Promising outcome of a hip school for patients with hip dysfunction




To evaluate, in an exploratory study, a hip school led by physical therapists for patients with hip dysfunction including pain.


Physicians in primary care and orthopedic units consecutively recruited patients by residential area to a treatment group (T group; n = 77) and a control group (C group; n = 68), mean age 61.8 years, 59 men, 86 women. Ninety-nine of the 145 subjects had radiologic hip osteoarthritis. Outcome measures were the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) and the Nottingham Health Profile (NHP).


After hip school (6 months) the T group showed greater reduction on the subscales of pain and activity limitations (WOMAC) than the C group. The subscales sleep disturbances, pain, and physical mobility (NHP) showed improvement for the T group compared with at start. The results for the T group persisted after an additional 6 months.


A hip school may reduce pain and activity limitations and improve health-related quality of life for patients with hip dysfunction including pain.


Hip osteoarthritis (OA) has sometimes been considered as a simple mechanical wear and tear phenomenon, with ultimate progression of cartilage and bone destruction and pain (1), and with total hip replacement (THR) as the treatment of choice (2). However, hip OA is a heterogeneous disease, with diagnosis up to now based on radiologic abnormalities, which might coexist with impairment, activity limitation, and participation restriction (3) or might even be asymptomatic (4). The progression of both joint destruction and pain can be stopped or even reversed (5, 6). When pain is severe, THR is the treatment of choice, but necessary only for a minority of patients and preceded by years of less severe symptoms.

The goals of current treatment of patients with symptomatic hip OA include pain control and maintenance or improvement of physical ability and health-related quality of life. Nonpharmacologic modalities, including patient education, are suggested (7). A literature review concerning education of patients with arthritis concluded that improvement of arthritis symptoms by 15–30% is attainable through patient education (8), and a metaanalysis showed that patient education provides, on average, 20% more pain relief than nonsteroidal antiinflammatory drugs alone (9).

Not all persons with hip dysfunction have radiographic signs of hip OA. The prevalence of radiographic hip OA changes in persons attending primary care for the first time for hip pain was 78% (10).

To our knowledge, no studies have assessed an educational intervention solely intended for patients with hip dysfunction, including pain, not sufficiently severe to require THR. The present aim was to evaluate the effect of a hip school for this category of patient with regard to their pain, stiffness, activity limitations, and health-related quality of life.



The sample comprised 171 patients with hip dysfunction, including pain, who were consecutively recruited by physicians in primary care and orthopedic units and assigned to a treatment group (T group) or a control group (C group) according to residential area in 1992–1995. The inclusion criterion was hip dysfunction, which lacking established diagnostic tests or clinical criteria (11–13), we defined as pain in the hip region lasting more than 3 months and manifestations of impaired hip joint range of motion and/or muscle function. All the subjects had a radiologic examination, although radiologic evidence of hip OA was not required for inclusion. The radiologic examinations were performed in several units, with no consistency regarding views or standards. The radiographs were interpreted by the radiologist on duty following normal routines. The diagnosis of OA was made from reading the radiological report. In the physician's inclusion form, the following exclusion criteria were enumerated: trauma; fractures; congenital malalignments; other hip joint diseases; inflammatory joint or neuromuscular diseases; and low back, sacroiliac, or knee problems overshadowing the hip problems. In addition, the subjects could not be listed or fulfill the inclusion criterion for THR defined as severe pain and persisting resting pain despite pharmacologic treatment; tried all other kinds of pain treatments; disturbed night sleep; and walking ability not exceeding 200–300 meters, even with walking aid. The C group was not permitted physical therapy treatment 6 months before or during the study, nor previous participation in a hip school.

In the T group, tests were run before entrance (test 1) to the hip school and after 6 months (test 2), with the tester blind as to prior results. The same test interval applied for the C group. The T group was retested after 1 year (test 3). In the T group, 17 persons dropped out after the initial test, 6 due to THR and 11 declined or did not show up. In the C group, 9 persons dropped out. Five did not want to wait for physical therapy treatment and 4 declined or did not show up. Finally, the sample comprised 145 persons, 77 in the T group and 68 in the C group; mean age was 61.8 years, with 59 men and 86 women. Ninety-nine had OA (Tables 1 and 2).

