- Email: [email protected]
Analysis of regional variation in hip and knee joint replacement rates in England using Hospital Episodes Statistics
Public Health (2006) 120, 83–90
Analysis of regional variation in hip and knee joint replacement rates in England using Hospital Episodes Statistics T. Dixona, M.E. Shawb, P.A. Dieppeb,* a
National Centre for Monitoring Diabetes, Australian Institute of Health and Welfare, Canberra, Australia Department of Social Medicine, University of Bristol, Canynge Hall, Whiteladies Road, Bristol BS8 2PR, UK
b
Received 25 November 2004; received in revised form 10 May 2005; accepted 28 June 2005 Available online 29 September 2005
KEYWORDS primary total hip replacement; primary total knee replacement; regional variation; inequality
Summary Objectives: Total hip and knee joint replacements are effective interventions for people with severe arthritis, and demand for these operations appears to be increasing as our population ages. This study explores regional variations in health care and inequalities in the provision of these expensive interventions, which are high on the UK Government’s health agenda. Study design: The Hospital Episode Statistics (HES) for England were analysed. The HES database holds information on patients who are admitted to National Health Service (NHS) hospitals in England. Methods: Age-standardized procedure rates were calculated using 5-year age groups with the English mid-year population of 2000 as the reference. Univariate associations between age-standardized operation rates and regional characteristics were assessed using Pearson’s correlation coefficient. Results: Age and sex-standardized surgery rates vary by 25-30%. For both hip and knee replacement, rates are highest in the South West and Midlands and lowest in the North West, South East and London regions. In the case of knee replacement, there are also marked differences in the sex ratios between regions. The variable that explained most variation in hip replacement rates was the proportion of older people in the region. In the case of knee replacement, the number of NHS centres offering surgery in the region was the main explanatory variable, with regions with fewer centres having the highest provision rates. Conclusion: These data can help to inform planning of services. They suggest that there may be inequities as well as inequalities in the provision of primary joint replacement surgery in England. Q 2005 The Royal Institute of Public Health. Published by Elsevier Ltd. All rights reserved.
* Corresponding author. Tel.: C44 117 928 7343; fax: C44 117 928 7236. E-mail address: [email protected] (P.A. Dieppe).
0033-3506/$ - see front matter Q 2005 The Royal Institute of Public Health. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.puhe.2005.06.003
84
Introduction Regional variations in health care and inequities in the provision of expensive interventions, including joint replacement, are high on the UK Government’s health agenda and attract extensive media coverage.1,2 Total hip and knee joint replacements (THR/TKR) are effective interventions for people with severe arthritis,3–5 but there are no clear indications for these interventions, making it difficult to examine the appropriateness of surgery in individuals or populations.6,7 Demand for these operations appears to be increasing as our population ages,8 and there are continuing concerns about the ability of supply to meet these demands in the UK, waiting times and outcomes.9,10 One Government response to these pressures in England has been to set up new ‘independent sector treatment centres’ (ISTCs) to increase the capacity for surgery. However, there is relatively little data in the public sector on the volume of or trends in joint replacement surgery, or on regional variations in England, and information on their determinants is particularly sparse. These data are needed to inform policy decisions such as the siting of ISTCs. We are attempting to fill that knowledge gap. The Hospital Episode Statistics (HES) for England database contains records of all surgical procedures undertaken in English National Health Service (NHS) hospitals. We have recently examined this database to determine volumes of THR and TKR surgery (both primary and revision operation rates) and trends over the 10-year period from 1991 to 2001.11 In addition to substantial increases in operation rates over the 10-year period, we found substantial variations by socio-economic status, with the most deprived fifth of the population experiencing significantly lower rates of surgery than those in higher socio-economic groups. We have now explored these data further in order to examine regional variations in England and their determinants. The particular aims of the research reported here were: (1) to determine the extent of regional variation in the provision of THR and TKR in England; (2) to find out whether any such variations had increased or decreased over the last decade; and (3) if such variation exists, to describe characteristics associated with regions of high and low provision.
Methods The HES for England were analysed (Source: Department of Health). The HES database holds
T. Dixon et al. information on patients who are admitted to NHS hospitals in England, either as a day case or as an ordinary admission; private hospital cases are excluded. Each record in the database refers to one ‘finished consultant episode’, i.e. the period of time that an individual spends under the care of one NHS consultant. Information held in HES includes age, sex, area of usual residence, reason for admission (diagnostic codes) and procedure performed (OPCS codes).12 Data for this analysis were obtained from the HES extract held by the Medical Research Council (MRC) Health Services Research Collaboration in Bristol; at the time of the work described here, this contained data from April 1991 to March 2001. We have presented and analysed the data for 2000 to look at contemporary regional variations and associations, and data from 1991, 1996 and 2000 to compare trends over time. Episodes involving total joint replacement (TJR) were identified as those with any of the OPCS4 codes W37, W38, W39 (hip) or W40, W41 or W42 (knee) recorded in any of the four procedure fields. Operations that were partial or complete revisions, rather than primary replacements, were excluded by omitting anyone with an admission code indicating adjustment of, or complication to, an existing prosthesis (ICD-9 codes 996.4, 996.6 or 996.7 and ICD-10 codes T84 or T85). The geographic areas used were the eight administrative NHS regions of England existing in 2000. These regions can be further broken down into health authorities (HAs). Between 1991 and 2000, the geographical boundaries of the HAs and NHS regions were redrawn several times. Using information from the Department of Health regarding re-assignment of HA boundaries over time,13 the HES data were assigned to HAs and then amalgamated into regions according to year 2001 boundaries. Age-standardized procedure rates were calculated using 5-year age groups with the English mid-year population of 2000 as the reference. Univariate associations between age-standardized operation rates and regional characteristics were assessed using Pearson’s correlation coefficient. Variables with a correlation P value of less than 0.1 were then entered into a forward stepwise multiple linear regression model (selection criteria: entry P!0.2, exit PS0.25) to determine what proportion of the variation in procedure rates could be explained by available regional characteristics. All analyses were performed using Stata version 8.2.
