Measuring the population impact of knee pain and disability with the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC)

Pain 100 (2002) 55–64 www.elsevier.com/locate/pain

Measuring the population impact of knee pain and disability with the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) Clare Jinks*, Kelvin Jordan, Peter Croft Primary Care Sciences Research Centre, Hornbeam Building, Keele University, Keele, Staffordshire ST5 5BG, UK Received 21 February 2002; accepted 21 June 2002

Abstract This study has used the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) in an unsolicited postal questionnaire to investigate the impact of knee pain and disability in the general older population. The study provides WOMAC population data for those aged over 50 and demographic and psychosocial associations with severity of WOMAC scores. A pilot survey (n ¼ 240) and repeatability study (n ¼ 80) were undertaken to test completion of the WOMAC in this new setting. The main questionnaire was mailed to 8995 men and women aged over 50 registered with three general practices in North Staffordshire, UK. Completion rates for WOMAC items were high. Substantial reliability was found for pain and physical function scales (both .0.80). Fourteen percent of the over 50 population in this study had severe knee pain, 20% had severe difficulty with at least one area of physical functioning, 12% had both. The strongest link with severe difficulty with physical functioning was chronicity (odds ratio (OR) ¼ 6.49, 95% CI 4.65, 9.04). Other independent links were age over 75 years (odds ratio (OR) ¼ 4.11, 95% confidence interval (CI) 3.03, 5.58), depression (OR ¼ 2.80, 95% CI 2.22, 3.54), bilateral knee injury (OR ¼ 2.23, 95% CI 1.63, 3.06) and body mass index . 30 (OR ¼ 2.00, 95% CI 1.51, 2.64). Similar associations were found for severe pain. The findings suggest that the WOMAC is a reliable measure for use in postal surveys. It has advantages over other instruments when measuring pain and physical function difficulty related to the knee. Chronicity, older age, injury, obesity and depression were all linked with higher WOMAC scores for knee pain severity and disability among knee pain sufferers in the general older population. q 2002 International Association for the Study of Pain. Published by Elsevier Science B.V. All rights reserved. Keywords: Western Ontario and McMaster Universities Osteoarthritis Index; Knest; Knee pain; Disability; Postal survey; Epidemiology

1. Introduction Osteoarthritis is a term used to describe both a degenerative disease of synovial joints with its accompanying radiographic signs, and a clinical syndrome of pain, stiffness and restricted movement of the joints. In general, the clinical syndrome becomes more common with increasing age and in older adults, it is assumed to be related to the degenerative disease. The knee is the most common joint affected by osteoarthritis (OA) (Graham et al., 1995; Urwin et al., 1998), and knee pain has been found to be more important than radiographic severity in determining disability in those with clinical knee OA (Jordan et al., 1996; Odding et al., 1998; McAlindon et al., 1993). The goal of contemporary management of knee osteoarthritis is, therefore, control of pain and improvement in function and health related quality

* Corresponding author. Tel.: 144-1782-583926; fax: 144-1782583911. E-mail address: [email protected] (C. Jinks).

of life (Felson et al., 2000a). This assumes growing public health importance with an increasing proportion of older people in the population (Greengross et al., 1997; Pendleton et al., 2000; Felson et al., 2000b). Previous population studies of knee pain have used the Short Form 36 (SF36), Health Assessment Questionnaire (HAQ/mHAQ) and Lequesne Index to assess physical function (McAlindon et al., 1992; Tennant et al., 1995; Urwin et al., 1998; O’Reilly et al., 1998a). Agreement on the best measure to use is, however, lacking (Symmons, 2001). One measure of knee pain and disability not used in general population postal surveys of knee pain is the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) (Bellamy, 1996). This disease-specific, self-administered instrument was developed for use as an outcome measure in trials of treatments of patients with hip and knee osteoarthritis. The WOMAC contains 24 questions (five pain, two stiffness and 17 difficulty with physical functions), is available in a Likert or Visual analogue scale and can be completed in less than 5 min. It has been widely tested in surgical or hospital-

0304-3959/02/$20.00 q 2002 International Association for the Study of Pain. Published by Elsevier Science B.V. All rights reserved. PII: S 0304-395 9(02)00239-7

56

C. Jinks et al. / Pain 100 (2002) 55–64

based populations (McConnell et al., 2001) and extensively used in clinical trials because of its sensitivity to change and construct validity (Wolfe and Kong, 1999). Use of the WOMAC in a general population study may offer advantages. Firstly, it is disease-specific and targeted to ask about pain and difficulty related to the knee. This is true also for the Lequesne index; however, this measure was developed for interview administration and the originators highlight disappointing self-report findings (Lequesne et al., 1998). Generic and disease-specific instruments measure different dimensions of health experience, the use of both in quality of life studies has been recommended (Bombardier et al., 1995; Brazier et al., 1999). The item contents of the WOMAC offer four advantages. Firstly, the WOMAC was developed for hip and knee OA studies, and does not contain items about upper limb function. The lack of relevance of these questions for studies of knee pain and lower limb OA has been noted (Brazier et al., 1996; Griffiths et al., 1995; Hart and Spector, 1995). Second is the advantage of length, the WOMAC contains only 24 questions. The HAQ disability index, for example, contains 20 questions and a list of 20 aids and devices. The issue of length is important in a postal survey to enhance response rates. Thirdly, the WOMAC items are reported in three subscales (pain, stiffness and physical function) separately. This contrasts to the HAQ/mHAQ and Lequesne, which report a single score. Single scores have the disadvantage of aggregation. They ‘lead to loss of information as the same score can be obtained from many different combinations of the subdomains of the scale’ (Bowling, 1997). Fourthly, by using the WOMAC, it is possible to gather additional detail on restriction of specific tasks and the ‘real life’ impact of knee pain and related disability, and at differing levels of severity. Finally, if used in the general older population, normative WOMAC data will be available. These data will provide an important context for hospital-based studies and trials that use the WOMAC to assess treatment outcomes in knee related samples. We have carried out a survey in the general population (aged over 50) as the first phase of a prospective study of knee pain in this age group. In this paper, our objective is to describe the prevalence and severity of knee pain and its associated disability in older adults using the WOMAC index and to report on the reliability of the WOMAC in this setting. In addition, we have investigated the extent to which the severity of knee pain and disability as measured by the WOMAC is linked to other pain status and to demographic, socioeconomic and psychological status.

