Modeling the need for hip and knee replacement surgery. Part 1. A two-stage cross-cohort approachкод для вставкиСкачать
Arthritis & Rheumatism (Arthritis Care & Research) Vol. 61, No. 12, December 15, 2009, pp 1657–1666 DOI 10.1002/art.24892 © 2009, American College of Rheumatology ORIGINAL ARTICLE Modeling the Need for Hip and Knee Replacement Surgery. Part 1. A Two-Stage Cross-Cohort Approach ANDY JUDGE,1 NICKY J. WELTON,1 JAT SANDHU,2 AND YOAV BEN-SHLOMO1 Objective. To explore inequalities in the need for hip/knee replacement surgery using a 2-stage cross-cohort approach. Methods. In the ﬁrst stage, a small-area population-based survey, the Somerset and Avon Survey of Health, was used to provide a high-quality measure of need for hip/knee replacement using the New Zealand (NZ) score. Receiver operating characteristic curve analyses were used to validate a simpliﬁed NZ score, excluding information from clinical examination. In the second stage, a nationally representative population-based survey, the English Longitudinal Study of Ageing, was used to explore inequalities in need for hip/knee replacement using the simpliﬁed NZ score. Multilevel Poisson regression modeling was used to estimate rates of need for surgery. Exposures considered were age, sex, social class, ethnicity, obesity, Index of Multiple Deprivation 2004 deprivation quintiles, rurality, and ethnic mix of area. Results. Rates of need for hip/knee replacement increase with age and are lower in men than in women (rate ratio [RR] 0.7, 95% conﬁdence interval [95% CI] 0.6 – 0.9 for hips; RR 0.8, 95% CI 0.7–1.0 for knees). Those of lowest social class have greater need. Need was greatest for people living in more deprived areas. Individual ethnic group did not predict the need for surgery. For hip replacement, there was no rurality effect; for knee replacement, those in town and fringe areas had greater need. Obesity was a strong predictor of need for surgery (RR 2.3, 95% CI 1.9 –2.8 for hips; RR 2.4, 95% CI 2.0 –2.8 for knees). Conclusion. This study provides evidence of greater variations of inequalities in need for hip/knee replacement than previous studies. Further research should explore geographic variation and produce small-area estimates of need to inform local health planning. It is important to complement data on need with willingness to undergo surgery. Joint replacements make a substantial contribution to public health and are among the most common elective procedures. They are cost effective (1,2), with good prosthesis survival rates (3,4), reduced pain, increased mobility, and improved quality of life (5–12). However, health needs will not be the same across different areas of a country and will vary according to the demographic characteristics of the area. In the UK, the Musculoskeletal Services Framework produced by the Department of Health recognizes this, describing needs assessment in the context of understanding the prevalence and incidence of musculoskeletal disorders, where patients are, and their use of services (13). Priority action must be taken to distribute resources The opinions expressed by the authors are theirs alone and do not represent the opinions of supporting organizations. The developers and funders of the English Longitudinal Study of Ageing and the UK Data Archive do not bear any responsibility for the analyses or interpretations presented here. Supported by the Medical Research Council/Health Services Research Collaboration at the Department of Social Medicine, University of Bristol. Dr. Welton’s work was supported by the UK Medical Research Council. Dr. Sandhu’s work was supported by a National Coordinating Centre for Research Capacity Development Department of Health Public Health Initiative 2003. The data are from the English Longitudinal Study of Ageing, which receives funding from the US National Institute on Aging and a consortium of UK government departments coordinated by the Ofﬁce for National Statistics, and the Somerset and Avon Survey of Health, which receives funding from the Department of Health and the South and West National Health Service Research and Development Directorate. 1 Andy Judge, PhD (current address: University of Oxford, Oxford, UK), Nicky J. Welton, PhD, Yoav Ben-Shlomo, MBBS, PhD: University of Bristol, Bristol, UK; 2Jat Sandhu, PhD: University of Bristol, Bristol, UK, and University of British Columbia, Vancouver, British Columbia, Canada. Address correspondence to Andy Judge, PhD, Botnar Research Centre, Institute of Musculoskeletal Sciences, Nufﬁeld Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences, University of Oxford, Windmill Road, Headington, Oxford, OX3 7LD, UK. E-mail: andrew. firstname.lastname@example.org. Submitted for publication January 27, 2009; accepted in revised form July 30, 2009. INTRODUCTION 1657 1658 relative to health need; otherwise, inequities in accessing services occur that may lead to health inequalities. A number of population-based prevalence studies have attempted to explore inequalities in need for hip and knee replacement. They found that need increased with age and is greater in women than in men (14 –21). Poor people were more likely to need joint replacement (18,19), and need was associated with less education and lower income (22). There was no evidence that rurality was associated with need for hip and knee replacement (18,19). However, the majority of previous studies were conducted in geographically small areas, and estimates of need may vary geographically. Second, studies exclude people from the estimate of need based on a list of comorbidities that could potentially make them unﬁt for surgery. Having a listed comorbidity does not necessarily mean a patient is ineligible for surgery. Third, to our knowledge, no studies have looked at whether need for joint replacement varies by ethnic group. Finally, studies only look at a few sociodemographic domains, such as age and sex (23). In