Regional variation in hip and knee joint replacement rates in England
85
Table 1 Age-standardized rates per 100 000 population (and 95% confidence intervals) of primary hip and knee replacement by region, 2000 Total hip replacement
Region
Males North & Yorkshire Trent West Midlands North West Eastern London South East South West England
68.6 65.5 74.5 59.6 66.6 55.4 60.1 83.2 65.5
(65.5, (62.2, (71.0, (56.7, (63.4, (52.5, (57.6, (79.6, (64.4,
71.7) 68.8) 78.0) 62.4) 69.9) 58.3) 62.6) 86.8) 66.5)
Total knee replacement Females
Males
87.5 (84.4, 90.6) 83.6 (80.3, 87.0) 98.1 (94.5, 101.7) 79.1 (76.2, 82.0) 88.3 (85.0, 91.7) 80.3 (77.2, 83.3) 82.8 (80.2, 85.3) 108.2 (104.5, 111.9) 87.1 (86.0, 88.2)
68.9 72.2 68.5 61.3 55.9 54.3 55.8 67.0 62.1
Results Regional variation In the year 2000, a total of 47 932 THR and 35 939 TKR procedures were performed in England. Agestandardized rates of both THR and TKR in 2000 showed substantial variation by region for both men and women (Table 1). The geography of the two procedures was broadly similar (Figures 1 and 2), with the lowest rates for both procedures being seen in the North West, South East and London regions and the highest rates seen in the South West and Midlands (TKR rates are also high in Trent). Variation by sex was different for the two procedures: for THR, the pattern of variation is similar
Fig. 1
(65.8, (68.8, (65.1, (58.4, (52.9, (51.4, (53.4, (63.8, (61.0,
Females 72.0) 75.7) 71.8) 64.2) 58.9) 57.3) 58.2) 70.2) 63.1)
69.0 80.0 80.6 61.9 69.3 70.3 63.9 80.3 70.7
(66.3, (76.7, (77.3, (59.4, (66.4, (67.4, (61.7, (77.1, (69.7,
71.8) 83.2) 83.8) 64.5) 72.3) 73.2) 66.2) 83.4) 71.7)
for men and women, but there were big differences for TKR. For example, for men, Trent had the highest TKR rate and London had the lowest, whereas rates were highest in the West Midlands and lowest in the North West for women. The amount of regional variation decreased between 1991 and 2000. However, in 2000, the highest rates were still about 25-30% higher than those in regions with the lowest rates (Figures 3 and 4).
Associations with regional variation The potential explanatory variables that we were able to investigate include population size, age structure, ethnicity, social class, unemployment
Regions of relatively high and low provision of primary total hip replacement, all persons, 2000.
86
T. Dixon et al.
Fig. 2
Regions of relatively high and low provision of primary total knee replacement, all persons, 2000.
rates, numbers with limiting long-term illness, standardized mortality rates, and numbers of centres in the region offering TJR. As shown in Table 2, most of these variables were not strongly correlated with age-standardized operation rates. Generally, higher procedure rates tended to occur in regions with a smaller total population, a larger proportion of older residents, a higher proportion of
the working-age population in manual social classes (IIIM, IV and V), and in those regions with fewer centres offering joint replacement. The final linear regression models of explanatory variables were different for THR and TKR (see Table 3). For THR, the single variable that explained about half of the regional variation in operation rates was the proportion of older people
ASR per 100 000 1991 1996
120
2000 100
80
60
40
20
0 North & Yorkshire
Trent
West Midlands
North West
East
London
South East South West
Region
Fig. 3 Age-standardized total hip replacement rates by region, all persons, 1991, 1996 and 2000 (error bars refer to 95% confidence intervals).
Regional variation in hip and knee joint replacement rates in England
87
ASR per 100 000
1991 1996
90
2000 80 70 60 50 40 30 20 10 0 North & Yorkshire
Trent
West Midlands
North West
East
London
South East South West
Region
Fig. 4 Age-standardized total knee replacement rates by region, all persons, 1991, 1996 and 2000 (error bars refer to 95% confidence intervals).
(aged 65-84 years) in spite of the fact that the data used were age-standardized. For TKR, the explanatory variable was the number of NHS centres offering surgery, with those regions with fewer centres doing more operations; this alone explains 58% of the regional variation.