Table 1. Background variables of treatment (T group) and control group (C group)*
 T group (n = 77)C group (n = 68)T/Z value
  • *

    t-test used for age and BMI; Mann-Whitney U test for the rest. BMI = body mass index; freq = frequency.

  • P ≥ 0.05.

Age, years, mean61.562.2 
BMI, kg/m2, mean27.127.0 
Time since onset, freq   
 <6 months32 
 6–12 months1514 
 1–2 years1714 
 >2 years <5 years159 
 5 years <10 years1514 
 10+ years1215−0.63
First visit to doctor, freq   
 <6 months2426 
 6–12 months129 
 1–2 years86 
 2–5 years159 
 5–10 years96 
 10 years or more912−0.23
Medicine intake per  month, freq   
 All days137 
 Most days68 
 Some days1711 
 Few days1011 
 No days3131−0.90
Walking distance, freq   
 ∼100 m31 
 200–300 m51 
 ∼500 m66 
 ∼1 km1711 
 ∼3 km1918 
 >3 km2731−1.66
Physical activity index   
 1 hardly anything1 
 4 moderate2426 
 6 hard physical activity1−1.28
Satisfaction with physical activity   
 Very satisfied138 
 Somewhat satisfied1010 
 Neither satisfied nor dissatisfied1721 
 Somewhat dissatisfied2127 
 Very dissatisfied162−0.91
Table 2. Background variables of treatment group (T group) and control group (C group)*
 T group frequency (n = 77)C group frequency (n = 68)Chi square
  • *

    Osteoarthritis (OA) verified by radiograph. Unilat = unilateral; bilat = bilateral; Phys = physical; pens = pension.

  • P ≥ 0.05.

  • P ≥ 0.05, Fischer's exact test.

Sex, male/female37/4022/463.69
OA, unilat/bilat OA35/1830/16 
Hip pain, right hip/left hip48/4845/420.23/0.01
Phys therapy visit earlier, “yes”28162.81
Others “yes”142.28
Pain other joints “yes”45370.24
Civil status, married/unmarried/divorced60/3/351/1/5 
 Widow + widower11110.17
Living, not alone/alone60/1751/170.17
Housing, flat/house26/5117/511.97
Education, university/senior high12/138/15 
 Senior level52450.89
Occupation, employed/sick-listed32/826/2 
 Old age/disablement pens32/432/62.05
Sports activity, no/yes37/4035/33 
 Series, competitions22113.58
Walking aid, indoors/outdoors1/81/40.01/0.97
Outdoor bicycling, can/do67/5158/480.09/0.32
Ergometer cycling, can/do73/1263/100.29/0.03


Weight and self recorded height were noted and a questionnaire regarding age, self reported demographics, socioeconomic variables, physical activity, hip pain history, and other medical problems was completed. There were no differences in these variables between the 2 groups (Tables 1 and 2).

The Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) (14) and the generic instrument Nottingham Health Profile (NHP) (15) were used as primary outcome measures (16).

The WOMAC probes symptoms in the areas of pain (5 questions, possible subscale score 0–20), stiffness (2 questions, 0–8), and activity limitations (17 questions, 0–68) (17). A Swedish translation (18) with the Likert version and a time frame “past 48 hours” was used. To enhance the interpretation between the subscales, they were transformed to range from 0 to 100, best to worst, by summing the total score of each subscale, dividing by the possible score for the subscale, and multiplying with 100 (19). In the present study, Cronbach's alpha was 0.95 for WOMAC test 1 and 0.96 for test 2.

The NHP is a health-related quality of life instrument covering areas of emotional reaction (9 statements), sleep disturbance (5 statements), lack of energy (3 statements), pain (8 statements), physical mobility (8 statements), and social isolation (5 statements) (15). Each item score is weighted, thus enabling the sum score in each subscale to range from 0 to 100, best to worst. In this study, Cronbach's alpha was 0.71 for test 1 and 0.85 for test 2.

The hip school consisted of an individual start up session, 3 group meetings (6–7 patients per group) lasting 1 hour each, and an individual followup session 2 months later, all led by a physical therapist with 20 hours of education in the method. The corpus of information was developed from literature reviews, interviews with physicians (20) and patients, and many years of reflection in action (21) (Appendix).

