowering onto a chair and squats at different speeds and with weight transfers
LungesStrength, stretching, and balanceFive minutes of lunges forward and sideways on alternative legs
Single-leg stanceStrength and balanceFive minutes of single-leg stance on alternative legs while moving the other leg
Standing on foam balance padBalance and strengthTen minutes of squats, forward, backward, and sideways, with increasing angles in the hips and knees
Step up/step downBalance, strength, and flexibilityFive minutes of ascending and descending a step, forward and backward and at different speeds and different step heights
Stair climbingBalance, strength, and flexibilityFive minutes of going up and down 5 steps with different heights and at different speeds
Obstacle courseBalanceTen minutes of stepping over obstacles; stepping onto, along, and down from an aerobic step and Bosu ball; walking over a foam mat and progressing by increasing the speed, height, and number of obstacles
Throwing ballBalance and coordinationFive minutes of throwing and catching a ball to each other in a circle while moving around
WalkingEndurance, balance, and flexibilityFive minutes of walking in a crowded corridor at different speeds and step lengths, with turns and progression to maximal walking speed
StretchingFlexibilityFive minutes of stretching the calf, leg, thigh, neck, and shoulder muscles

Control group.

The control group did not attend any supervised physiotherapy programs during the same time period, but were encouraged to continue with the exercises they had learned in the hospital or during their rehabilitation stay, and to keep generally active.

Performance-based measures of physical functioning.

The 6-minute walk test (6MWT) measures the distance in meters walked indoors at a comfortable speed for 6 minutes (33), and is considered an adequate measure of physical functioning in subjects with OA after THA (28, 34). The patients walked back and forth along a 40-meter hospital corridor for 6 minutes. The test is considered reliable and valid (35). A change in walking distance of 50 meters has been proposed to be of substantial clinical relevance (27), and the number of patients in the training and control groups who increased their walking distance by 50 meters or more was registered.

In the stair climbing test (ST), the patients ascend and descend 8 steps with a step height of 16 cm as fast as they can without running. The patients were instructed to use alternate legs and were allowed to support themselves by holding onto the stair rail. The time was measured in seconds.

The figure-of-eight test consists of a double set of circles, where the outer circles have a diameter of 180 cm and the inner circles have a diameter of 150 cm. During walking the feet must be placed in the 15-cm space between the lines. Every step on and outside the lines was registered, and the higher the number the worse the score. The test is reported to be reliable and valid (36).

The Index of Muscle Function (IMF) consists of tests of general mobility, muscle strength, balance/coordination, and endurance. A patient's performance is evaluated by the assessor on a 3-point scale (range 0–2) (37). The total score ranges from 0 (best) to 40 (worst). The IMF has been tested for validity and reliability for patients with OA (37).

Active range of motion (ROM) of hip flexion, extension, and abduction was measured by a goniometer according to the procedures of Norkin and White (38).

Self-report measures of physical functioning and self-efficacy.

The Harris Hip Score (HHS) is a widely used disease-specific measure of hip disabilities after THA. The physiotherapist administers the test in the form of a structured interview with the patient. The domains include pain, functions of daily living, and gait (39). The rating scale is from 0 (worse) to 100 points (best). The HHS is considered to have good validity and reliability (40).

According to Bandura, self-efficacy for activities is best evaluated when the questions are tailored to the particular domains of interest (41). Therefore, we constructed 10 questions about challenging everyday ambulatory activities. The questions were: “How certain are you that you can climb two floors of stairs without using a rail, walk 2 km in the woods at normal speed, bend down, squat, kneel, go shopping in a crowd, avoid falling indoors and outdoors, cross the street at a green light, and lead an active life style?” Each question had a scale ranging from 0 (very uncertain) to 10 (very certain). The responses were calculated as a sum score (range 0–100). The internal consistency of the responses to the items was found to have a Cronbach's alpha of 0.78 in our material.

The Hip Dysfunction and Osteoarthritis Outcome Score (HOOS) LK 2.0 is a disease-specific questionnaire that consists of 5 subscales that address pain, other symptoms, activities of daily living (ADL), functions of sport and recreation, and hip-related quality of life (QOL). The scores range from 0 (worse) to 100 points (best). The HOOS has been found to be valid and responsive (42). We translated the Swedish HOOS 2.0 version into Norwegian according to a standard procedure (43).