2. Methods 2.1. Format of the full questionnaire The development of the questionnaire used in our survey

is detailed elsewhere (Jinks et al., 2001). It contained a Knee Pain Screening Tool (KNEST), a manikin for shading body pain, the SF36 (Ware et al., 1993), the Hospital and Anxiety Depression Scale (HADS) (Zigmond and Snaith, 1983) and demographic questions. Deprivation status was calculated using Townsend Deprivation Index (Townsend et al., 1988), a small area neighborhood deprivation measure derived from the national census and applied to individuals according to their residential code. The score is calculated from four variables: houses not occupied by the owner, households with no car, households with more than one person per room (overcrowding) and number of persons who were unemployed. An area with a positive score is considered deprived, a negative score more affluent. Those who responded positively to the KNEST screening question ‘have you had pain in the last year in or around the knee?’ were asked to complete the WOMAC (Likert version). The original WOMAC asks respondents to consider pain and difficulty in the last 48 h because of arthritis. For the purposes of this study, the word ‘arthritis’ was not used. Instead, respondents were asked to think about ‘the amount of pain that you have experienced in your knee’ in the last 48 h. This minor change made the questionnaire relevant to people in the general population who had not consulted their General Practitioner (GP) or been given a diagnosis of OA. There is no agreed cut off to define severity of pain or disability in the WOMAC literature (O’Reilly et al., 1998b). We have arbitrarily dichotomised the responders who scored ‘severe’ or ‘extreme’ on at least one item on the pain and physical function scales into ‘severe’ pain and ‘severe’ physical function groups respectively. For the purposes of analysis, this generates reasonable distinct groups who share perceptions that some aspect of their lives are severely restricted by pain or disability. The remainder of responders were described as ‘non-severe pain’ or ‘non-severe physical function’. 2.2. Pilot and repeatability studies A pilot study was carried out to assess response/completion rates and to establish reliability in this setting. The questionnaire was sent to a random sample of 240 older individuals registered with two general practices. Another identical questionnaire was sent 2 weeks later to a random subsample (n ¼ 80) of responders to test repeatability of the instrument. 2.3. Main study The main questionnaire was mailed to all adults aged over 50 years (n ¼ 8995) registered with three different general practices. The practices were in a mix of urban and rural locations, and represented a broad range of socioeconomic status. After 2 weeks, a postcard reminder was sent to non-responders. After a further 2 weeks, a second

C. Jinks et al. / Pain 100 (2002) 55–64

57

Table 1 Test–retest reliability of the WOMAC in the general adult population (aged over 50) with self-reported knee pain Scale

No. of cases

Mean difference a (95% CI)

ICC (2,1) b (one sided 95% CI)

Pain Stiffness Physical function

18 21 21

0.71 (20.40 to 1.81) 2 0.14 (20.84 to 0.55) 2.41 (21.47 to 6.3)

0.88 (r . 0.74) 0.69 (r . 0.44) 0.85 (r . 0.71)

a b

Baseline-retest at 2 weeks. ICC based on two-way random effects ANOVA: ICC ranges from 0 to 1 (perfect reliability).

questionnaire and reminder letter were sent to non-responders. The local Research Ethics Committee approved the survey.

2.4. Statistical analysis Test–retest reliability for the WOMAC in the pilot study was assessed using the intraclass correlation coefficient (ICC (2,1)) (Shrout and Fleiss, 1979). Internal reliability of the WOMAC subscales in the main survey was measured by Cronbach’s alpha (Bland and Altman, 1997). Body mass index (BMI) was calculated from self-reported weight and height (BMI ¼ weight (kg)/[height (m)] 2). Widespread pain was defined as pain shaded on the pain manikin in the axial skeleton or lower back and in at least two areas of two contralateral limbs. This definition has been devised, validated and used by Macfarlane et al. (1996) and requires the presence of more diffuse limb pain. Item responses on the WOMAC were summed to produce subscale scores (pain 0– 20, stiffness 0–8, physical function 0–68) with higher scores indicating worst health. Recommended guidelines for dealing with missing data were followed (Bellamy, 1996). The HADS was trichotomised (least, moderate and most anxiety and depression). Those in the top tertile score were considered to be ‘more’ anxious and depressed and the bottom two tertiles were combined into a ‘less’ anxious or depressed group. Univariate analyses were performed to assess relationships between pain, psychosocial and demographic factors and WOMAC scores. Associations were assessed using unpaired t-tests or analysis of variance (ANOVA) and Pearson’s correlation coefficients. Multiple logistic regression analysis was performed to obtain odds ratios for associations between pain, psychosocial and demographic factors and the category of severe WOMAC pain and difficulty with physical function. Variables included in the model were those identified as being associated with WOMAC subscales in previous hospital studies or those found to be significantly associated (P , 0:05) with WOMAC scales in the univariate analyses. Independent variables were entered jointly into the model. Statistical analysis was performed using SPSS 10.0 (SPSS inc). Prevalence estimates were standardised by age and gender to the whole target population using direct standardisation.