this study, we explore all relevant sociodemographic variables and risk factors in a multivariable regression analysis that considers the effects of interactions, overdispersion, and geographic variation. The advantage of small-area population-based studies is that they are speciﬁcally designed to estimate the population requirement for joint replacement and have a highquality measure of need. For example, in the Somerset and Avon Survey of Health (16,17) and Ontario studies (20), the need for surgery is conﬁrmed radiographically and through clinical examination. However, small-area studies are limited in terms of their generalizability. On the other hand, large nationally representative population surveys are more generalizable, but are often not designed to examine a speciﬁc health problem and rarely have detailed clinical data and/or radiography. One possible solution that combines the strengths of these 2 types of study is a 2-stage cross-cohort approach whereby in the ﬁrst stage, a small-area population-based study is obtained with a high-quality measure of need (scoring tool) to identify people requiring hip and knee replacement. A simpler measure of need (shorter version of the scoring tool) is then created using information available in a national survey, which would typically exclude data from clinical examinations and radiographs. Receiver operating characteristic (ROC) curve analyses are then performed to validate the simpler measure of need against the gold standard in the small-area survey. In the second stage, the nationally representative survey can be used to identify people in need of hip and knee replacement using the simpler scoring tool. We illustrate this approach with an example of joint replacement from the UK. However, the methods are general and applicable in other settings and conditions where health care services need to make provision decisions based on the estimated need. MATERIALS AND METHODS Stage 1. We obtained access to the Somerset and Avon Survey of Health, a small-area population-based survey Judge et al conducted in 1994 –1995, used to estimate need for hip (17) and knee (16) replacement. The sample was obtained using a multistage sampling strategy (24) of 28,080 people age ⱖ35 years from 40 general practices in the former UK counties of Avon and Somerset. Questionnaires were completed on symptoms of hip/knee pain and stiffness, comorbidity, limitations to activities of daily living, previous use of health services, preferences and priorities for treatment, and indicators of socioeconomic status. A full clinical examination of the hip, knee, and lower back was carried out by a medically trained researcher and a team of nurses with orthopaedic experience who had undergone a standard training program and who followed an examination schedule. If clinically signiﬁcant hip or knee disease was discovered at the clinic, study participants were invited for a radiograph of the joint. Because the aim was to assess the need for primary joint replacement, respondents reporting a previous hip/knee replacement on the same side as the reported symptoms were excluded. For people with hip/knee pain, the severity of their joint disease was assessed using the New Zealand (NZ) score (25) (Table 1), which is a continuous score from 0 to 100 comprised of 4 main components: degree and occurrence of pain, functional limitations, pain on clinical examination and other abnormal ﬁndings, and involvement of other joints and the degree to which independence was threatened. Higher scores reﬂect more severe disease. A cutoff of ⱖ55 was previously used to identify those in need of surgery (16,17). The score was developed through professional consensus in New Zealand to determine access to and priority for joint replacement, with a recent study conﬁrming the validity of the score (26). In the UK, some primary care trusts have begun using the NZ score as a tool to determine access to an orthopaedic surgeon (27) to ensure that referrals to a consultant are appropriate. Within this study, the radiographic ﬁndings were not incorporated into the NZ score because of the well-known uncertainty about the relationship between radiographic ﬁndings and symptoms, and uncertainty as to how best to incorporate radiographic ﬁndings into the NZ criteria. We acquired data from the English Longitudinal Study of Ageing, a nationally representative population-based sample of 11,392 people age ⱖ50 years living in private households in the UK and developed by a team of researchers based at the National Centre for Social Research, University College London, and the Institute for Fiscal Studies (28). The sample was drawn from the households that previously responded to the Health Survey for England in 1998, 1999, or 2001 (the 1999 survey included a boost sample that represented ethnic minorities). Several waves of data are available from the Economic and Social Data Service: wave 0 contains data for the English Longitudinal Study of Ageing participants from the Health Survey for England data sets, and wave 1 was conducted in 2002–2003 and contains information from individual interviews and self-completion questionnaires. The health module contains information on the severity of hip/knee pain and activities of daily living. A weighting variable is included with the data set. The aim of weighting is to take account of any bias from nonresponse in order to make the Inequalities in the Need for Hip and Knee Replacement 1659 Table 1. The New Zealand Priority Criteria for Major Joint Replacement Surgery (maximum score 100) Clinical features Pain (40%) Degree (patient must be on the maximum medical therapy at the time of rating) None Mild: slight or occasional pain; patient has not altered patterns of activity or work Mild to moderate: moderate or frequent pain; patient has not altered patterns of activity or work Moderate: patient is active but has had to modify or give up some