Adverse events during the training sessions, such as injuries and falls, were registered by the physiotherapist. At posttest 2 the patients filled in a questionnaire to register whether they had experienced any falls, hip dislocation, or loosening of the prosthesis, thrombophlebitis, or deep vein thrombosis.

Statistical analyses.

We used the software program SPSS, version 18.0, for the statistical analyses. Descriptive data are shown as means with their 95% confidence intervals (95% CIs). Data from all randomized subjects were included in the data analysis. Few dropped out during the study period (Figure 1), and the missing data were considered to be unrelated to the intervention and the outcome of the subjects. In order to obtain a complete data set, the last observation of the subject was carried forward to replace the missing data. The chi-square test was used for comparison of categorical variables, and the continuous data were analyzed by Student's sample t-tests. There were no statistically significant differences between the groups in the change from posttest 1 to posttest 2. At pretest, there was a statistically significant difference between the groups in HOOS ADL. There was also a trend toward more women in the training group than in the control group, and that the training group had worse pretest scores. Therefore, the effects of therapy were analyzed by a general linear model with the posttest scores as dependent variables, and sex, group variable, and pretest scores as covariates. The results are given as adjusted means, their 95% CIs, and the partial eta-squared effect sizes. A partial eta-squared value between 0.01 and 0.05 (range 1–5%) is considered a small effect, between 0.06 and 0.13 (range 6–13%) is considered a moderate effect, and between 0.14 and 1 (range 14–100%) is considered a large effect (44). A P value of less than or equal to 0.05 was considered statistically significant.

Figure 1.

Flow chart of the participants through the study.

RESULTS

Approximately 250 patients with residence close to the hospital were operated on in the 2 hospitals during the inclusion period. Some of those did not fulfill the inclusion criteria, and others were not asked because of the busy time schedules at the hospitals. The number of these patients was not registered. There were 128 patients fulfilling the inclusion criteria and were asked to participate preoperatively. Ninety-two patients agreed to participate before surgery, but 24 of them declined to participate when asked again at 3 months postoperatively. These patients, 19 of them women, had worse scores preoperatively than the participants on the 6MWT, ST, hip ROM in flexion and extension, HHS, and HOOS ADL, sport, and QOL (P < 0.05) (data not shown). Sixty-eight patients were randomized to 2 groups, with 35 patients in the training group and 33 in the control group. Before starting the intervention, 1 patient in each group withdrew from the study, 1 patient in the training group attended 5 sessions and then dropped out because of a busy working schedule, and 1 patient in the training group was lost to followup at posttest 2 because of cancer treatment (Figure 1).

The study sample comprises 35 women and 33 men ages 45–81 years, with a mean age of 66 years (95% CI 64, 67 years). The patient characteristics are shown in Table 2. There were no statistically significant differences between the groups at pretest, apart from those for the HOOS ADL (P < 0.05) (Table 3), but there was a trend toward more women in the training group than in the control group, which was adjusted for in the analysis. In the training group, 30 patients attended 12 sessions and 4 attended 8 to 11 sessions.

Table 2. Characteristics of the patients who had undergone total hip arthroplasty*
 Training group (n = 35)Control group (n = 33)
  • *

    Values are the number (percentage) unless otherwise indicated. There were no statistically significant group differences. 95% CI = 95% confidence interval; BMI = body mass index.

Age, mean (95% CI) years65 (63, 68)66 (63, 69)
Women21 (60)14 (42)
Men14 (40)19 (58)
BMI, mean (95% CI) kg/m227 (26, 29)27 (25, 28)
Educational level ≤12 years14 (40)15 (46)
Educational level >12 years21 (60)18 (55)
Exeter prosthesis26 (74)24 (73)
Spectron prosthesis9 (26)9 (27)
Married/cohabitating27 (77)25 (76)
Comorbidity, no.  
 Cancer12
 Osteoporosis02
 Musculoskeletal disorders63
 Stomach/intestinal problem22
 Lung disease01
 Psychological disorder01
Table 3. Within-group changes in walking and physical functioning from pretest (3 months) to 5 months (posttest 1) and 12 months (posttest 2) in patients who had undergone total hip arthroplasty*
 Training group (n = 35)Control group (n = 33)
PretestPosttest 1Posttest 2PretestPosttest 1Posttest 2
  • *

    Values are the mean (95% confidence interval). 6MWT = 6-minute walk test; ST = stair climbing test; IMF = Index of Muscle Function; ROM = range of motion; HHS = Harris Hip Score; HOOS = Hip Dysfunction and Osteoarthritis Outcome Score; ADL = activities of daily living; QOL = quality of life.