Discussion Examination of the HES database has shown that the age-standardized rates of primary THR and TKR varied between English regions by about 25-30% in
2000. For THR rates, the number of older people in the population was the main explanatory variable (higher rates in regions with more older people, even after adjustment for the age profiles of regions), but the numbers of centres offering the procedure explained most of the variation (fewer centres being associated with higher operation rates) in TKR rates. The methodology of this study has both strengths and weaknesses. The absence of data from the private sector means that an estimated 20-30% of procedures are missing from our analyses;13 a figure that is large enough to potentially obliterate most
Table 2 Pearson correlations between age-standardized joint replacement rates and regional sociodemographic variables, all persons, 2000. Variable
Mean value (range)
Correlation with total hip replacement rate (P value)
Correlation with total knee replacement rate (P value)
Population (thousands) % in least-deprived quartile % in most-deprived quartile % aged 65-84 years Number of centres offering total joint replacement % ethnic population % in manual social classes Unemployment rate % of people over 65 years old with limiting long-term illness SMR
6123.8 (4909.3-8607.2) 24.9 (7.2-39.3) 24.6 (8.1-52.4) 14.1 (11.0-16.1) 22 (15-29)
-0.64 (0.085) 0.28 (0.498) -0.45 (0.260) 0.72 (0.045) -0.66 (0.073)
-0.76 (0.028) -0.11 (0.794) -0.07 (0.866) 0.48 (0.231) -0.80 (0.017)
7.2 (1.7-25.9) 35.9 (25.8-40.7) 5.2 (3.3-7.0) 41 (36-49)
-0.38 (0.354) 0.49 (0.215) -0.16 (0.705) -0.01 (0.972)
-0.17 (0.695) 0.70 (0.054) 0.21 (0.622) 0.50 (0.209)
98 (89-109)
-0.30 (0.464)
0.04 (0.919)
88
T. Dixon et al.
Table 3 Results of stepwise multiple linear regression models predicting age-standardized rate of joint replacement procedures, all persons, 2000 Model and variable For total hip replacement: % aged 65-84 years For total knee replacement: Number of centres offering surgery
P value
Adj R2
4.686
0.044
0.436
-0.887
0.017
0.584
Coefficient
of the regional variation seen. For example, relatively low rates of NHS operations in the South East and London regions, where the population is generally better off and more likely to have private health insurance than in other regions, might be more than compensated for by higher than average rates of private surgery in these regions. In addition, the accuracy and completeness of the HES database is by no means perfect,14 and the changes made to regional boundaries over the study period may be another source of inaccuracy. On the positive side, the database is very large (with over 80 000 TJRs in 2000 alone), and it is likely that coding of TJRs is reasonably good in view of the importance and clarity of these procedures. For the analysis of explanatory variables, we were dependent on national census data and were not able to include some factors that might be of importance, such as regional variations in the enthusiasm for TJRs amongst patients and medical professionals. In order to obtain a joint replacement in the NHS, a patient has to be willing to undergo the procedure and be referred from a general practitioner to the surgeon. It is clear from studies in other countries that surgeons, general practitioners and patients vary in their enthusiasm for and willingness to undergo TJRs15,16 and such factors could explain regional variations. Waiting lists may influence provision but it is difficult to obtain appropriate information about them on a regional basis, and we believe it would be difficult to interpret data that included them in the analysis. It is well known that surgical rates show wide variations, both between and within countries.17,18 Our data show that English regional variations in rates of provision are a little lower than they were a decade ago, but they remain high. There are no clear indications regarding who should or should not have these procedures, or at what stage of the development of arthritis (relatively early or relatively late) it is best to consider surgery.19,20 Given
the lack of consensus on indications, it is perhaps not surprising that individual surgeon enthusiasm for the procedures was found to be an important explanatory variable in one study,15 whilst variations in the willingness of people to consider surgery appeared to explain variations in provision rates in another study.21 We found some marked differences in the variability of THR and TKR in England, and different explanatory variables for each of them. In the case of THR, all regions have similar sex ratios, and the main explanatory variable for regional variation is the number of older people in the region, in spite of the data having been age-standardized. We interpret this as suggesting that regions with a high proportion of older people have deliberately increased provision of THR in the hope of meeting the presumed high need. The situation with TKR is quite different, even though the mean age at which the operation is done is similar to that for the hip. In addition to regional variations in provision rates, we see quite marked sex differences between regions. Also, there is the seemingly paradoxical finding that provision is highest in regions with the fewest numbers of centres carrying out the operations. Our interpretation of this finding relates to the difference in the development of services for primary THR and TKR in England over the last 50 years. Soon after Charnley’s pioneering work on hip replacements, the operation became popular throughout the country and most orthopaedic surgeons began to use it. Knee replacement came later, and proved to be a more demanding and difficult procedure.22 As a result, it developed more slowly and became the province of more specialized centres rather than being taken on by all orthopaedic departments in every district hospital. It would appear that this concentration of the technique in specialist centres might have progressed more successfully in some English regions than in others. We have shown that there are variations in the provision of joint replacements in different parts of England, but the more pressing question is whether there are inequities in their provision. These findings, taken in conjunction with our previous analyses, suggest that there might be, particularly in the case of TKR. We have previously shown that although we may be doing approximately the appropriate numbers of THR procedures in England, we are likely to be underproviding for TKR procedures.7,23 In addition, our previous work on the HES database showed that those in the most deprived fifth of the population experience lower rates than everyone else.11 Those data, in addition to the new findings on variations by region in the numbers and sex distribution of TKR procedures,
Regional variation in hip and knee joint replacement rates in England suggest disparities and probably inequities in provision. Furthermore, recently published data from a cohort study of 15 000 people aged 65 years or over in Sheffield and Wiltshire show that the most socio-economically deprived people have the greatest potential need for joint replacement but are less likely to receive services.24,25 Such inequities could be due to problems of patient access to surgery, variations in individual preferences and willingness to undergo the procedure, or the enthusiasm or behaviour of referring general practitioners and orthopaedic surgeons. Our data do not allow us to say which of these are most likely to explain the variations in provision and likely inequity. There are three important contextual issues that pertain to these findings. First, increasing numbers of these operations are being undertaken in England, where the population continues to age. Second, new means of provision of surgery, such as ISTCs, are appearing, and waiting times are going down. Third, the Government and the NHS are committed to equitable distribution of health care. Our data suggest that there are inequalities, and probably inequities, in the provision of joint replacements in England. We believe that there is an urgent need for the development of consensus on the indications for these operations and of local appropriateness criteria6 that could be used to examine inequities and their causes.