activities because of pain Moderate to severe: fairly severe pain with substantially limited activities Severe: major pain and serious limitation Occurrence None or with ﬁrst steps only Only after long walks (30 minutes) With all walking, mostly day pain Signiﬁcant, regular night pain Functional activity (20%) Time walked Unlimited 31–60 minutes (e.g., longer shopping trips to the mall) 11–30 minutes (e.g., gardening, grocery shopping) 2–10 minutes (e.g., trip to the letter box) ⬍2 minutes or indoors only (more or less housebound) Unable to walk Other functional limitations (e.g., putting on shoes, managing stairs, sitting to standing, sexual activity, recreation or hobbies, walking aids needed) None Mild Moderate Severe Movement and deformity (20%) Pain on examination (overall results are both active and passive range of motion) None Mild Moderate Severe Other abnormal ﬁndings (limited to orthopaedic problems, e.g., reduced range of motion, deformity, limp, instability, progressive radiograph ﬁndings) None Mild Moderate Severe Other factors (20%) Multiple joint disease No, single joint Yes, each affected joint mild to moderate in severity Yes, severe involvement (e.g., severe rheumatoid arthritis) Ability to work, give care to dependants, live independently (difﬁculty must be related to affected joint) Not threatened or difﬁcult Not threatened but more difﬁcult Threatened but not immediately Immediately threatened Total respondent sample more representative of the population. Calibration weighting was used, which attaches an estimated probability of response to each household that explains the discrepancy between the survey and the distribution of age and sex in the population. We went back to the original Somerset and Avon Survey of Health questionnaires and determined how the information they contained was used to create an NZ score. Questions used were then compared with those in the English Longitudinal Study of Ageing and matched as Score 0 4 6 9 14 20 0 4 10 20 0 2 4 6 8 10 0 2 4 10 0 2 5 10 0 2 5 10 0 4 10 0 4 6 10 100 closely as possible, enabling a simpler (proxy) NZ score to be created (see Supplementary Table 1, available in the online version of this article at http://www3.interscience. wiley.com/journal/77005015/home). Patients in the English Longitudinal Study of Ageing were assigned an NZ score if they had had a limiting long-term illness over a period of time, were often troubled with pain, and had pain in either their hips or knees. Because a clinical examination was not performed, it was not possible to complete the section on “pain on examination and other ab- 1660 Judge et al normal ﬁndings” of the NZ score; hence, a simpler score out of 80 was used. Stage 2. The nationally representative survey (English Longitudinal Study of Ageing) was used to explore inequalities in the need for hip and knee replacement, using as an outcome variable the simpler NZ score out of 80 with a cutoff of 48 to create a binary variable of whether or not a person was in need of surgery. The individual and ecologic exposure variables were explored. Individual. The following patient-level information was extracted: age (50 –54, 55–59, 60 – 64, 65– 69, 70 –74, 75– 79, 80 – 84, and ⱖ85 years), sex, occupational social class (I ⫽ professional, II ⫽ managerial and technical, IIIN ⫽ skilled nonmanual, IIIM ⫽ skilled manual, IV ⫽ partly skilled, and V ⫽ unskilled), ethnicity (white and nonwhite), and obesity (body mass index ⬍30 kg/m2 and ⱖ30 kg/m2; measured by an interviewer who took height/ weight measurements). Ecologic. Geographic information was not readily available in the archived English Longitudinal Study of Ageing data set on the Economic and Social Data Service Web site. An application was made to the National Centre for Social Research to obtain additional geographic data. The Census Area Statistics ward (anonymized) and the district the patient lives in allowed multilevel modeling to be used to explore geographic variation and control for clustering in the data. The following ecologic data have been linked to the census ward a patient lives in: the Index of Multiple Deprivation 2004 deprivation quintiles (weighted to the ward population because each census ward varies in size: 1 [least deprived], 2, 3, 4, and 5 [most deprived]) (29), rurality (urban [population ⱖ10,000], town and fringe, and village/isolated), and ethnic mix of the area (white [ⱖ10% white and ⱕ0.5% African American, Asian, and other] and nonwhite [all remaining groups]). Statistical methods. The outcome of interest was a binary/dichotomous variable of whether or not the patient was in need of hip or knee replacement. Exposure variables consisted of age, sex, social class, ethnicity, obesity, Index of Multiple Deprivation 2004 deprivation quintiles, rurality, and ethnic mix of the area. A univariable Poisson regression model was ﬁtted in Stata statistical software (StataCorp, College Station, TX) to examine the association between the rates of need for joint replacement and each of the sociodemographic variables. We also ﬁtted a multivariable model controlling for all variables. Analyses were weighted to control for bias from nonrandom nonresponse in the English Longitudinal Study of Ageing sample. Wald tests were used to explore linear trends by ﬁtting models with the variable as a score. To assess for nonlinear trend, likelihood ratio tests were used, comparing a model with a categorical variable with a model with the variable as a score. Effect modiﬁcation was considered by using likelihood ratio tests for interaction between each of the sociodemographic variables. Separate models were ﬁtted for hip and knee replacement. The hierarchical structure of the data consisted of 11,392 individuals nested within 2,913 wards within 348 Figure 1. Receiver operating characteristic (ROC) curve for hips. districts. Multilevel Poisson regression models were ﬁtted in the statistical software MLwiN (University