All the tests were conducted by 1 specially trained physical therapist that knew which group the patients belonged to but was not involved in the intervention.

All patients gave their informed consent before testing and the Medical Research Ethics Committee, Örebro, Sweden, approved the study.


For group comparisons, Student's t-tests were used for age and body mass index, chi-square tests for the categorical background variables, and Mann-Whitney U tests for the WOMAC and NHP at test 1 and 2, respectively. For change over time, difference variables were constructed by subtracting the scores from test 2 with the scores from test 1; and comparison between the T group and C group was conducted with the Mann-Whitney U test. A second post hoc analysis of the change over time was performed within the 2 groups separately using the Wilcoxon's matched pairs signed rank test. Cronbach's alpha determined the internal consistency. Significance level was defined as P ≤ 0.05.



There were no differences between the dropouts in the T group and the C group on any variable tested, so the dropouts were excluded from further analyses. However, there were several differences between the dropouts in both groups (n = 17 + 9) and the participants in both groups (n = 77 + 68), indicating that the dropouts had more pain and activity limitations, and experienced lower health-related quality of life compared with the participants.

Group comparisons for preintervention test.

At test 1, the scores for stiffness, activity limitations, total WOMAC, and NHP-pain differed between the 2 groups, with the T group experiencing more problems (Table 3).

Table 3. Median and interquartile range for the subscales and totals in WOMAC and NHP*
 Treatment groupControl groupP
Test 1Test 2Test 2-1Test 1Test 2Test 2-1Test 1Test 2-1
  • *

    Median and interquartile ranges for test 1, test 2, and for the change score (test 2-1). All scores are given as 0–100, best to worst. Minus signs indicate improvement. Also reported are the P values when comparing the scores at test 1 and the change score (test 2-1) between the treatment group and the control group tested with Mann-Whitney U tests. WOMAC = Western Ontario and McMaster Universities Osteoarthritis Index; NHP = Nottingham Health Profile; WOMAC-P = WOMAC pain subscale; WOMAC-S = WOMAC stiffness subscale; WOMAC-A = WOMAC activity limitations; NHP-EM = NHP emotion reaction subscale; NHP-SL = NHP sleep disturbances subscale; NHP-EN = NHP energy subscale; NHP-P = NHP pain subscale; NHP-PM = NHP physical mobility subscale; NHP-SO = NHP social isolation subscale.

WOMAC-P30 (20)30 (25)−5 (15)30 (25)35 (30)5 (19)0.4090.032
WOMAC-S50 (38)50 (25)0 (19)38 (34)38 (38)0 (25)0.0200.068
WOMAC-A32 (27)25 (24)−1 (15)20 (29)25 (35)3 (18)0.0040.044
Total WOMAC33 (24)27 (23)−2 (16)23 (36)28 (35)1 (20)0.0090.030
NHP-EM0 (9)0 (8)0 (4)0 (9)0 (8)0 (0)0.9840.907
NHP-SL11 (33)11 (22)0 (11)11 (33)11 (31)0 (19)0.9770.816
NHP-EN0 (37)0 (24)0 (0)0 (0)0 (37)0 (0)0.1490.034
NHP-P39 (33)29 (31)0 (36)29 (38)29 (50)0 (21)0.0360.067
NHP-PM21 (25)11 (21)0 (11)20 (29)20 (35)0 (21)0.0670.083
NHP-SO0 (0)0 (0)0 (0)0 (0)0 (0)0 (0)0.5670.596
Total NHP16 (20)11 (17)−4 (14)13 (18)13 (20)−1 (10)0.1030.056

Change over time.

When the change over time scores were compared between the T group and C group using the computed difference variable for WOMAC, there were differences for pain, activity limitations, total WOMAC, and NHP (lack of) energy subscale in favor of the T group (Table 3).

Post hoc analysis revealed that there were no differences between the groups on any of the subscales or total scores of WOMAC or NHP at test 2 (Table 3).

In the second post hoc analysis when the change over time scores were analyzed by within group comparisons, there were no significant changes for the C group, but the T group saw reductions in activity limitations and total WOMAC scores (P ≤ 0.05, respectively). There was also reduction for the T group in NHP scores for sleeping disturbances (P ≤ 0.05), pain (P ≤ 0.001), physical mobility (P ≤ 0.01), and total (P ≤ 0.001).