Conclusion English regional variations in TJR provision remain large, are not easily explained, and differ for hips and knees. It is likely that considerable inequalities and inequities in provision exist, particularly in the case of primary TKR.
Acknowledgements HES data were made available by the Department of Health. HES analyses conducted within the Department of Social Medicine are supported by the South West Public Health Observatory and the MRC Health Services Research Collaboration. The Department of Social Medicine is the lead centre of the MRC Health Services Research Collaboration. At the time of writing, Tracy Dixon was supported by the MRC Health Services Research Collaboration. Mary Shaw is funded by the South West Public Health Observatory.
89
References [1] Department of Health. Reducing health inequalities action report. London: TSO; 1999. [2] National Audit Office. Hip replacements: getting it right first time. 43rd report. London: TSO; 2000. [3] Scott D, Smith C, Lohmander S, Chard J. Osteoarthritis In: Clinical evidence, issue 10. London: BMJ Publishing Group; 2003. [4] Faulkner A, Kennedy LG, Baxter K, Donovan J, Wilkinson M, Bevan G. Effectiveness of hip prosthesis in primary total hip replacement: a critical review of the evidence and an economic model. Health Technol Assess 1998;2:1–33. [5] Callahan C, Drake B, Heck D, Dittus R. Patient outcomes following tricompartmental total knee replacement. A meta-analysis. J Am Med Assoc 1994;271:1349–57. [6] Quintana J, Azkarate J, Goenaga J, Arostegui I, Beldarrain I, Vil J. Evaluation of the appropriateness of hip joint replacement techniques. Int J Technol Assess Health Care 2000;16:165–77. [7] Frankel S, Eachus J, Pearson N, et al. Population requirement for primary hip-replacement surgery: a crosssectional study. Lancet 1999;353:1304–9. [8] Birrell F, Johnell O, Silman A. Projecting the need for hip replacement over the next three decades: influence of changing demography and threshold for surgery. Ann Rheum Dis 1999;58:569–72. [9] Williams O, Fitzpatrick R, Hajat S, et al. Mortality, morbidity and 1-year outcomes of primary elective total hip arthroplasty. J Arthroplasty 2002;2:165–71. [10] The Royal College of Surgeons of England. An Investigation of the Performance of the 3M Capital Hip System. London: The Royal College of Surgeons of England; 2001. [11] Dixon T, Shaw M, Ebrahim S, Dieppe P. Trends in hip and knee joint replacement: socio-economic inequalities and projections of need. Ann Rheum Dis 2004;63:825–30. [12] National Health Service Information Authority. Organisation codes service CD-ROM, August 2003 edition. Exeter: NHSIA; 2003. [13] Williams B, Whatmough P, McGill J, Rushton L. Patients and procedures in short-stay independent hospitals in England and Wales, 1997–1998. J Public Health Med 2000;22:68–73. [14] Department of Health. Hospital Episodes Statistics Data Quality Index Report 2000–01. Department of Health. [15] Wright J, Hawker G, Bombardier C, Croxford R, Dittus RS, Freund DA, et al. Physician enthusiasm as an explanation for area variation in the utilisation of knee replacement surgery. Med Care 1999;37:946–56. [16] Clark J, Hudak P, Hawker G, Coyte PC, Mahomed N, Kreder HJ, et al. The moving target: a qualitative study of elderly patients’ decision-making regarding total joint replacement surgery. J Bone Joint Surg Br 2004;86-A: 1366–74. [17] Katz B, Freund D, Heck D, Dittus RS, Paul JE, Wright J, et al. Demographic variation in the rate of knee replacement: a multi-year analysis. Health Serv Res 1996;31:125–40. [18] Merx H, Dreinhofer Schrader P, Sturmer T, Puhl W, Gu ¨nther K-P, Brenner H. International variation in hip replacement rates. Ann Rheum Dis 2003;62:222–6. [19] Fortin P, Penrod J, Clarke A, St-Pierre Y, Joseph L, Belisle P, et al. Timing of total joint replacement affects clinical outcomes among patients with osteoarthritis of the hip or knee. Arthritis Rheum 2002;46:3327–30. [20] Kennedy G, Newman J, Ackroyd C, Dieppe P. When should we do knee replacements? Knee 2002;10:161–6.
90 [21] Hawker G, Wright J, Coyte P, Williams J, Harvey B, Glazier R, et al. Determining the need for hip and knee arthroplasties: the role of clinical severity and patients preferences. Med Care 2001;39:206–16. [22] Fisher J. Surgery for arthritis: total hip and knee joint replacement. Chesterfield: Arthritis Research; 2004. [23] Juni P, Dieppe P, Donovan J, Peters T, Eachus J, Pearson N, et al. Population requirement for primary knee
T. Dixon et al. replacement surgery: a cross-sectional study. Rheumatology 2003;42:516–21. [24] Young PFK, Milner PC, Payne JN, et al. Inequalities in access to knee joint replacements for people in need. Ann Rheum Dis 2004;63:1483–9. [25] 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.