of Bristol, Bristol, UK). For the hip model, the intraclass correlation coefﬁcient was 0.032; hence, 99.0% of the variation in rates of need for hip replacement was at the individual level (for the knee model, 99.8% of the variation was at the individual level). There was no evidence of clustering across either wards (P ⫽ 1.00 for hip and knee models) or districts (P ⫽ 0.56 hip, P ⫽ 0.90 knee); hence, the simpler ﬁxed-effects regression model was adequate. RESULTS We conducted an ROC curve analysis on the Somerset and Avon Survey of Health data set, taking the NZ score out of 100 with a cutoff of ⱖ55 to be the gold standard to determine a threshold for the simpler score out of 80. For both hip and knee replacement, the area under the curve was maximized using a threshold of 44 (providing the best tradeoff between sensitivity and speciﬁcity); hence, the simpler score can reliably be used. However, a threshold of 48 was chosen for analysis because this correctly classiﬁes the greatest number of people as to whether they are in need of surgery (sensitivity 89.5%, speciﬁcity 98.8%, cor- Figure 2. Receiver operating characteristic (ROC) curve for knees. Inequalities in the Need for Hip and Knee Replacement 1661 Table 2. Rates of need for hip replacement by sociodemographic groups* Individual-level variables Age groups, years 50–54 55–59 60–64 65–69 70–74 75–79 80–84 ⱖ85 P linear trend P nonlinear trend† Sex Female Male Social class I. professional II. managerial and technical IIIN. skilled nonmanual IIIM. skilled manual IV. partly skilled V. unskilled P linear trend P nonlinear trend† Ethnicity White Nonwhite Ecologic variables IMD 2004 1 (least deprived) 2 3 4 5 (most deprived) P linear trend P nonlinear trend† Rurality Urban (population ⱖ10,000) Town and fringe Village/isolated P linear trend P nonlinear trend† Ethnic mix of area White Nonwhite Risk factors Obesity BMI ⬍30 kg/m2 BMI ⱖ30 kg/m2 Number (%) Crude rate of need per 1,000 (95% CI) 1,981 (17.4) 2,185 (19.2) 1,688 (14.8) 1,711 (15.0) 1,471 (12.9) 1,094 (9.6) 806 (7.1) 456 (4.0) Crude RR (95% CI) Adjusted RR (95% CI) 29.0 (22.6–37.3) 39.7 (32.2–48.8) 41.2 (32.8–51.7) 43.5 (34.9–54.3) 51.7 (41.4–64.4) 61.6 (48.9–77.7) 77.0 (60.7–97.8) 71.2 (50.9–99.4) 1.00 1.37 (0.99–1.89) 1.42 (1.01–1.99) 1.50 (1.07–2.09) 1.78 (1.27–2.48) 2.12 (1.51–2.99) 2.65 (1.88–3.75) 2.45 (1.61–3.72) ⬍ 0.001 0.84 1.00 1.33 (0.94–1.89) 1.15 (0.79–1.67) 1.23 (0.85–1.77) 1.37 (0.94–1.99) 1.95 (1.34–2.83) 2.44 (1.65–3.61) 2.84 (1.75–4.59) ⬍ 0.001 0.30 6,205 (54.5) 5,187 (45.5) 54.6 (49.2–60.6) 36.5 (31.7–41.9) 1.00 0.67 (0.56–0.80) 1.00 0.72 (0.58–0.90) 497 (4.4) 2,997 (26.3) 2,618 (23.0) 2,218 (19.5) 1,779 (15.6) 785 (6.9) 7.5 (2.8–20.0) 29.3 (24.0–35.9) 43.0 (35.9–51.5) 57.7 (48.8–68.3) 66.1 (55.6–78.7) 60.1 (45.4–79.6) 1.00 3.91 (1.44–10.62) 5.73 (2.12–15.51) 7.69 (2.85–20.77) 8.81 (3.26–23.84) 8.01 (2.89–22.20) ⬍ 0.001 0.003 1.00 2.51 (0.92–6.84) 2.98 (1.09–8.18) 4.40 (1.62–11.92) 3.72 (1.36–10.21) 3.03 (1.06–8.65) 0.002 0.006 10,996 (96.5) 320 (2.8) 44.8 (41.1–48.9) 81.4 (56.4–117.4) 1.00 1.81 (1.24–2.64) 1.00 1.43 (0.86–2.37) 2,573 (22.6) 2,530 (22.2) 2,348 (20.6) 2,161 (19.0) 1,779 (15.6) 25.6 (20.1–32.5) 35.1 (28.5–43.1) 47.9 (40.1–57.4) 57.1 (48.1–67.9) 75.6 (64.3–88.9) 1.00 1.37 (1.00–1.88) 1.87 (1.39–2.53) 2.23 (1.66–3.00) 2.96 (2.21–3.95) ⬍ 0.001 0.89 1.00 1.22 (0.85–1.75) 1.67 (1.19–2.35) 2.08 (1.48–2.93) 2.39 (1.68–3.40) ⬍ 0.001 0.88 8,606 (75.5) 1,393 (12.2) 1,392 (12.2) 48.9 (44.5–53.7) 44.2 (34.6–56.4) 31.2 (23.3–41.9) 1.00 0.90 (0.69–1.17) 0.64 (0.47–0.87) 0.004 0.42 1.00 1.24 (0.92–1.67) 1.00 (0.69–1.45) 0.57 0.22 2,163 (19.0) 9,228 (81.0) 45.3 (37.3–55.1) 46.4 (42.3–50.9) 1.00 1.02 (0.83–1.27) –‡ –‡ 7,556 (66.3) 2,566 (22.5) 31.3 (27.6–35.6) 72.9 (63.6–83.7) 1.00 2.33 (1.93–2.81) 1.00 2.28 (1.88–2.76) * 95% CI ⫽ 95% conﬁdence interval; RR ⫽ rate ratio; IMD ⫽ Index of Multiple Deprivation; BMI ⫽ body mass index. † To assess for nonlinear trend, likelihood ratio tests were used, comparing a model with a categorical variable with a model with the variable as a score. ‡ Variable excluded from fully adjusted model because no evidence was associated with rates of need for hip replacement. rectly classiﬁed 97.9% for the hip; sensitivity 85.9%, speciﬁcity 97.6%, correctly classiﬁed 96.3% for the knee) (Figures 1 and 2). When screening individuals for surgery, we wanted to reduce the risk of false-positive results (operating on someone that does not need surgery). By favoring increased speciﬁcity, an individual considered not in need of surgery using the gold standard is less likely to be considered in need using the simpler score. The overall rate of need for hip replacement in the UK, adjusted for all of the variables in the multivariable model, 1662 Judge et al Table 3. Rates of need for knee replacement by sociodemographic groups* Individual-level variables Age groups, years 50–54 55–59 60–64 65–69 70–74 75–79 80–84 ⱖ85 P linear trend P nonlinear trend† Sex Female Male Social class I. professional II. managerial and technical IIIN. skilled nonmanual IIIM. skilled manual IV. partly skilled V. unskilled P linear trend P nonlinear trend† Ethnicity White Nonwhite Ecologic variables IMD 2004 1 (least deprived) 2 3 4 5 (most deprived) P linear trend P nonlinear trend† Rurality Urban (population ⱖ10,000) Town and fringe Village/isolated P linear trend P nonlinear trend† Ethnic mix of area White Nonwhite Risk factors Obesity BMI ⬍30 kg/m2 BMI ⱖ30 kg/m2 Number (%) Crude rate of need per 1,000 (95% CI) 1,981 (17.4) 2,185 (19.2) 1,688 (14.8) 1,711 (15.0) 1,471 (12.9) 1,094 (9.6) 806 (7.1) 456 (4.0) Crude RR (95% CI) Adjusted RR (95% CI) 38.9 (31.3–48.3) 48.0 (39.8–58.0) 54.1 (44.4–66.1) 59.7 (49.5–71.9) 66.9 (55.2–81.2) 88.1 (72.8–106.5) 104.1 (85.0–127.4) 91.9 (68.5–123.2) 1.00 1.24 (0.93–1.65) 1.39 (1.04–1.87) 1.53 (1.15–2.04) 1.72 (1.29–2.30) 2.26 (1.70–3.02) 2.67 (1.99–3.60) 2.36 (1.64–3.40) ⬍ 0.001 0.74 1.00 1.26 (0.92–1.71) 