  • Within-group differences from pretest to posttest 1 and from posttest 1 to posttest 2 at the 0.1% level.

  • Within-group differences from pretest to posttest 1 and from posttest 1 to posttest 2 at the 1% level.

  • §

    Within-group differences from pretest to posttest 1 and from posttest 1 to posttest 2 at the 5% level.

  • Between-group difference at pretest at the 5% level.

Performance measures      
 6MWT, minutes431 (403, 459)507 (478, 537)530 (501, 559)446 (415, 477)468 (436, 500)489 (457, 521)
 ST, seconds13 (12, 14)11 (10, 12)11 (10, 11)§12 (11, 14)12 (10, 13)§11 (10, 13)
 Figure-of-eight test, steps11 (6, 16)8 (4, 12)§7 (4, 11)9 (5, 13)9 (5, 13)7 (4, 11)
 IMF13 (11, 15)9 (7, 10)8 (6, 10)11 (9, 14)10 (8, 13)§9 (7, 11)
 Active ROM in hip flexion, °85 (81, 89)88 (85, 91)95 (91, 99)87 (84, 90)92 (89, 94)94 (90, 98)
 Active ROM in hip extension, °−1 (−2, 1)1 (−1, 2)0 (−1, 1)−2 (−4, 0)−2 (−4, −1)−1 (−3, 0)
 Active ROM in hip abduction, °18 (16, 19)23 (21, 25)24 (22, 27)§20 (18, 22)22 (20, 25)25 (23, 28)
Self-report measures      
 HHS83 (79, 88)92 (90, 95)94 (92, 97)§87 (84, 90)91 (88, 94)93 (89, 97)§
 Self-efficacy71 (64, 79)84 (79, 89)86 (81, 90)79 (72, 85)83 (77, 90)86 (79, 93)
 HOOS symptoms77 (73, 81)80 (76, 84)85 (81, 89)79 (76, 82)82 (77, 86)88 (84, 92)
 HOOS pain85 (80, 90)91 (87, 95)93 (90, 96)89 (86, 93)91 (87, 95)95 (92, 98)
 HOOS ADL81 (77, 86)88 (85, 92)91 (88, 94)87 (84, 90)90 (88, 93)92 (89, 96)
 HOOS sport61 (54, 69)76 (70, 83)77 (70, 84)69 (61, 76)76 (69, 82)§80 (73, 88)
 HOOS QOL66 (59, 73)76 (70, 82)80 (74, 85)71 (65, 76)77 (71, 83)§84 (78, 89)§

Effects of the walking skill training program.

Both groups improved from pretest to posttest 1 on most of the outcome measures (P < 0.05) and on several outcome measures from posttest 1 to posttest 2 (P < 0.05) (Table 3).

Compared with pretest scores, the training group had a statistically significant larger improvement at posttest 1 on the 6MWT (P < 0.001), ST (P = 0.01), figure-of-eight test (P = 0.02), IMF (P = 0.001), active hip ROM in extension (P = 0.02), HHS (P = 0.05), and self-efficacy (P = 0.04) than the control group (Table 4). For the 6MWT the effect size was 0.24, for the ST the effect size was 0.10, and for the other variables the effect sizes varied from 0.06–0.12. Twenty-three patients (66%) in the training group and 5 (15%) in the control group had increased their walking distance to ≥50 meters at posttest 1 compared to the pretest distance (P < 0.001). There were no adverse events registered in the training group at posttest 1.

Table 4. Descriptive data at posttests and effects of a walking skill training program in patients who had undergone total hip arthroplasty*
 Posttest 1 (5 months after surgery)Posttest 2 (12 months after surgery)
Training group (n = 35)Control group (n = 33)Group differenceTraining group (n = 35)Control group (n = 33)Group difference
  • *

    Values are the adjusted mean (95% confidence interval). The data are adjusted for sex and pretest scores. 6MWT = 6-minute walk test; ST = stair climbing test; IMF = Index of Muscle Function; ROM = range of motion; HHS = Harris Hip Score; HOOS = Hip Dysfunction and Osteoarthritis Outcome Score; ADL = activities of daily living; QOL = quality of life.

  • Between-group differences at posttest 1 and posttest 2 at the 0.1% level.