Analysis of regional variation in hip and knee joint replacement rates in England using Hospital Episodes Statistics T. Dixona, M.E. Shawb, P.A. Dieppeb,* a
National Centre for Monitoring Diabetes, Australian Institute of Health and Welfare, Canberra, Australia Department of Social Medicine, University of Bristol, Canynge Hall, Whiteladies Road, Bristol BS8 2PR, UK
b
Received 25 November 2004; received in revised form 10 May 2005; accepted 28 June 2005 Available online 29 September 2005
KEYWORDS primary total hip replacement; primary total knee replacement; regional variation; inequality
Summary Objectives: Total hip and knee joint replacements are effective interventions for people with severe arthritis, and demand for these operations appears to be increasing as our population ages. This study explores regional variations in health care and inequalities in the provision of these expensive interventions, which are high on the UK Government’s health agenda. Study design: The Hospital Episode Statistics (HES) for England were analysed. The HES database holds information on patients who are admitted to National Health Service (NHS) hospitals in England. Methods: Age-standardized procedure rates were calculated using 5-year age groups with the English mid-year population of 2000 as the reference. Univariate associations between age-standardized operation rates and regional characteristics were assessed using Pearson’s correlation coefficient. Results: Age and sex-standardized surgery rates vary by 25-30%. For both hip and knee replacement, rates are highest in the South West and Midlands and lowest in the North West, South East and London regions. In the case of knee replacement, there are also marked differences in the sex ratios between regions. The variable that explained most variation in hip replacement rates was the proportion of older people in the region. In the case of knee replacement, the number of NHS centres offering surgery in the region was the main explanatory variable, with regions with fewer centres having the highest provision rates. Conclusion: These data can help to inform planning of services. They suggest that there may be inequities as well as inequalities in the provision of primary joint replacement surgery in England. Q 2005 The Royal Institute of Public Health. Published by Elsevier Ltd. All rights reserved.
* Corresponding author. Tel.: C44 117 928 7343; fax: C44 117 928 7236. E-mail address: [email protected] (P.A. Dieppe).
0033-3506/$ - see front matter Q 2005 The Royal Institute of Public Health. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.puhe.2005.06.003
84
Introduction Regional variations in health care and inequities in the provision of expensive interventions, including joint replacement, are high on the UK Government’s health agenda and attract extensive media coverage.1,2 Total hip and knee joint replacements (THR/TKR) are effective interventions for people with severe arthritis,3–5 but there are no clear indications for these interventions, making it difficult to examine the appropriateness of surgery in individuals or populations.6,7 Demand for these operations appears to be increasing as our population ages,8 and there are continuing concerns about the ability of supply to meet these demands in the UK, waiting times and outcomes.9,10 One Government response to these pressures in England has been to set up new ‘independent sector treatment centres’ (ISTCs) to increase the capacity for surgery. However, there is relatively little data in the public sector on the volume of or trends in joint replacement surgery, or on regional variations in England, and information on their determinants is particularly sparse. These data are needed to inform policy decisions such as the siting of ISTCs. We are attempting to fill that knowledge gap. The Hospital Episode Statistics (HES) for England database contains records of all surgical procedures undertaken in English National Health Service (NHS) hospitals. We have recently examined this database to determine volumes of THR and TKR surgery (both primary and revision operation rates) and trends over the 10-year period from 1991 to 2001.11 In addition to substantial increases in operation rates over the 10-year period, we found substantial variations by socio-economic status, with the most deprived fifth of the population experiencing significantly lower rates of surgery than those in higher socio-economic groups. We have now explored these data further in order to examine regional variations in England and their determinants. The particular aims of the research reported here were: (1) to determine the extent of regional variation in the provision of THR and TKR in England; (2) to find out whether any such variations had increased or decreased over the last decade; and (3) if such variation exists, to describe characteristics associated with regions of high and low provision.
Methods The HES for England were analysed (Source: Department of Health). The HES database holds
T. Dixon et al. information on patients who are admitted to NHS hospitals in England, either as a day case or as an ordinary admission; private hospital cases are excluded. Each record in the database refers to one ‘finished consultant episode’, i.e. the period of time that an individual spends under the care of one NHS consultant. Information held in HES includes age, sex, area of usual residence, reason for admission (diagnostic codes) and procedure performed (OPCS codes).12 Data for this analysis were obtained from the HES extract held by the Medical Research Council (MRC) Health Services Research Collaboration in Bristol; at the time of the work described here, this contained data from April 1991 to March 2001. We have presented and analysed the data for 2000 to look at contemporary regional variations and associations, and data from 1991, 1996 and 2000 to compare trends over time. Episodes involving total joint replacement (TJR) were identified as those with any of the OPCS4 codes W37, W38, W39 (hip) or W40, W41 or W42 (knee) recorded in any of the four procedure fields. Operations that were partial or complete revisions, rather than primary replacements, were excluded by omitting anyone with an admission code indicating adjustment of, or complication to, an existing prosthesis (ICD-9 codes 996.4, 996.6 or 996.7 and ICD-10 codes T84 or T85). The geographic areas used were the eight administrative NHS regions of England existing in 2000. These regions can be further broken down into health authorities (HAs). Between 1991 and 2000, the geographical boundaries of the HAs and NHS regions were redrawn several times. Using information from the Department of Health regarding re-assignment of HA boundaries over time,13 the HES data were assigned to HAs and then amalgamated into regions according to year 2001 boundaries. Age-standardized procedure rates were calculated using 5-year age groups with the English mid-year population of 2000 as the reference. Univariate associations between age-standardized operation rates and regional characteristics were assessed using Pearson’s correlation coefficient. Variables with a correlation P value of less than 0.1 were then entered into a forward stepwise multiple linear regression model (selection criteria: entry P!0.2, exit PS0.25) to determine what proportion of the variation in procedure rates could be explained by available regional characteristics. All analyses were performed using Stata version 8.2.