1.23 (0.89–1.70) 1.24 (0.91–1.70) 1.42 (1.03–1.96) 2.17 (1.58–2.98) 2.45 (1.74–3.45) 2.86 (1.87–4.36) ⬍ 0.001 0.31 6,205 (54.5) 5,187 (45.5) 70.2 (64.0–77.0) 50.7 (45.1–57.1) 1.00 0.72 (0.62–0.84) 1.00 0.78 (0.65–0.95) 497 (4.4) 2,997 (26.3) 2,618 (23.0) 2,218 (19.5) 1,779 (15.6) 785 (6.9) 11.3 (5.1–25.2) 38.9 (32.6–46.4) 56.1 (47.9–65.7) 75.5 (65.3–87.3) 84.8 (72.8–98.7) 88.5 (70.6–110.8) 1.00 3.43 (1.52–7.76) 4.94 (2.19–11.15) 6.65 (2.96–14.97) 7.47 (3.32–16.83) 7.80 (3.40–17.86) ⬍ 0.001 0.005 1.00 2.61 (1.07–6.37) 3.15 (1.28–7.72) 4.30 (1.77–10.45) 4.06 (1.66–9.95) 3.77 (1.51–9.42) ⬍ 0.001 0.02 10,996 (96.5) 320 (2.8) 59.7 (55.4–64.3) 108.9 (79.5–149.4) 1.00 1.82 (1.32–2.52) 1.00 1.17 (0.74–1.86) 2,573 (22.6) 2,530 (22.2) 2,348 (20.6) 2,161 (19.0) 1,779 (15.6) 33.7 (27.4–41.5) 44.5 (37.1–53.5) 58.3 (49.5–68.6) 77.1 (66.6–89.3) 108.2 (94.6–123.6) 1.00 1.32 (1.00–1.74) 1.73 (1.33–2.25) 2.29 (1.77–2.95) 3.21 (2.51–4.11) ⬍ 0.001 0.97 1.00 1.21 (0.89–1.66) 1.55 (1.15–2.08) 2.14 (1.60–2.87) 2.70 (2.01–3.64) ⬍ 0.001 0.94 8,606 (75.5) 1,393 (12.2) 1,392 (12.2) 64.3 (59.3–69.8) 60.5 (49.1–74.4) 42.2 (32.9–54.2) 1.00 0.94 (0.75–1.17) 0.66 (0.50–0.85) 0.002 0.26 1.00 1.36 (1.05–1.75) 1.11 (0.81–1.52) 0.14 0.09 2,163 (19.0) 9,228 (81.0) 63.1 (53.6–74.3) 60.7 (56.0–65.8) 1.00 0.96 (0.80–1.15) –‡ –‡ 7,556 (66.3) 2,566 (22.5) 39.6 (35.3–44.3) 99.3 (88.3–111.6) 1 2.51 (2.13–2.95) 1 2.41 (2.04–2.84) * 95% CI ⫽ 95% conﬁdence interval; RR ⫽ rate ratio; IMD ⫽ Index of Multiple Deprivation; BMI ⫽ body mass index. † To assess for nonlinear trend, likelihood ratio tests were used, comparing a model with a categorical variable with a model with the variable as a score. ‡ Variable excluded from fully adjusted model because no evidence was associated with rates of need for knee replacement. was 31.9 per 1,000 (95% conﬁdence interval [95% CI] 28.4 –35.8), and for knee replacement was 41.0 per 1,000 (95% CI 37.1– 45.4). In univariable analysis, rates of need for hip and knee replacement increase with age before falling slightly in those age ⱖ85 years (Tables 2 and 3). However, multivariable adjustment for other variables, predominantly obesity, strengthened the effect of those age ⱖ85 years such that they have the greatest need. For both hips and knees, women have greater need than men, although the effect was attenuated somewhat in multivari- Inequalities in the Need for Hip and Knee Replacement able analyses. For both joints, there was a strong linear effect of individual social class, with those of lowest social class having the greatest need. In multivariable models, the effect of individual social class was attenuated, mainly by adjustment for area deprivation and obesity. For hip and knee models, univariable analyses suggested that people of nonwhite ethnicity had greater need for joint replacement, but this was due to confounding by area deprivation and social class. There was a clear effect of area deprivation, where the need for hip and knee replacement was greatest for people living in the most deprived areas. This was not as strong as the effect of individual social class. For hip replacement, multivariable models suggest no rurality effect, this being attenuated by area deprivation. For knee replacement, there was some evidence that those in town and fringe areas had greater need. The ethnic mix of the area people live in was not associated with need for hip or knee replacement. Obesity was a strong predictor of need for hip and knee replacement, where obese people have more than twice the need compared with those who are not obese. For the knee model, a number of interactions were observed (see Supplementary Figure 1 and Supplementary Table 2, available in the online version of this article at http://www3.interscience.wiley.com/journal/77005015/ home). There was evidence of an interaction between age and sex (P ⫽ 0.002). Although the need for knee replacement increases with age, the association is weaker in men. The effect of sex is stronger in those age ⱖ70 years than in the younger age groups. There was an interaction between age and deprivation (P ⫽ 0.002) where the age gradient was weaker in the most deprived areas. Sensitivity analyses. In our analysis, we used a cutoff of 48 to classify people as being in need of joint replacement. Different choices of threshold will classify those with more/less severe disease as being in need of surgery. Regression analyses were therefore repeated using both lower and higher choices of threshold (scores of 43 and 53, respectively) as a sensitivity analysis. Our conclusions remained unchanged (results not shown); evidence of inequalities in the need for joint replacement by various sociodemographic groups was consistent regardless of the cutoff used. DISCUSSION This study has demonstrated evidence of inequalities in need for hip and knee replacement across different sociodemographic groups. We have shown how a 2-stage approach can be used to identify patients in need of surgery in large-scale nationally representative studies where cost is an issue. Although we have used 2 independent surveys, it would be better if in the ﬁrst stage, a subsample of patients was selected from the main nationally representative survey. These patients can then undergo a detailed clinical assessment of their need for surgery using the full version of a scoring system. ROC curve analyses can then be used to validate a simpler version of the scoring tool against the detailed version. Then in the second stage, the 1663 simpler scoring system can be applied to all of the patients in the national study. The 2-stage approach could be used to identify patients in need of hip and knee replacement using nationally representative surveys in other countries, and can also be applied to other important clinical indicators. Previous research has found that the need for hip and knee replacement increased with age and is greater in women than in men (14 –21), which