When the T group was divided into 4 groups by percentiles of the initial total WOMAC score, the percentile with initially highest mean scores on pain, stiffness, activity limitations, and total WOMAC, respectively, showed more change after 6 months.

Medication, physical activity, and walking distance.

There were no statistical differences for the change scores for medication, physical activity level, satisfaction with physical activity, and walking distance between the T group and C group; nor were there any within-group changes except for walking distance. The C group reduced their walking distance over the 6-month period (P ≤ 0.01).

One-year followup.

At test 3, 1 year after the first test, the patients in the T group had not worsened their scores, compared with test 2, either in the subscales of WOMAC or in the subscales of NHP (Figure 1).

Figure 1.

Median (interquartile range) for total Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) and total Nottingham Health Profile (NHP) from test 1 and 2 for the treatment group (T-gr) and control group (C-gr). Scores from test 3 for the treatment group are also shown.


This study indicated that the hip school intervention reduced the patients' pain and activity limitations and improved their health-related quality of life. This is not the only study to consider the early disease stages (10, 22), but as far as we know, it is the first to study early contact and health education solely for this patient group and with the aim of enabling them to cope and improve or maintain physical ability.

To put the study results in context, certain methodologic issues must be considered. For primary outcome measures, we used a generic health-related quality of life instrument (the NHP) that evaluates pain and physical mobility in general terms as well as aspects of mental and social well being. We also used an OA-specific instrument (the WOMAC) that evaluates aspects of pain, stiffness, and activity limitations. Our study group had their main problems in pain and more physical aspects, enabling all the subscales in WOMAC and only the pain and physical mobility subscales in NHP to show change over time.

Lacking consensus about early diagnostic tests for hip dysfunction or clinical criteria for hip OA, an operational definition of hip dysfunction was made and used by the physicians who introduced the patients to the study.

The patients were assigned to the T group and the C group according to residential area. The main reasons for not assigning them at random were the practical problems of having many units involved, and the risk that patients in the T group and C group would communicate with each other in small urban districts. Thus, recruitment according to residential area seemed to be the only way to conduct the study, and the ethics committee approved.

We ended up with an initial incongruity between the T group and C group in that the former had more pain, activity limitations, and health-related quality of life problems. This could be a consequence of the methodology used, admitting a physicians' bias. This is because the assigning physician knew that he or she would be assigning patients either to the C group with a 6-month wait for physical therapy or to the T group with a newly introduced intervention. The physicians in the T group area may have been more willing to assign persons with severe hip pain. This would explain why we had 6 dropouts due to THR in the T group, and none in the C group.

The results could illustrate the tendency of measures to regress towards the mean. The subjects were, however, recruited to the study when they were seeing a physician, which they probably do when they experience deterioration. Six months after the hip school, the T group reported reduced pain, activity limitations, and health-related quality of life problems, whereas the C group reported increased problems. The between-group comparisons of the change scores (test 2–test 1) showed greater reduction for the T group for pain, activity limitations, and total hip problems. The second post hoc analysis of within-group comparisons showed reduction of activity limitations and total hip problems in WOMAC, and in 3 of the NHP's subscales, together with total NHP score for the T group. For the T group, scores from an additional 6 months followup show sustained scores compared with test 2, for all subscales and total scores in both WOMAC and NHP. All these facts strengthen the conclusion that there was a real intervention effect, not simply regression toward the mean. However, on the basis of the results from this pilot study, alternative explanations cannot be excluded.

Finally, the hip school seeks to affect behavior, thoughts, and feelings. The degree of adherence is not known. Hip problems pose no immediate risks or threats, but require lifestyle changes. This might give high levels of nonadherence (23). On the other hand, if the proposed behavior changes make sense and require little time or effort, this might facilitate adherence. Earlier studies show weak association between changes in behavior and improved health (24). Instead, psychological theories have been used to explanation improved health (25). Different emotions have been associated with specific behavioral tendencies (26). According to clinical experience, sadness and fear are the most common negative emotions inducing inactivity and avoidance, respectively (26). Both these behavioral tendencies are dealt with in the hip school. Reduction of fear so as to change “illness behavior to wellness behavior” has been discussed earlier in a physical therapy perspective (27). The simple mechanical wear and tear model can reinforce inactivity and avoidance (28), further contributing to progression of disability. In educational interventions it is difficult to make a distinction between true effects and placebo effects (29, 30). Some of the effects of the hip school might be due to a change from 1 thought pattern with nocebo effects to another with powerful placebo effects. The trend seen in the scores from test 3 for the T group gives reason to believe there are effects that are meaningful for the patients.