Regional variation in hip and knee joint replacement rates in England
85
Table 1 Age-standardized rates per 100 000 population (and 95% confidence intervals) of primary hip and knee replacement by region, 2000 Total hip replacement
Region
Males North & Yorkshire Trent West Midlands North West Eastern London South East South West England
68.6 65.5 74.5 59.6 66.6 55.4 60.1 83.2 65.5
(65.5, (62.2, (71.0, (56.7, (63.4, (52.5, (57.6, (79.6, (64.4,
71.7) 68.8) 78.0) 62.4) 69.9) 58.3) 62.6) 86.8) 66.5)
Total knee replacement Females
Males
87.5 (84.4, 90.6) 83.6 (80.3, 87.0) 98.1 (94.5, 101.7) 79.1 (76.2, 82.0) 88.3 (85.0, 91.7) 80.3 (77.2, 83.3) 82.8 (80.2, 85.3) 108.2 (104.5, 111.9) 87.1 (86.0, 88.2)
68.9 72.2 68.5 61.3 55.9 54.3 55.8 67.0 62.1
Results Regional variation In the year 2000, a total of 47 932 THR and 35 939 TKR procedures were performed in England. Agestandardized rates of both THR and TKR in 2000 showed substantial variation by region for both men and women (Table 1). The geography of the two procedures was broadly similar (Figures 1 and 2), with the lowest rates for both procedures being seen in the North West, South East and London regions and the highest rates seen in the South West and Midlands (TKR rates are also high in Trent). Variation by sex was different for the two procedures: for THR, the pattern of variation is similar
Fig. 1
(65.8, (68.8, (65.1, (58.4, (52.9, (51.4, (53.4, (63.8, (61.0,
Females 72.0) 75.7) 71.8) 64.2) 58.9) 57.3) 58.2) 70.2) 63.1)
69.0 80.0 80.6 61.9 69.3 70.3 63.9 80.3 70.7
(66.3, (76.7, (77.3, (59.4, (66.4, (67.4, (61.7, (77.1, (69.7,
71.8) 83.2) 83.8) 64.5) 72.3) 73.2) 66.2) 83.4) 71.7)
for men and women, but there were big differences for TKR. For example, for men, Trent had the highest TKR rate and London had the lowest, whereas rates were highest in the West Midlands and lowest in the North West for women. The amount of regional variation decreased between 1991 and 2000. However, in 2000, the highest rates were still about 25-30% higher than those in regions with the lowest rates (Figures 3 and 4).
Associations with regional variation The potential explanatory variables that we were able to investigate include population size, age structure, ethnicity, social class, unemployment
Regions of relatively high and low provision of primary total hip replacement, all persons, 2000.
86
T. Dixon et al.
Fig. 2
Regions of relatively high and low provision of primary total knee replacement, all persons, 2000.
rates, numbers with limiting long-term illness, standardized mortality rates, and numbers of centres in the region offering TJR. As shown in Table 2, most of these variables were not strongly correlated with age-standardized operation rates. Generally, higher procedure rates tended to occur in regions with a smaller total population, a larger proportion of older residents, a higher proportion of
the working-age population in manual social classes (IIIM, IV and V), and in those regions with fewer centres offering joint replacement. The final linear regression models of explanatory variables were different for THR and TKR (see Table 3). For THR, the single variable that explained about half of the regional variation in operation rates was the proportion of older people
ASR per 100 000 1991 1996
120
2000 100
80
60
40
20
0 North & Yorkshire
Trent
West Midlands
North West
East
London
South East South West
Region
Fig. 3 Age-standardized total hip replacement rates by region, all persons, 1991, 1996 and 2000 (error bars refer to 95% confidence intervals).
Regional variation in hip and knee joint replacement rates in England
87
ASR per 100 000
1991 1996
90
2000 80 70 60 50 40 30 20 10 0 North & Yorkshire
Trent
West Midlands
North West
East
London
South East South West
Region
Fig. 4 Age-standardized total knee replacement rates by region, all persons, 1991, 1996 and 2000 (error bars refer to 95% confidence intervals).
(aged 65-84 years) in spite of the fact that the data used were age-standardized. For TKR, the explanatory variable was the number of NHS centres offering surgery, with those regions with fewer centres doing more operations; this alone explains 58% of the regional variation.