3. Results of the pilot study 3.1. Response/completion rates and reliability An 85% (187/198) baseline response rate was achieved for the first pilot questionnaire. Of these, 45% (84/187) reported knee pain in the last 12 months and proceeded to complete the WOMAC. Completion rates were high for all WOMAC items ranging from 88.1 to 100%. Subscale scores could be calculated for 89% (pain), 99% (stiffness) and 95% (physical function) of responders. A random sample of 80 responders (irrespective of knee pain status) was mailed the repeatability questionnaire. The response rate was 74% (n ¼ 59). Of these, 26 (44%, mean age 68.7, SD 10.02, 16 females) reported knee pain. For each scale, 24 out of the 26 questionnaires could be scored. Table 1 shows test–retest reliability of the WOMAC subscales. Substantial reliability was found for the pain and physical functions scales (both above 0.80) based on Shrout’s classifications (Shrout, 1998). The stiffness scale shows moderate reliability. 4. Results of main study 4.1. Response and completion rates Six thousand seven hundred and ninety two individuals returned the main study questionnaire. After accounting for wrong address or removals, the adjusted response rate was 77%. Respondents had a mean age of 65.4 years (sd 10.10, range 50–100), 56% were female, 99.6% white UK/ European. Females and those aged between 65 and 74 years had slightly higher response rates. The 12-month period prevalence of pain ‘in or around the knee’ was 47% (male 44%, female 49%, n ¼ 3023). Completion scores for WOMAC items for these responders were high with all items at least 94% completed. Subscale scores could be calculated for 95% of subjects for both the pain and physical functioning scales and 97% of subjects for the stiffness scale. 4.2. Pain and disability scores and prevalence Distributions of scores are displayed in Fig. 1. Severity scores as measured by WOMAC subscales are presented in

58

C. Jinks et al. / Pain 100 (2002) 55–64

Fig. 1. Distribution of WOMAC subscale scores in the general population (aged over 50) with self-reported knee pain.

Table 2 classified by age and gender. Mean scores tended to be highest (worse) for females and for the older age groups. The proportions of responders who scored at the floor (zero) on each WOMAC scale were 8.9% pain, 18.3% stiffness and 7.7% physical function (the lower the score on the WOMAC, the less severe the problem). Conversely, the proportions of responders who scored at the ceiling were 0.6% pain, 1.8% stiffness and 0.4% physical function. 4.3. Internal validity of WOMAC scales Cronbach’s Alpha for WOMAC subscales were 0.92 pain, 0.90 stiffness and 0.98 physical function.

4.4. Population impact Going up or down stairs was the activity causing most pain with 23% of responders with knee pain reporting severe or extreme pain with this task. The prevalence of other pain items was approximately 1 in 10. The greatest degree of difficulty in physical functioning was reported for heavy domestic duties (27% of responders with knee pain reporting severe or extreme difficulty) followed by bending (26%) and getting in or out of the bath (21%). Overall, 31% of responders with knee pain were classified as having severe pain, 45% had severe difficulty with physical function, and 28% had both. When extrapolated to the general practice popula-

Table 2 Normative WOMAC Scores in a general older population (aged over 50) with self-reported knee pain

Male

50–64

65–74

75 1

Female

50–64

65–74

Total

a b c

Possible range 0–20. Possible range 0–8. Possible range 0–68.

Pain a

Stiffness b

Physical function c

Mean (sd) Median n Mean (sd) Median n Mean (sd) Median n

5.5 (4.34) 5.00 663 6.47 (4.49) 6.00 362 7.16 (4.26) 7.00 176

2.47 (1.97) 2.00 678 3.09 (2.03) 3.00 367 3.08 (1.89) 4.00 179

17.62 (15.4) 15.00 675 22.63 (15.35) 22.00 359 25.9 (15.2) 27.1 172

Mean (sd) Median n Mean (sd) Median n

6.05 (4.52) 5.00 800 7.21 (4.49) 7.00 522

2.65 (2.06) 2.00 813 3.12 (1.98) 3.00 535

18.76 (16.03) 14.73 797 24.42 (15.57) 24.00 518

Mean (sd) Median n

6.57 (4.52) 6.00 2876

2.92 (2.04) 3.00 2947

22.21 (16.38) 20.00 2870

C. Jinks et al. / Pain 100 (2002) 55–64

59

Table 3 WOMAC scores for men and women (aged over 50) with self-reported knee pain in relation to other pain factors (n ¼ 3023) a Percentage

Chronic knee pain $ 3 months , 3 months

54 1271

Pain*

Stiffness*

Physical function*

n

Mean

s.d.

n

Mean

s.d.

n

Mean

s.d.

1528 4.22

8.61 3.71

4.17 1301

1550 1.95

3.76 1.72

1.92 1284

1497 13.70

29.65 13.18

15.32

Bilateral pain Yes No

52

1458 1356

7.39 5.71

4.62 4.25

1482 1386

3.36 2.45

2.03 1.95

1445 1350

25.65 18.53

16.71 15.27

Hip pain Yes No

34

980 1896

7.80 5.49

4.62 4.34

1002 1945

3.43 2.66

2.04 1.98

974 1896

26.82 19.84

16.44 15.84

Lower back pain Yes No

45

1288 1588

7.34 5.94

4.66 4.31

1316 1691

3.28 2.63

2.03 1.99

1278 1592

25.36 19.68

16.78 15.61

Neck pain Yes No

29

832 2044

7.77 6.08

4.62 4.39

851 2096

3.48 2.69

2.02 2.00

829 2041

26.96 20.28

16.92 15.76

Hand pain Yes No

27

794 2082

7.82 6.09

4.60 4.40

811 2136

3.55 2.68

2.02 1.99

785 2085

27.71 20.14

16.66 15.79

Foot and ankle pain Yes No

28

804 2072

8.14 5.96

4.45 4.04

822 2125

3.66 2.63

1.98 1.98

795 2075

28.36 19.85

16.19 15.84

Widespread pain Yes No

17

503 2373

9.06 6.04

4.57 4.34

514 2433

4.02 2.68

1.97 1.97

496 2374

31.86 20.19

16.70 15.58

a

All differences between those with and without pain are significant, P , 0:001:

tion (standardised by age and gender), the community burden of self-reported knee pain can be estimated. Of the over 50 population in this study, 14% (95% CI 13.0%, 14.8%) had severe pain, 20% (95% CI 19.4%, 21.6%) had severe difficulty with physical functioning and 12% (95% CI 11.6%, 13.3%) had both. With reference to the use of health care, 49% of responders with severe pain or difficulty with physical functioning had not consulted their GP in the last 12 months and 36% reported using no treatments or healthcare services in the last 12 months. Standardising to the whole sample population, 8% of this older population has severe knee pain or difficulty with physical function for which they do not consult their GP or use treatments or other services in 1 year. 4.5. Characteristics associated with WOMAC scores: univariate analyses WOMAC scores are shown in Tables 3 and 4 in relation to pain, demographic and psychosocial factors. Chronicity (knee pain $ 3 months), bilaterality, regional pain (hip, lower back, neck, hand, foot and ankle) and widespread

pain were all associated with significantly higher WOMAC scores. There were clear significant trends of rising severity across all three domains with increasing age and HADS scores. 4.6. Characteristics associated with severe pain and difficulty with functioning: categorical analyses We identified the subgroup with severe difficulty with physical function (Table 5) and severe knee pain (Table 6). We examined associations between these and widespread pain and psychosocial and sociodemographic variables, by comparing the severe groups with the rest of the knee pain sufferers. In the unadjusted analyses, severe pain and severe difficulty with physical function were each associated with age, gender, deprivation, obesity, anxiety and depression, living alone, lower education level, knee injury and widespread pain. Knee pain of 3 months or more was the most strongly associated characteristic. Multiple regression analysis was performed with each multivariate model set at 95% confidence interval with all covariates entered jointly. WOMAC severe pain and physical function categories were dependent variables in

60

C. Jinks et al. / Pain 100 (2002) 55–64

Table 4 WOMAC scores for men and women aged over 50 with self-reported knee pain in relation to sociodemographic and psychosocial factors a Percentage

Pain*

Stiffness*

Physical Function*

n

Mean

s.d.

n

mean

s.d.

n

Mean

s.d.

Female Male

59

1675 1201

6.95 6.04

4.56 4.41

1723 1224

3.04 2.75

2.06 1.99

1664 1206

23.60 20.29

16.75 15.67

Age: 50–64 65–74 751

50 31 19

1376 820 466

5.67 6.86 8.10

4.43 4.44 4.34

1376 820 466

2.54 3.08 3.62

2.01 3.98 1.94

1376 820 466

18.13 23.60 31.40

15.78 15.62 15.92

Normal weight (BMI .20–25) Underweight (BMI ,20 Overweight (BMI .25–30) Obese (BMI .30)

34 4 41 21

883 90 1114 575

5.68 7.71 6.41 7.57

4.33 5.21 4.38 4.65

883 90 1114 575

2.51 3.33 2.87 3.45

1.96 2.22 1.99 2.08

883 90 1114 575

18.79 29.70 21.96 26.44

16.04 18.40 15.95 16.59

Affluent area Deprived area

28

808 2068

5.20 7.10

4.02 4.60

824 2123

2.44 3.10

1.95 2.04

806 2064

17.45 24.07

15.11 16.49

9

259 2505

5.35 6.64

4.76 4.48

269 2556

2.34 2.96

2.13 2.02

263 2496

17.00 22.60

16.86 16.29

Cohabiting Not Cohabiting

68

1950 909

6.24 7.18

4.50 4.43

1992 934

2.77 3.19

2.01 2.03

1955 896

20.42 25.84

15.91 16.60

Previous injury No previous injury

45

1256 1516

7.17 5.95

4.53 4.41

1281 1552

3.21 2.63

2.02 2.00

1237 1527

24.57 19.83

16.27 16.14

Anxiety b Most Moderate Least

36 36 29

989 1014 795

8.16 6.23 5.02

4.56 4.29 4.09

1017 1032 819

3.53 2.87 2.21

2.05 1.96 1.84

981 1013 804

28.34 21.26 15.83

16.51 15.51 14.47

Depression b Most Moderate Least

33 37 30

918 1026 858

8.94 6.31 4.36

4.61 4.04 3.65

947 1056 874

3.87 2.88 1.94

2.04 1.87 1.70

903 1030 868

32.50 21.28 12.62

16.27 14.20 12.24

Further education No further education

a b

*All scores are significant at P , 0:001. Hospital Anxiety and Depression Scale (Zigmond and Snaith, 1983) based on tertiles.

separate multiple logistic regressions. After adjustment for the presence of other variables, the strongest link was with persistence of pain (OR for those with pain for 3 months or more ¼ 5.69, 95% CI 3.84, 8.42 for pain, OR ¼ 6.49, 95% CI 4.65, 9.04 for physical function). Other main links, after adjustment, were older age, (OR for those over 75 years ¼ 2.20, 95% CI 1.64, 2.95 for pain, OR ¼ 4.11, 95% CI 3.03, 5.58 for physical function), depression (OR for those most depressed ¼ 2.30, 95% CI 1.83, 2.89 for pain, OR ¼ 2.80, 95% CI 2.22, 3.54 for physical function for physical function), bilateral injury (OR ¼ 1.41, 95% CI 1.13, 1.77 for pain, OR ¼ 2.23, 95% CI 1.63, 3.06 for physical function) and BMI $ 30 (OR ¼ 1.78, 95% CI 1.35, 2.34 for pain, OR ¼ 2.00, 95% CI 1.51, 2.64 for physical function). Weaker but significant associations persisted for widespread pain, anxiety, deprivation score and gender.