  • Between-group differences at posttest 1 and posttest 2 at the 1% level.

  • §

    Between-group differences at posttest 1 and posttest 2 at the 5% level.

Performance measures      
 6MWT, minutes513 (497, 529)462 (445, 478)52 (29, 74)535 (516, 555)483 (463, 503)52 (24, 80)
 ST, seconds11 (10, 11)12 (11, 13)−1 (−2, 0)10 (9, 11)12 (11, 13)−1 (−3, 0)§
 Figure-of-eight test, steps7 (5, 9)10 (8, 12)−3 (−5, 0)§7 (5, 8)8 (7, 10)−1 (−3, 1)
 IMF8 (7, 9)11 (9, 12)−3 (−4, −1)7 (6, 9)10 (8, 11)−2 (−5, 0)
 Active ROM in hip flexion, °88 (86, 91)91 (89, 94)−3 (−7, 0)95 (91, 98)94 (90, 98)1 (−5, 6)
 Active ROM in hip extension, °0 (−1, 2)−2 (−3, −1)2 (0, 4)§0 (−2, 1)−1 (−3, 0)1 (−1, 3)
 Active ROM in hip abduction, °24 (22, 25)22 (20, 23)2 (0, 4)25 (23, 27)25 (23, 27)0 (−3, 3)
Self-report measures      
 HHS93 (91, 96)90 (87, 92)3 (0, 7)§96 (93, 98)92 (90, 95)3 (−1, 7)
 Self-efficacy86 (82, 90)81 (77, 84)6 (0, 11)§88 (83, 92)84 (79, 88)4 (−2, 10)
 HOOS symptoms81 (77, 84)81 (78, 84)0 (−5, 4)86 (82, 89)87 (84, 91)−2 (−6, 3)
 HOOS pain92 (89, 95)90 (87, 93)2 (−2, 6)94 (91, 96)94 (92, 97)−1 (−4, 3)
 HOOS ADL90 (88, 92)89 (86, 91)2 (−1, 5)92 (90, 95)91 (88, 94)1 (−3, 5)
 HOOS sport78 (73, 83)74 (68, 79)4 (−3, 12)79 (73, 86)78 (72, 84)1 (−8, 10)
 HOOS QOL77 (73, 82)76 (71, 80)2 (−5, 9)81 (76, 86)83 (78, 88)−2 (−9, 5)

There were no statistically significant differences between the groups in the change from posttest 1 to posttest 2 (P > 0.05). However, the improvements from pretest to posttest 2 were larger in the training group than in the control group on the 6MWT (P < 0.001) and ST (P = 0.05) (Table 4). The partial eta-squared effect size was 0.18 for 6MWT and 0.06 for ST. Twenty-six patients (74%) in the training group and 15 (46%) in the control group had increased their walking distance to ≥50 meters at posttest 2 compared to the pretest distance (P < 0.001). At posttest 2, nine patients in the training group and 15 patients in the control group reported falls (P > 0.05). No complications were reported in terms of hip dislocation, loosening of the prosthesis, thrombophlebitis, or deep vein thrombosis.

DISCUSSION

Immediately after completing the walking skill training program, the training group had better scores than the control group on performance-based and self-reported physical functioning, pain, and self-efficacy scores. The improvement was found to be particularly great for walking distance, and at 12 months after surgery the differences between the groups on walking distance and stair climbing time still persisted. No complications were reported, and the program seemed to be well tolerated by the patients.

The walking skill training program was designed to improve walking. Our results suggest that the program served its purpose, as both walking distance and stair climbing time had improved more in the training group than in the control group immediately after the intervention. Despite the fact that both groups improved further from immediately after the intervention to 1 year after surgery, the training group kept their advantage above the control group at 1 year. The question is, however, whether the differences between the groups were of any clinical importance. The mean difference in effects on walking distance between the groups from pretest to posttest 1 and 2 was approximately 50 meters in favor of the training group at both time points, which is considered to be a substantial meaningful improvement (27). A larger number of patients in the training group than in the control group showed improvements above 50 meters compared with pretest scores at both posttest 1 and 2, and the effect sizes at both posttests were large (44). It is unknown, though, how much change in stair climbing time is necessary to reach clinical significance, but the effect sizes indicate that the changes were moderate. Less fall events were reported by the patients in the training group than in the control group 1 year after surgery. Taken together, these results suggest that the effects of the walking skill training program had clinical importance.