Discussion Examination of the HES database has shown that the age-standardized rates of primary THR and TKR varied between English regions by about 25-30% in
2000. For THR rates, the number of older people in the population was the main explanatory variable (higher rates in regions with more older people, even after adjustment for the age profiles of regions), but the numbers of centres offering the procedure explained most of the variation (fewer centres being associated with higher operation rates) in TKR rates. The methodology of this study has both strengths and weaknesses. The absence of data from the private sector means that an estimated 20-30% of procedures are missing from our analyses;13 a figure that is large enough to potentially obliterate most
Table 2 Pearson correlations between age-standardized joint replacement rates and regional sociodemographic variables, all persons, 2000. Variable
Mean value (range)
Correlation with total hip replacement rate (P value)
Correlation with total knee replacement rate (P value)
Population (thousands) % in least-deprived quartile % in most-deprived quartile % aged 65-84 years Number of centres offering total joint replacement % ethnic population % in manual social classes Unemployment rate % of people over 65 years old with limiting long-term illness SMR
6123.8 (4909.3-8607.2) 24.9 (7.2-39.3) 24.6 (8.1-52.4) 14.1 (11.0-16.1) 22 (15-29)
-0.64 (0.085) 0.28 (0.498) -0.45 (0.260) 0.72 (0.045) -0.66 (0.073)
-0.76 (0.028) -0.11 (0.794) -0.07 (0.866) 0.48 (0.231) -0.80 (0.017)
7.2 (1.7-25.9) 35.9 (25.8-40.7) 5.2 (3.3-7.0) 41 (36-49)
-0.38 (0.354) 0.49 (0.215) -0.16 (0.705) -0.01 (0.972)
-0.17 (0.695) 0.70 (0.054) 0.21 (0.622) 0.50 (0.209)
98 (89-109)
-0.30 (0.464)
0.04 (0.919)
88
T. Dixon et al.
Table 3 Results of stepwise multiple linear regression models predicting age-standardized rate of joint replacement procedures, all persons, 2000 Model and variable For total hip replacement: % aged 65-84 years For total knee replacement: Number of centres offering surgery
P value
Adj R2
4.686
0.044
0.436
-0.887
0.017
0.584
Coefficient
of the regional variation seen. For example, relatively low rates of NHS operations in the South East and London regions, where the population is generally better off and more likely to have private health insurance than in other regions, might be more than compensated for by higher than average rates of private surgery in these regions. In addition, the accuracy and completeness of the HES database is by no means perfect,14 and the changes made to regional boundaries over the study period may be another source of inaccuracy. On the positive side, the database is very large (with over 80 000 TJRs in 2000 alone), and it is likely that coding of TJRs is reasonably good in view of the importance and clarity of these procedures. For the analysis of explanatory variables, we were dependent on national census data and were not able to include some factors that might be of importance, such as regional variations in the enthusiasm for TJRs amongst patients and medical professionals. In order to obtain a joint replacement in the NHS, a patient has to be willing to undergo the procedure and be referred from a general practitioner to the surgeon. It is clear from studies in other countries that surgeons, general practitioners and patients vary in their enthusiasm for and willingness to undergo TJRs15,16 and such factors could explain regional variations. Waiting lists may influence provision but it is difficult to obtain appropriate information about them on a regional basis, and we believe it would be difficult to interpret data that included them in the analysis. It is well known that surgical rates show wide variations, both between and within countries.17,18 Our data show that English regional variations in rates of provision are a little lower than they were a decade ago, but they remain high. There are no clear indications regarding who should or should not have these procedures, or at what stage of the development of arthritis (relatively early or relatively late) it is best to consider surgery.19,20 Given
the lack of consensus on indications, it is perhaps not surprising that individual surgeon enthusiasm for the procedures was found to be an important explanatory variable in one study,15 whilst variations in the willingness of people to consider surgery appeared to explain variations in provision rates in another study.21 We found some marked differences in the variability of THR and TKR in England, and different explanatory variables for each of them. In the case of THR, all regions have similar sex ratios, and the main explanatory variable for regional variation is the number of older people in the region, in spite of the data having been age-standardized. We interpret this as suggesting that regions with a high proportion of older people have deliberately increased provision of THR in the hope of meeting the presumed high need. The situation with TKR is quite different, even though the mean age at which the operation is done is similar to that for the hip. In addition to regional variations in provision rates, we see quite marked sex differences between regions. Also, there is the seemingly paradoxical finding that provision is highest in regions with the fewest numbers of centres carrying out the operations. Our interpretation of this finding relates to the difference in the development of services for primary THR and TKR in England over the last 50 years. Soon after Charnley’s pioneering work on hip replacements, the operation became popular throughout the country and most orthopaedic surgeons began to use it. Knee replacement came later, and proved to be a more demanding and difficult procedure.22 As a result, it developed more slowly and became the province of more specialized centres rather than being taken on by all orthopaedic departments in every district hospital. It would appear that this concentration of the technique in specialist centres might have progressed more successfully in some English regions than in others. We have shown that there are variations in the provision of joint replacements in different parts of England, but the more pressing question is whether there are inequities in their provision. These findings, taken in conjunction with our previous analyses, suggest that there might be, particularly in the case of TKR. We have previously shown that although we may be doing approximately the appropriate numbers of THR procedures in England, we are likely to be underproviding for TKR procedures.7,23 In addition, our previous work on the HES database showed that those in the most deprived fifth of the population experience lower rates than everyone else.11 Those data, in addition to the new findings on variations by region in the numbers and sex distribution of TKR procedures,
Regional variation in hip and knee joint replacement rates in England suggest disparities and probably inequities in provision. Furthermore, recently published data from a cohort study of 15 000 people aged 65 years or over in Sheffield and Wiltshire show that the most socio-economically deprived people have the greatest potential need for joint replacement but are less likely to receive services.24,25 Such inequities could be due to problems of patient access to surgery, variations in individual preferences and willingness to undergo the procedure, or the enthusiasm or behaviour of referring general practitioners and orthopaedic surgeons. Our data do not allow us to say which of these are most likely to explain the variations in provision and likely inequity. There are three important contextual issues that pertain to these findings. First, increasing numbers of these operations are being undertaken in England, where the population continues to age. Second, new means of provision of surgery, such as ISTCs, are appearing, and waiting times are going down. Third, the Government and the NHS are committed to equitable distribution of health care. Our data suggest that there are inequalities, and probably inequities, in the provision of joint replacements in England. We believe that there is an urgent need for the development of consensus on the indications for these operations and of local appropriateness criteria6 that could be used to examine inequities and their causes.