5. Discussion We have used the WOMAC, a self-administered instrument already established for measuring pain and physical function in osteoarthritis, to investigate the impact and associated features of knee pain in the older general population. We expanded the application of this tool to knee pain rather than diagnosed OA, as there are many older people with knee pain who do not consult their GP and get a diagnosis (Jinks et al., 2001) but who have difficulty with activities of daily life. This broad group is important in identifying potential targets for prevention of progression of pain and disability and for effective treatments and strategies of care. This may also be relevant to prevention of progression of radiographic OA as epidemiological studies have shown that baseline knee pain predicts subsequent radiographic OA (Hart et al., 1999).

C. Jinks et al. / Pain 100 (2002) 55–64

61

Table 5 Severe difficulty a with physical function by demographic and psychosocial factors Severe difficulty

Not severe difficulty

OR b (95% CI)

OR c (95% CI)

Male Female

489 (40%) 835 (49%)

732 875

1.00 1.43 (1.23, 1.66)

1.00 1.30 (1.06, 1.61)

Age: 50–64 65–74 75 1

518 (35%) 442 (49%) 364 (67%)

972 454 181

1.00 1.83 (1.54, 2.16) 3.77 (3.07, 4.65)

1.00 1.83 (1.46, 2.30) 4.11 (3.03, 5.58)

Normal weight (BMI .20–25) Underweight (BMI ,20) Overweight (BMI .25–30) Obese (BMI .30)

341 (36%) 57 (59%) 511 (44%) 338 (56%)

594 39 648 261

1.00 2.55 (1.66, 3.91) 1.37 (1.15, 1.64) 2.26 (1.83, 2.78)

1.00 1.43 (0.81, 2.52) 1.33 (1.05, 1.67) 2.00 (1.51, 2.64)

Cohabiting Not cohabiting

807 (41%) 504 (54%)

1179 421

1.00 1.75 (1.49, 2.05)

1.00 1.08 (0.85, 1.35)

Further education No further education

93 (35%) 1163 (46%)

174 1380

1.00 1.58 (1.21, 2.05)

1.00 1.10 (0.77, 1.57)

Chronicity Less 7 days 1–4 weeks 1 month ,3 months $ 3 months

73 (18%) 76 (18%) 143 (32%) 989 (64%)

338 356 308 552

1.00 0.99 (0.69, 1.41) 2.15 (1.56, 2.97) 8.30 (6.31, 10.91)

1.00 0.94 (0.61, 1.43) 1.97 (1.34, 2.90) 6.49 (4.65, 9.04)

Affluent area Deprived area

282 (34%) 1042 (49%)

537 1070

1.00 1.85 (1.57, 2.19)

1.00 1.52 (1.21, 1.91)

Less anxious Most anxious

682 (37%) 599 (59%)

1162 409

1.00 2.50 (2.13, 2.92)

1.00 1.63 (1.29, 2.05)

Less depressed Most depressed

630 (33%) 656 (70%)

1288 283

1.00 4.74 (4.00, 5.61)

1.00 2.80 (2.22, 3.54)

Laterality Unilateral Bilateral

512 (37%) 773 (52%)

865 702

1.00 1.86 (1.60, 2.16)

1.00 1.17 (0.94, 1.46)

Not widespread pain Widespread pain

997 (41%) 327 (64%)

1424 183

1.00 2.55 (2.09, 3.11)

1.00 1.68 (1.28, 2.20)

No knee injury ever Knee injury ever – one knee Knee injury ever – both knees

591 (38%) 428 (47%) 233 (63%)

954 474 138

1.00 1.46 (1.23, 1.72) 2.73 (2.16, 3.45)

1.00 1.48 (1.18, 1.86) 2.23 (1.63, 3.06)

a b c

Severe or extreme difficulty on at least one WOMAC PF item. Unadjusted. Adjusted for the other variables.

We used a sample of all those aged over 50 registered with three general practices. Approximately 98% of the British population is registered with a GP (Bowling, 1997) and the register provides a convenient sampling frame of a local population. As our survey achieved a high response rate (77%), the effects of non-response bias should be limited, although there were some differences between responders and non-responders in terms of age, gender and general practice registered. The mean BMI score in our population sample was 26.2, in line with national trends (Department of Health, 1999). However, 99% of responders were caucasian (white UK/European origin), reflecting the nature of our local population and our results are only generalisable to this ethnic group.

Demographic characteristics of the sample were compared to the population of North Staffordshire and England and Wales (based on population estimates mid-2000 (ONS, 2001). In the current study, 44% of responders were males and 56% females, compared to 46% (male) and 54% (female) of over 50s, respectively, in North Staffordshire and England and Wales. The proportion of people aged over 75 1 in the survey sample was 21%, compared to 22% in North Staffordshire and 23% in England and Wales. The Index of Deprivation (DETR, 2000) shows that the general practices used are in a range of socioeconomic areas, and we adjusted for Townsend deprivation scores in the analysis. The only main restriction on generalisability is the small proportion of ethnic subgroups compared to some areas of Britain.