is consistent with our ﬁndings. A Canadian study observed an interaction where there were no differences between men and women with low socioeconomic status, but in people with high socioeconomic status, women had greater need (22). No such interaction was observed in this study. Research suggests that poor people were more likely to be in need of joint replacement (18,19) and that need was associated with less education and lower income (22). We also found that need was greater in more deprived areas and in people of lower social class. In line with other studies, we found no evidence that rurality was associated with need for hip replacement (18,19), but did ﬁnd evidence that those in the town and fringe areas may have greater need of knee surgery. A population-based study by Ang et al found that in a sample of elderly male veterans with moderate to severe hip and knee pain, ethnicity was not a signiﬁcant predictor of Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) pain or function (30). Our ﬁndings also suggest that individual ethnic group is not associated with the need for joint replacement. A previous study by Steel et al used the English Longitudinal Study of Ageing to identify patients in need of joint replacement and explore inequalities in the need for surgery (21). They used a simpler measure of need based on the National Institutes of Health method that only uses information on the degree of hip and knee pain. Applying the questions used to assess need in this earlier study to the NZ criteria would create a score out of 20. Repeating ROC curve analyses using the score out of 20 shows that the area under the curve is smaller (see Supplementary Figures 2 and 3, available in the online version of this article at http://www3.interscience.wiley.com/journal/ 77005015/home), and hence there is a weaker measure of need. The added value of this study is that by using a 2-stage approach, we were able to create a more detailed score out of 80, including additional information on occurrence of pain and activities of daily living recorded in the questionnaires. With a weaker measure of need, there may be a greater degree of misclassiﬁcation and therefore underestimation of the effects of covariates such as age, sex, and social class (31). When comparing the ﬁndings of Steel et al with the results of our study, they found no association with age, women had greater need than men, a weak nonsigniﬁcant effect of social class was observed where lower social classes have greater need, and need was greater in obese people and in areas of poorest wealth. In our study, the associations were all in the same direction, but the effects of age, sex, social class, and obesity were far stronger, suggesting that using a strong measure of need provides a better picture of inequalities. However, there are other differences between the studies that could explain the discrepancy in the strength of inequalities 1664 observed. In the prior study, people age ⬍60 years were excluded, broad age categories were not used, and they looked at the overall need for hip or knee replacement rather than considering both joints separately. The beneﬁt of using the English Longitudinal Study of Ageing is that it is a large nationally representative sample with data on both individual and ecologic sociodemographic variables. However, it was not speciﬁcally designed to identify those in need of joint replacement. Because self-reported questions contain information on the severity of hip and knee pain and activities of daily living, it was possible to assign people a simpler version of the NZ score to measure disease severity without the need for a detailed clinical assessment. However, it was not possible to estimate side-speciﬁc scores to right and left hips and knees because this information is not available. Nor can we restrict the estimate of need to be for primary operations only. Although there is information on previous operations for those age ⬍60 years, we do not know which side the operation relates to; for example, if the operation was on the left hip, they may still need an operation on the right hip. By using a 2-stage approach, we have been able to describe evidence of inequalities in the need for hip and knee replacement using self-reported data from a nationally representative survey, which does not contain data from radiographs or a detailed clinical examination. This is set in the context of providing information at an aggregate level to inform health planning, and although the relationship between self-reported symptoms and radiographic severity has been questioned (32,33), at the individual level, a clinician would use additional information from clinical examination and radiographs regarding the decision for surgery. A limitation of our analysis and of all such studies estimating the need for joint replacement is the different methods used to assess the severity of joint disease, such as the WOMAC score (34), the NZ score (25), or the Lequesne Index (35). Authors derive an arbitrary cutoff based on these measures to determine whether patients require joint replacement. Clearly, different choices of cutoff will lead to different estimates of need for surgery, but sensitivity analyses can be used, repeating the analysis with different choices of cutoff to see if patterns of inequality remain unchanged. There appears to be a general lack of consensus about which indicator or threshold of disease severity to use (36), and until this is resolved, it will remain a limitation of future studies. Previous studies have adjusted their estimates of need according to a patient’s willingness and ﬁtness for having surgery. If adjusting for such factors attenuates inequalities, it may help explain why they exist. Older people are less willing to seek or want joint replacement (16,20,22,37– 40), considering the symptoms of arthritis a normal part of aging and adapting their lifestyles to