Conclusion English regional variations in TJR provision remain large, are not easily explained, and differ for hips and knees. It is likely that considerable inequalities and inequities in provision exist, particularly in the case of primary TKR.
Acknowledgements HES data were made available by the Department of Health. HES analyses conducted within the Department of Social Medicine are supported by the South West Public Health Observatory and the MRC Health Services Research Collaboration. The Department of Social Medicine is the lead centre of the MRC Health Services Research Collaboration. At the time of writing, Tracy Dixon was supported by the MRC Health Services Research Collaboration. Mary Shaw is funded by the South West Public Health Observatory.
89
References [1] Department of Health. Reducing health inequalities action report. London: TSO; 1999. [2] National Audit Office. Hip replacements: getting it right first time. 43rd report. London: TSO; 2000. [3] Scott D, Smith C, Lohmander S, Chard J. Osteoarthritis In: Clinical evidence, issue 10. London: BMJ Publishing Group; 2003. [4] Faulkner A, Kennedy LG, Baxter K, Donovan J, Wilkinson M, Bevan G. Effectiveness of hip prosthesis in primary total hip replacement: a critical review of the evidence and an economic model. Health Technol Assess 1998;2:1–33. [5] Callahan C, Drake B, Heck D, Dittus R. Patient outcomes following tricompartmental total knee replacement. A meta-analysis. J Am Med Assoc 1994;271:1349–57. [6] Quintana J, Azkarate J, Goenaga J, Arostegui I, Beldarrain I, Vil J. Evaluation of the appropriateness of hip joint replacement techniques. Int J Technol Assess Health Care 2000;16:165–77. [7] Frankel S, Eachus J, Pearson N, et al. Population requirement for primary hip-replacement surgery: a crosssectional study. Lancet 1999;353:1304–9. [8] Birrell F, Johnell O, Silman A. Projecting the need for hip replacement over the next three decades: influence of changing demography and threshold for surgery. Ann Rheum Dis 1999;58:569–72. [9] Williams O, Fitzpatrick R, Hajat S, et al. Mortality, morbidity and 1-year outcomes of primary elective total hip arthroplasty. J Arthroplasty 2002;2:165–71. [10] The Royal College of Surgeons of England. An Investigation of the Performance of the 3M Capital Hip System. London: The Royal College of Surgeons of England; 2001. [11] Dixon T, Shaw M, Ebrahim S, Dieppe P. Trends in hip and knee joint replacement: socio-economic inequalities and projections of need. Ann Rheum Dis 2004;63:825–30. [12] National Health Service Information Authority. Organisation codes service CD-ROM, August 2003 edition. Exeter: NHSIA; 2003. [13] Williams B, Whatmough P, McGill J, Rushton L. Patients and procedures in short-stay independent hospitals in England and Wales, 1997–1998. J Public Health Med 2000;22:68–73. [14] Department of Health. Hospital Episodes Statistics Data Quality Index Report 2000–01. Department of Health. [15] Wright J, Hawker G, Bombardier C, Croxford R, Dittus RS, Freund DA, et al. Physician enthusiasm as an explanation for area variation in the utilisation of knee replacement surgery. Med Care 1999;37:946–56. [16] Clark J, Hudak P, Hawker G, Coyte PC, Mahomed N, Kreder HJ, et al. The moving target: a qualitative study of elderly patients’ decision-making regarding total joint replacement surgery. J Bone Joint Surg Br 2004;86-A: 1366–74. [17] Katz B, Freund D, Heck D, Dittus RS, Paul JE, Wright J, et al. Demographic variation in the rate of knee replacement: a multi-year analysis. Health Serv Res 1996;31:125–40. [18] Merx H, Dreinhofer Schrader P, Sturmer T, Puhl W, Gu ¨nther K-P, Brenner H. International variation in hip replacement rates. Ann Rheum Dis 2003;62:222–6. [19] Fortin P, Penrod J, Clarke A, St-Pierre Y, Joseph L, Belisle P, et al. Timing of total joint replacement affects clinical outcomes among patients with osteoarthritis of the hip or knee. Arthritis Rheum 2002;46:3327–30. [20] Kennedy G, Newman J, Ackroyd C, Dieppe P. When should we do knee replacements? Knee 2002;10:161–6.
90 [21] Hawker G, Wright J, Coyte P, Williams J, Harvey B, Glazier R, et al. Determining the need for hip and knee arthroplasties: the role of clinical severity and patients preferences. Med Care 2001;39:206–16. [22] Fisher J. Surgery for arthritis: total hip and knee joint replacement. Chesterfield: Arthritis Research; 2004. [23] Juni P, Dieppe P, Donovan J, Peters T, Eachus J, Pearson N, et al. Population requirement for primary knee
T. Dixon et al. replacement surgery: a cross-sectional study. Rheumatology 2003;42:516–21. [24] Young PFK, Milner PC, Payne JN, et al. Inequalities in access to knee joint replacements for people in need. Ann Rheum Dis 2004;63:1483–9. [25] 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.