62

C. Jinks et al. / Pain 100 (2002) 55–64

Table 6 Severe knee pain a by demographic, psychosocial and pain factors Severe pain a

Not severe pain

OR b(95% CI)

OR c (95% CI)

Male Female

324 (27%) 578 (34%)

881 1104

1.00 1.42 (1.21, 1.67)

1.00 1.28 (1.03, 1.58)

Age 50–64 65–74 75 1

376 (26%) 292 (33%) 234 (44%)

1091 594 300

1.00 1.43 (1.19, 1.71) 2.26 (1.84, 2.79)

1.00 1.44 (1.14, 1.81) 2.20 (1.64, 2.95)

Normal weight (BMI .20–25) Underweight (BMI ,20) Overweight (BMI .25–30) Obese (BMI .30)

224 (24%) 35 (36%) 348 (30%) 245 (42%)

698 62 802 339

1.00 1.76 (1.13, 2.73) 1.35 (1.11, 1.65) 2.25 (1.80, 2.81)

1.00 1.07 (0.61, 1.88) 1.22 (0.96, 1.55) 1.78 (1.35, 2.34)

Cohabiting Not cohabiting

557 (29%) 333 (37%)

1400 580

1.00 1.44 (1.22, 1.71)

1.00 1.02 (0.81, 1.28)

Further education No further education

73 (28%) 790 (31%)

188 1723

1.00 1.18 (0.89, 1.57)

1.00 0.78 (0.55, 1.12)

Chronicity Less 7 days 1 – 4 weeks 1 month ,3 months $ 3 months

43 (11%) 49 (12%) 78 (17%) 717 (47%)

361 372 369 819

1.00 1.11 (0.72, 1.71) 1.78 (1.19, 2.65) 7.35 (5.28, 10.24)

1.00 1.23 (0.75, 2.01) 1.66 (1.04, 2.64) 5.69 (3.84, 8.42)

Affluent area Deprived area

182 (22%) 720 (35%)

628 1357

1.00 1.83 (1.52, 2.21)

1.00 1.52 (1.20, 1.93)

Less anxious Most anxious

437 (24%) 440 (44%)

1379 553

1.00 2.51 (2.13, 2.96)

1.00 1.55 (1.24, 1.95)

Less depressed Most depressed

407 (22%) 472 (51%)

1483 451

1.00 3.81 (3.22, 4.52)

1.00 2.30 (1.83, 2.89)

Unilateral knee pain Bilateral knee pain

318 (23%) 570 (39%)

1043 893

1.00 2.09 (1.78, 2.47)

1.00 1.41 (1.13, 1.77)

No widespread pain Widespread pain

655 (28%) 247 (49%)

1727 258

1.00 2.52 (2.07, 3.07)

1.00 1.53 (1.19, 1.98)

No knee injury ever Knee injury – one knee Knee injury – both knees

401 (26%) 281 (31%) 171 (47%)

1119 614 197

1.00 1.28 (1.07, 1.53) 2.42 (1.92, 3.06)

1.00 1.33 (1.05, 1.68) 1.68 (1.25, 2.27)

a b c

Severe or extreme pain on at least one WOMAC pain item. Unadjusted. Adjusted for other variables.

This study aimed to investigate the impact of knee pain and disability in the general population rather than in patients with specific medical conditions. A forthcoming study will investigate the primary care medical records of responders and link diagnostic subgroups to the selfreported survey data.

5.1. Completion and response rates This study reported response and completion rates for all WOMAC items that were higher than those reported in a Swedish survey of people who had previously undergone knee arthroplasty (Dunbar et al., 2001). The high response and completion rates achieved in this study suggest that the WOMAC is an acceptable tool to use in community surveys.

5.2. Reliability Previous studies have investigated reliability of the WOMAC in clinical and hospital settings (McConnell et al., 2001). Cronbach alpha statistics in our study were higher than those reported by the developers and in other knee OA studies (Bellamy, 1996; Dunbar et al., 2001; Roos et al., 1999; Stucki et al., 1998). Our study showed test–retest scores, using the intraclass correlation coefficient, higher than those in patients with knee OA on a hospital physiotherapist waiting list (Fransen and Edmonds, 1999) and comparable to other clinical knee OA studies (Bellamy, 1996; Dunbar et al., 2001; Roos et al., 1999; Stucki et al., 1998). 5.3. WOMAC dimension scores We have presented the first WOMAC subscale data for a

C. Jinks et al. / Pain 100 (2002) 55–64

general older (aged over 50) population with knee pain. These data, although limited by the ethnic mix, provide a valuable context for assessing results from hospital studies and trials of knee OA that have used the WOMAC. As expected, mean scores are lower in our study than those reported from other studies where patients were awaiting knee arthroplasty or attending hospitals or clinics (Brazier et al., 1996; Bellamy et al., 1988a,b). Also, in this study, the number of people scoring at the floor (best health) is less than that reported in subjects who had previously undergone knee replacement surgery (Bombardier et al., 1995; Dunbar et al., 2001), but more than in a study of patients awaiting knee surgery or attending hospitals/ clinics (Brazier et al., 1996). The number of responders who scored at the ceiling (worst health) was small and consistent with hospital-based studies of knee OA (Bombardier et al., 1995; Brazier et al., 1996; Wolfe and Kong, 1999; Dunbar et al., 2001). The profile of difficulty with tasks reported in this study is also similar to that reported using the WOMAC (VAS) in hospital outpatients with knee OA (Creamer et al., 2000).