cope (41– 45), whereas in younger people, the impact on work and social lives is greater so they seek surgery to get back to normal (16,37,41,45,46). Women may be less willing to have joint surgery (others suggest they are equally willing) (16,20,22,37,38), as are people of lower socioeconomic status (37) and African Americans (47). The common explanation is that these groups are less positive about the Judge et al beneﬁts and outcomes of surgery, as largely inﬂuenced by friends and family and those they know who had surgery (41,43,44). We have no information on willingness, and although data on comorbidities are available, it is unclear what would make a patient an unsuitable candidate for surgery, given improvements in modern anesthesia, surgical techniques, and prosthesis survival. Local health planners may argue that such estimates of need should be adjusted for these factors so that they can determine the level of provision for their area based only on those who are willing and ﬁt for surgery. Although this may be of use in resource allocation, willingness and ﬁtness for surgery are likely determinants of why inequities exist and planning services in such a way will not ensure fairness in accessing care, because these factors may change over time. In the UK, local health planners (primary care trusts) are responsible for the planning, commissioning, and delivery of National Health Service services (48). They must assess the health needs of people in their local area, ensuring that services are available to and can be accessed by everyone who needs them. The results of this study describe inequalities in the need for hip and knee replacement, but such data are of no use on their own. Local health planners require small-area estimates of need to plan an appropriate level of surgical provision. The second part of this study takes this work forward, producing estimates of need for joint replacement across small areas of the UK to inform local health planning. ACKNOWLEDGMENTS We would like to thank Dr. Mary Shaw at the Department of Social Medicine, University of Bristol, for support and advice throughout the project, and Professor Kelvyn Jones at the Department of Geographical Sciences, University of Bristol, for statistical advice on multilevel modeling. We would also like to thank all of the study participants and the partners and practice staff of participating general practices. We are indebted to the entire Somerset and Avon Survey of Health research team. 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 submitted for publication. Dr. Judge 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. Judge, Sandhu, Ben-Shlomo. Acquisition of data. Judge, Sandhu, Ben-Shlomo. Analysis and interpretation of data. Judge, Welton, Ben-Shlomo. REFERENCES 1. Liang MH, Cullen KE, Larson MG, Thompson MS, Schwartz JA, Fossel AH, et al. Cost-effectiveness of total joint arthroplasty in osteoarthritis. Arthritis Rheum 1986;29:937– 43. 2. Chang RW, Pellisier JM, Hazen GB. A cost-effectiveness analysis of total hip arthroplasty for osteoarthritis of the hip. JAMA 1996;275:858 – 65. 3. Rand JA, Ilstrup DM. Survivorship analysis of total knee Inequalities in the Need for Hip and Knee Replacement 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. arthroplasty: cumulative rates of survival of 9200 total knee arthroplasties. J Bone Joint Surg Am 1991;73:397– 409. Scuderi GR, Insall JN, Windsor RE, Moran MC. Survivorship of cemented knee replacements. J Bone Joint Surg Br 1989; 71B:798 – 803. Kirwan JR, Currey HL, Freeman MA, Snow S, Young PJ. Overall long-term impact of total hip and knee joint replacement surgery on patients with osteoarthritis and rheumatoid arthritis. Br J Rheumatol 1994;33:357– 60. Hawker G, Wright J, Coyte P, Paul J, Dittus R, Croxford R, et al. Health-related quality of life after knee replacement. J Bone Joint Surg Am 1998;80:163–73. Callahan CM, Drake BG, Heck DA, Dittus RS. Patient outcomes following unicompartmental or bicompartmental knee arthroplasty: a meta-analysis. J Arthroplasty 1995;10:141–50. Callahan CM, Drake BG, Heck DA, Dittus RS. Patient outcomes following tricompartmental total knee replacement: a meta-analysis. JAMA 1994;271:1349 –57. Williams MH, Frankel SJ, Nanchahl K, Coast J, Donovan JL. Total hip replacement. In: Stevens A, Raftery J, editors. Health care needs assessment: the epidemiological based needs assessment review. Vol. 1. Oxford: Radcliffe Medical Press; 1994. NHS Centre for Reviews and Dissemination. Effective health care bulletin: total hip replacement. York (UK): NHS Centre for Reviews and Dissemination; 1996. Walker DJ, Heslop PS, Chandler C, Pinder IM. Measured ambulation and self-reported health status following total joint replacement for the osteoarthritic knee. Rheumatology (Oxford) 2002;41:755– 8. McMurray A, Grant S, Grifﬁths S, Letford A. Health-related quality of life and health service use following total hip replacement surgery. J Adv Nurs 2002;40:663–72. Department of Health. The musculoskeletal services framework. A joint responsibility: doing it differently. 2006. URL: http://www.dh.gov.uk/prod_consum_dh/groups/dh_digital assets/@dh/@en/documents/digitalasset/dh_4138412.pdf. Fear J, Hillman M, Chamberlain MA, Tennant A. Prevalence of hip problems in the population aged 55 years and over: access to specialist care and future demand for hip arthroplasty. Br J Rheumatol 1997;36:74 – 6. 