Improvements in walking have also been reported in previous studies (18–20, 23). However, it is difficult to compare the effects across studies because walking has been assessed in different ways and at different time periods after surgery. The exercise programs focused mostly on training of muscle strength and joint flexibility in non–weight-bearing positions (18–20). Only 1 study comprised weight-bearing exercises through aerobic dance (23). Furthermore, the sample sizes were mostly small, from 7 to 26 patients in the exercise groups, and consequently the studies tended to have low statistical power. In these studies, clinical relevance of the findings was not considered, and the long-term effects of the programs were not examined. Therefore, the substantial and maintained effects of our walking skill training program with no adverse events are promising. However, the study has to be replicated by others before definite conclusions about its effectiveness can be drawn.

Before starting the training, the patients in the training group reported that their highest priority was to improve walking and balance, and they considered that their walking was not as good as they would have liked it to be. We could therefore be sure that the walking skill training program was in line with the patients' priorities and likely to be motivating. In this particular training program, the patients practiced exercises that involved walking in many different ways and with several repetitions under guidance and feedback from a physiotherapist. The patients in the training group said that this kind of training was more mentally and physically demanding than any physiotherapy program they had experienced previously. Despite the fact that the program was demanding, it was well tolerated by the patients during the training sessions and no complications occurred. Consequently, the walking skill training program appears to be safe and well tolerated by the patients.

Although improvements were shown for most of the performance-based measures and self-reports of self-efficacy and the HHS at posttest 1, the groups differed only with respect to walking and stair climbing at posttest 2. Except for walking distance, the differences between the groups in the outcome measures were small at posttest 1. The scores on several measures of both groups were relatively high at all times of assessment. Therefore, ceiling effects may have appeared. The activities listed in the questionnaires as well as the performance-based measures might have been too easy or unsuitable for our patients being examined several months after surgery.

The physiotherapist's supervision and guidance was an important integrated part of the program, but it is well known that attention alone may have a substantial effect (45). This issue has to be kept in mind when interpreting the data. Furthermore, missing data are a concern for researchers because no statistical method will ever be able to replace missing information (46). However, only 4 patients dropped out, and they were likely to be missed at random and not threatening our randomization. The last observation carried forward is widely used in clinical trials (46), and it is not likely to overestimate the effects in our study because all of the patients improved during the study period. Some uncertainty is related to the results of self-reports, but the primary outcome is not affected by this and is considered to be robust.

Another important issue is the external validity of the study. The patients in the catchment area of the hospitals tend to have a higher educational level than the general Norwegian population (47). However, educational level did not influence the effect estimates when included as a covariate in our statistical model. More patients were operated on at the hospitals during the inclusion period than we have data from. In most cases these patients did not fulfill the inclusion criteria, but there were also some not asked because of a busy clinical time schedule. However, these patients are likely to be missed at random and should not cause any skewness in our data material. Those who were lost from preoperatively to 3 months postoperatively or declined to participate in the training study had worse preoperative scores than the participants, and they may also have been worse after surgery. It is probably a methodologic problem in most training studies that the unmotivated and less healthy people decline to participate. Our patients had a rather normal body mass index, which is in line with our impression that obesity is not necessarily typical among Norwegian patients receiving THA. Altogether we believe that our results can be generalized to relatively healthy, motivated, and nonobese patients who have undergone THA.

In conclusion, the walking skill training program performed 3 to 5 months after THA was effective in improving physical functioning, especially walking distance, and the effect on walking distance and stair climbing still persisted 12 months after surgery in relatively healthy and nonobese patients. No adverse events occurred, and the weight-bearing training was well tolerated by the patients.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Ms Heiberg had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Heiberg, Bruun-Olsen, Mengshoel.

Acquisition of data. Heiberg, Bruun-Olsen.

Analysis and interpretation of data. Heiberg, Ekeland, Mengshoel.

Acknowledgements

We would like to thank physiotherapist Mary Deighan Hansen at Martina Hansens Hospital and physiotherapist Anne Gunn Kallum at Bærum Hospital, Vestre Viken Hospital Trust, for their efforts in recruiting the patients, performing the measurements, and collecting the data. We also thank the physiotherapy staff at Martina Hansens Hospital and Bærum Hospital, Vestre Viken Hospital Trust, for their support and the patients for their participation.

Ancillary