5.4. Associations with WOMAC scores in the general population The associations described here for knee pain severity in the population reflect those reported elsewhere using other scales or in studies of radiographic OA sufferers. Hence the links with older age and female gender are in agreement with previous studies that used the HAQ (McAlindon et al., 1992) and the SF36 (O’Reilly et al., 1998a). The findings in relation to depression are consistent with populationderived data using the HADs and WOMAC (O’Reilly et al., 1998b). We have confirmed strong links between obesity and self-reported knee pain and related disability using the WOMAC in the general older population. This link has been previously reported from WOMAC studies using hospitalbased samples of patients with symptomatic knee OA (Creamer et al., 1999, 2000). Obesity provides one example of the rationale in studying non-specific knee pain in the population. Progression of lower limb disability in later life in the general population is linked to problems such as knee pain rather than the severity of radiographic disease in a small subgroup (Jordan et al., 1996). To reduce the burden and impact of disability related to knee pain, health service planners may reasonably centre intervention schemes and prevention programmes on the symptom itself and associated disability. The WOMAC, when used in the general population can: (i) serve as a context for trials of treatments; (ii) set a baseline of how knee pain affects people’s own perceptions of disability; (iii) provide a basis for a broader evaluation of outcome of health care; (iv) place knee pain within a wider psychosocial framework and inform a more integrated approach to development of health and social care.

63

6. Conclusion The WOMAC is a reliable disease-specific measure for use in large epidemiological postal surveys. It is well completed and has good test–retest reliability in the general older adult population. Our study using the WOMAC has shown that one-fifth of the population aged over 50 has severe difficulty with daily tasks because of their knee, and has described the main independent factors associated with severe pain and difficulty with physical function as chronicity, age over 75 years, depression, knee injury and BMI . 30. The WOMAC has the advantages of being a concise knee specific instrument that can measure the severity of pain and disability. Such data can provide a context for trials and other knee related studies that have used or intend to use the WOMAC. Acknowledgements The authors acknowledge the West Midlands New Blood Research Fellowship Committee who awarded funding for this study and also the Haywood Rheumatism Research and Development Foundation, North Staffordshire who supported the survey through a project grant. Professor Bie Nio Ong jointly supervised the New Blood Fellowship. We thank the doctors, staff and patients of the three health centres concerned. We would also like to thank the administration team in Primary Care Sciences who helped with the survey. References Bellamy N, Buchanan WW, Goldsmith CH, Campbell J, Stitt LW. Validation study of WOMAC: a health status instrument for measuring clinically important patient relevant outcomes to antirheumatic drug therapy in patients with osteoarthritis of the hip or knee. J Rheumatol 1988a;15(12):1833–1840. Bellamy N, Buchanan WW, Goldsmith CH, Campbell J. Validation study of WOMAC: a health status instrument for measuring clinically important patient relevant outcomes following total hip or knee arthroplasty in osteoarthritis. J Orthop Rheumatol 1988b;1:95–108. Bellamy N. WOMAC osteoarthritis index. A user’s guide, London, Ontario: London Health Services Centre, McMaster University, 1996. Bland JM, Altman DG. Gronbach’s alpha. BMJ 1997;314(7080):572. Bombardier C, Melfi CA, Paul J, Green R, Hawker G, Wright J, Coyte P. Comparison of a generic and a disease-specific measure of pain and physical function after knee replacement surgery. Med Care 1995;33(4 Suppl):AS131–AS144. Bowling A. Research methods in health, Buckingham: Open University Press, 1997. Brazier J, Snaith ML, Munro J. Measuring health outcome in people with osteoarthritis of the knee., Sheffield: Sheffield Centre for Health and Related Research, 1996. Brazier JE, Harper R, Munro J, Walters SJ, Snaith ML. Generic and condition-specific outcome measures for people with osteoarthritis of the knee. Rheumatology (Oxford) 1999;38(9):870–877. Creamer P, Lethbridge-Cejku M, Hochberg MC. Determinants of pain severity in knee osteoarthritis: effect of demographic and psychosocial variables using 3 pain measures. J Rheumatol 1999;26(8):1785–1792.