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. BMJ 1995;310:1291–3. Juni P, Dieppe P, Donovan J, Peters T, Eachus J, Pearson N, et al. Population requirement for primary knee replacement surgery: a cross-sectional study. Rheumatology (Oxford) 2003; 42:516 –21. Frankel S, Eachus J, Pearson N, Greenwood R, Chan P, Peters TJ, et al. Population requirement for primary hip-replacement surgery: a cross-sectional study. Lancet 1999;353:1304 –9. Yong PF, Milner PC, Payne JN, Lewis PA, Jennison C. Inequalities in access to knee joint replacements for people in need. Ann Rheum Dis 2004;63:1483–9. Milner PC, Payne JN, Stanﬁeld RC, Lewis PA, Jennison C, Saul C. Inequalities in accessing hip joint replacement for people in need. Eur J Public Health 2004;14:58 – 62. Hawker GA, Wright JG, Coyte PC, Williams JI, Harvey B, Glazier R, et al. Determining the need for hip and knee arthroplasty: the role of clinical severity and patients’ preferences. Med Care 2001;39:206 –16. Steel N, Melzer D, Gardener E, McWilliams B. Need for and receipt of hip and knee replacement: a national population survey. Rheumatology (Oxford) 2006;45:1437– 41. Hawker GA, Wright JG, Glazier RH, Coyte PC, Harvey B, Williams JI, et al. The effect of education and income on need and willingness to undergo total joint arthroplasty. Arthritis Rheum 2002;46:3331–9. Hawker GA, Wright JG, Coyte PC, Williams JI, Harvey B, Glazier R, et al. Differences between men and women in the rate of use of hip and knee arthroplasty. N Engl J Med 2002; 342:1016 –22. Peters TJ, Eachus JI. Achieving equal probability of selection 1665 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. under various random sampling strategies. Paediatr Perinat Epidemiol 1995;9:219 –24. Hadorn DC, Holmes AC. The New Zealand priority criteria project. Part 1: overview. BMJ 1997;314:131– 4. Toye F, Barlow J, Wright C, Lamb SE. A validation study of the New Zealand score for hip and knee surgery. Clin Orthop Relat Res 2007;464:190 –5. Report by the Comptroller and Auditor General. Hip replacements: an update. London: National Audit Ofﬁce; 2003. Marmot M, Banks J, Blundell R, Erens B, Lessof C, Nazroo J, et al. English Longitudinal Study of Ageing: wave 0 (1998, 1999 and 2001) and waves 1-3 (2002-2007). 10th ed. Essex: UK Data Archive; 2008. Noble M, Wright G, Dibben C, Smith G, McLennan D, Antilla C, et al. Indices of deprivation 2004: report to the Ofﬁce of the Deputy Prime Minister. London: Neighbourhood Renewal Unit; 2004. Ang DC, Ibrahim SA, Burant CJ, Kwoh CK. Is there a difference in the perception of symptoms between African Americans and whites with osteoarthritis? J Rheumatol 2003;30: 1305–10. Kelaher M, Paul S, Lambert H, Ahmad W, Smith GD. The impact of different measures of socioeconomic position on the relationship between ethnicity and health. Ann Epidemiol 2008;18:351– 6. Dieppe P, Judge A, Williams S, Ikwueke I, Guenther KP, Floeren M, et al. Variations in the pre-operative status of patients coming to primary hip replacement for osteoarthritis in European orthopaedic centres. BMC Musculoskelet Disord 2009;10:19. Ethgen O, Bruyere O, Richy F, Dardennes C, Reginster JY. Health-related quality of life in total hip and total knee arthroplasty: a qualitative and systematic review of the literature. J Bone Joint Surg Am 2004;86A:963–74. 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 1988;15:1833– 40. Lequesne MG, Mery C, Samson M, Gerard P. Indexes of severity for osteoarthritis of the hip and knee: validation. Value in comparison with other assessment tests. Scand J Rheumatol Suppl 1987;65:85–9. Dreinhofer KE, Dieppe P, Sturmer T, Grober-Gratz D, Floren M, Gunther KP, et al. Indications for total hip replacement: comparison of assessments of orthopaedic surgeons and referring physicians. Ann Rheum Dis 2006;65:1346 –50. Hawker GA, Wright JG, Badley EM, Coyte PC, for the Toronto Arthroplasty Health Services Research Consortium. Perceptions of, and willingness to consider, total joint arthroplasty in a population-based cohort of individuals with disabling hip and knee arthritis. Arthritis Rheum 2004;51:635– 41. Byrne MM, Souchek J, Richardson M, Suarez-Almazor M. Racial/ethnic differences in preferences for total knee replacement surgery. J Clin Epidemiol 2006;59:1078 – 86. Hawker GA, Guan J, Croxford R, Coyte PC, Glazier RH, Harvey BJ, et al. A prospective population-based study of the predictors of undergoing total joint arthroplasty. Arthritis Rheum 2006;54:3212–20. Dieppe P, Dixon D, Horwood J, Pollard B, Johnston M, and the MOBILE research team. MOBILE and the provision of total joint replacement. J Health Serv Res Policy 2008;13 Suppl 3:47–56. Sanders C, Donovan JL, Dieppe PA. Unmet need for joint replacement: a qualitative investigation of barriers to treatment among individuals with severe pain and disability of the hip and knee. Rheumatology (Oxford) 2004;43:353–7. Jinks C, Ong BN, Richardson J. A mixed methods study to investigate needs assessment for knee pain and disability: population and individual perspectives. BMC Musculoskelet Disord 2007;8:59. Hudak PL, Clark JP, Hawker GA, Coyte PC, Mahomed NN, Kreder HJ, et al. “You’re perfect for the procedure! Why don’t you want it?” Elderly arthritis patients’ unwillingness to con- 1666 sider total joint arthroplasty surgery: a qualitative study. Med Decis Making 2002;22:272– 8. 44. Ballantyne PJ, Gignac MA, Hawker GA. A patient-centered perspective on surgery avoidance for hip or knee arthritis: lessons for the future. Arthritis Rheum 2007;57:27–34. 45. Gignac MA, Davis AM, Hawker G, Wright JG, Mahomed N, Fortin PR, et al. “What do you expect? You’re just getting older”: a comparison of perceived osteoarthritis-related and aging-related health experiences in middle- and older-age adults. Arthritis Rheum 2006;55:905–12. 46. Sanders C, Donovan J, Dieppe P. The signiﬁcance and conse- Judge et al quences of having painful and disabled joints in older age: co-existing accounts of normal and disrupted biographies. Sociol Health Illn 2002;24:227–53. 47. Ibrahim SA, Siminoff LA, Burant CJ, Kwoh CK. Differences in expectations of outcome mediate African American/white patient differences in “willingness” to consider joint replacement. Arthritis Rheum 2002;46:2429 –35. 48. Department of Health. Delivering investment in general practice: implementing the new GMS contract. 2003. URL: http://www.dh.gov.uk/prod_consum_dh/groups/dh_digital assets/@dh/@en/documents/digitalasset/dh_4070231.pdf.