These results must be regarded as preliminary, and further randomized studies are desirable. Both quantitative and qualitative studies are needed to further develop and evaluate the hip school.


We thank Professor Nicholas Bellamy for his generosity regarding information and the use of WOMAC; physical therapist Eva Högman Harvig for help with data collection; Professor Stefan Lohmander, Professor Lennart Bodin, and Professor Marion Minor for consultation; and Ann Enander and Tim Crosfield for reviewing the manuscript.



Persons with hip dysfunction, whether or not with known osteoarthritis (OA) of the hip are the targets of the hip school. Hip dysfunction is a multifactorial process, influenced by disease or joint failure but also by inactivity and aging.1

Illustration 1.

The “Hip School” is led by physical therapists educated in the method. At the individual start up session interview, specific combined hip joint range of motion (ROM) and simple balance tests are used to examine the patient's current condition and suggest possible ways for improvement.

At the first group meeting, the patients are encouraged to ask all the questions they have about hip problems and then slides with texts, photos, and illustrations are shown to help answer these questions and cover the areas in Table 1. Written therapeutic exercises for preserving/enhancing hip ROM are distributed. The patients are also offered an individual followup session after 2 months.

Table 1. Content summary of the 3 group meetings*
First group meetingSecond group meetingThird group meeting
  • *

    ROM = range of motion; OA = osteoarthritis.

Where is the hip?Muscles involvedPain
Tissues belonging to a jointDiagnosing decreased ROMSelf management of pain
Diagnosing hip OAProposed physical activityPros and cons of  pain treatments
Who gets hip OA?Not too much and not too little Physical theray
Hip OA and painTo preserve/enhance ROM Pharmacology
Natural course at group levelTherapeutic exercise sheet Surgery

Key messages in the hip school

There is hope! Hip dysfunction can improve or at least does not always get worse. The majority have less pain over time (31).

Talking to the person about signs and symptoms associated with movement and performing simple examination of impairments is the way to diagnose hip dysfunction. As long as the treatment strategies do not depend on whether OA changes exist, diagnosing hip dysfunction is enough. Plain radiography is a rough method, but still the best for diagnosing hip OA. The tissues surrounding the joint report the signs and symptoms. This is why people with radiologic hip OA can have the same signs and symptoms as those without.

The treatment strategy for persons with hip dysfunction should start with nonpharmacologic interventions, with education (32) aiming at empowering patients to manage their pain relief themselves, prevent or reduce impairment, and improve or maintain physical ability.

For the joint structures, weight bearing is positive and necessary. Try to find the optimal loading every day, i.e., not too much and not too little (33). More pain the day after activity should be avoided: it tells you that you have done too much (34).

Both for your hip and your general health, try to find ways to stay fit, in spite of pain, with at least 30 minutes accumulated physical activity a day. Bicycling, swimming, and walking are examples (35, 36). Try shock-absorbing materials in shoe soles or insoles (37). Keep up your hip ROM, you need it for daily living (38), and practice balance training (39, 40). Decreasing/preventing hip impairment is beneficial for all persons with hip dysfunction—for those whose pain is getting worse, and for those with a more favorable course. Exercise is integral in reducing secondary disability, modifying possible risk factors in disease progression, and improving general health (41).

Seek help before the problems overwhelm you. You can do things yourself to affect signs and symptoms from the hip and you can use the health care service (42). It's not all a mystery. There are explanations for many if not all of the experienced signs and symptoms.

Patient-related instructions including education are important because they help the person to make decisions regarding his/her health and the use of health care resources.

Don't be afraid of the pain (43). For those whose pain is getting worse, total hip replacement (THR) is a good possibility (44). After THR you still have the same surrounding tissues, so try to keep up your physical ability in spite of pain.