64

C. Jinks et al. / Pain 100 (2002) 55–64

Creamer P, Lethbridge-Cejku M, Hochberg M. Factors associated with functional impairment in symptomatic knee OA. Rheumatology 2000;39:490–496. Department of Health. Health survey for England, London: HMSO, 1999. Department of the Environment, Transport and the Regions, Regeneration Research Summary: Indices of Deprivation 2000. Available from URL: http://www.neighbourhood.statistics.gov.uk/home.asp. Dunbar MJ, Robertsson O, Ryd L, Lidgren L. Appropriate questionnaires for knee arthroplasty. Results of a survey of 3600 patients from The Swedish Knee Arthroplasty Registry. J Bone Joint Surg Br 2001;83(3):339–344. Felson DT, Lawrence RC, Dieppe PA, Hirsch R, Helmick CG, Jordan JM, Kington RS, Lane NE, Nevitt MC, Zhang Y, Sowers M, McAlindon T, Spector TD, Poole AR, Yanovski SZ, Ateshian G, Sharma L, Buckwalter JA, Brandt KD, Fries JF Osteoarthritis: new insights. Ann Intern Med 2000a;133(9):726–737. Felson DT, Lawrence RC, Dieppe PA, Hirsch R, Helmick CG, Jordan JM, Kington RS, Lane NE, Nevitt MC, Zhang Y, Sowers M, McAlindon T, Spector TD, Poole AR, Yanovski SZ, Ateshian G, Sharma L, Buckwalter JA, Brandt KD, Fries JF. Osteoarthritis: new insights. Ann Intern Med 2000b;133(8):635–646. Fransen M, Edmonds J. Reliability and validity of the EuroQol in patients with osteoarthritis of the knee. Rheumatology 1999;38(9):807–813. Graham G, Fairclough JA, Klippel JH, Dieppe PA, editors. Practical rheumatology, vol. 9 (The Knee). London: Times Mirror International Publishers Limited, 1995. pp. 97–110. Greengross S, Murphy E, Quam L, Rochon P, Smith R. Aging: a subject that must be at the top of world agendas. Br Med J 1997;315(7115):1029– 1030. Griffiths G, Bellamy N, Bailey WH, Bailey SI, McLaren AC, Campbell J. A comparative study of the relative efficiency of the WOMAC, AIMS and HAQ instruments in evaluating the outcome of total knee arthroplasty. Inflammopharmacolgy 1995;3:1–6. Hart DJ, Spector TD. The classification and assessment of osteoarthritis. Baillieres Clin Rheumatol 1995;9(2):407–432. Hart DJ, Doyle DV, Spector TD. Incidence and risk factors for radiographic knee osteoarthritis in middle-aged women: the Chingford Study. Arthritis Rheum 1999;42(1):17–24. Jinks C, Lewis M, Ong BN, Croft P. A brief screening tool for knee pain in primary care. Rheumatology 2001;40(5):528–536. Jordan JM, Luta G, Renner JB, Linder GF, Dragomir A, Hochberg MC, Fryer JG. Self-reported functional status in osteoarthritis of the knee in a rural southern community: the role of sociodemographic factors, obesity, and knee pain. Arthritis Care Res 1996;9(4):273–278. Lequesne MG, Mery C, Samson M, Marty M. Comparison between the WOMAC and the Lequesne indices in patients with knee and hip osteoarthritis. Osteoarthritis Cartilage 1998;6(6):441–442. Macfarlane GJ, Croft PR, Schollum J, Silman AJ. Widespread pain: is an improved classification possible? J Rheumatol 1996;23(9):1628–1632. McAlindon TE, Cooper C, Kirwan JR, Dieppe PA. Knee pain and disability in the community. Br J Rheumatol 1992;31(3):189–192. McAlindon TE, Cooper C, Kirwan JR, Dieppe PA. Determinants of disability in osteoarthritis of the knee. Ann Rheum Dis 1993;52(4):258–262. McConnell S, Kolopack P, Davis A. The Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC): a review of its utility and measurement properties. Arthritis Care Res 2001;45(5):453–461. Odding E, Valkenburg HA, Algra D, Vandenouweland FA, Grobbee DE,

Hofman A. Associations of radiological osteoarthritis of the hip and knee with locomotor disability in the Rotterdam Study. Ann Rheum Dis 1998;57(4):203–208. Office for National Statistics. Population Mid-2000: Population change in size of selected age-groups within constituent countries of the United Kingdom, mid-1991 to mid-2000. HMSO. Available from URL: http// www.statistics.gov.uk/statbase. O’Reilly SC, Muir KR, Doherty M. Knee pain and disability in the Nottingham community: association with poor health status and psychological distress. Br J Rheumatol 1998a;37(8):870–873. O’Reilly SC, Jones A, Muir KR, Doherty M. Quadriceps weakness in knee osteoarthritis: the effect on pain and disability. Ann Rheum Dis 1998b;57(10):588–594. Pendleton A, Arden N, Dougados M, Doherty M, Bannwarth B, Bijlsma JW, Cluzeau F, Cooper C, Dieppe PA, Gunther KP, Hauselmann HJ, Herrero-Beaumont G, Kaklamanis PM, Leeb B, Lequesne M, Lohmander S, Mazieres B, Mola EM, Pavelka K, Serni U, Swoboda B, Verbruggen AA, Weseloh G, Zimmermann-Gotska I. EULAR recommendations for the management of knee osteoarthritis: report of a task force of the Standing Committee for International Clinical Studies Including Therapeutic Trials (ESCISIT). Ann Rheum Dis 2000;59(12):936–944. Roos EM, Klassbo M, Lohmander LS. WOMAC osteoarthritis index. Reliability, validity, and responsiveness in patients with arthroscopically assessed osteoarthritis. Western Ontario and MacMaster Universities. Scand J Rheumatol 1999;28(4):210–215. Shrout P, Fleiss JL. Intraclass correlations:uses in assessing rater reliability. Psychol Bull 1979;86:420–428. Shrout P. Measurement reliability and agreement in psychiatry. Stat Methods Med Res 1998;7:301–317. Stucki G, Sangha O, Stucki S, Michel BA, Tyndall A, Dick W, Theiler R. Comparison of the WOMAC (Western Ontario and McMaster Universities) osteoarthritis index and a self-report format of the self-administered Lequesne–Algofunctional index in patients with knee and hip osteoarthritis. Osteoarthritis Cartilage 1998;6(2):79–86. Symmons DP. Knee pain in older adults: the latest musculoskeletal epidemic. Ann Rheum Dis 2001;60:89–90. Tennant A, Fear J, Pickering A, Hillman M, Cutts A, Chamberlain MA. Prevalence of knee problems in the population aged 55 years and over: identifying the need for knee arthroplasty. Br Med J 1995;310(6990):1291–1293. Townsend P, Phillimore P, Beattie A. Health and deprivation: inequality and the North London Croom Helm. 1988. Urwin M, Symmons D, Allison T, Brammah T, Busby H, Roxby M, Simmons A, Williams G. Estimating the burden of musculoskeletal disorders in the community: the comparative prevalence of symptoms at different anatomical sites, and the relation to social deprivation [see comments]. Ann Rheum Dis 1998;57(11):649–655. Ware JE, Snow K, Gandek B, Kosinski M SF36 health survey. Manual and interpretation guide, Boston, MA: Health Institute, New England Medical Centre, 1993. Wolfe F, Kong SX. Rasch analysis of the Western Ontario MacMaster questionnaire (WOMAC) in 2205 patients with osteoarthritis, rheumatoid arthritis, and fibromyalgia. Ann Rheum Dis 1999;58(9):563–568. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand 1983;67(361):370.