Total hip arthroplasty (THA) is a common treatment of severe osteoarthritis (OA) of the hip. The operation is done primarily to improve health-related quality of life (HRQOL). Several studies have confirmed its effectiveness in reducing pain and improving function in patients with OA.1,12-14,17,22 However, there is a large variation in treatment effectiveness.18 Our understanding of the variance in THA results is limited mostly to patient and implant-related factors.3,7,8,20 The role of variation in service delivery practice, such as waiting time, is unclear.
Waiting lists for THA are the longest among elective surgical procedures in many industrialized nations, and particularly in Canada.28 Although there are indications that patients experience deterioration of health status while waiting,15,21,24 controversy exists regarding the effect of waiting on postoperative outcomes. To date, observational studies have given conflicting results.11,21,23,24 A Dutch study found no effect of waiting time on 3-month or 1-year postoperative outcomes.24 An Ontario study investigated the effects of waiting on 99 patients who had THAs.21 This study found greater improvement in Western Ontario McMaster Universities Osteoarthritis Index (WOMAC) scores in patients who waited less than 6 months than in those who waited longer.21 A Swedish study found that a 3-month difference in waiting time did not result in a difference in postoperative outcomes in 124 patients who had THAs for OA.23 In the UK, a large prospective cohort study involving 7151 patients compared preoperative and postoperative HRQOL scores.11 This study showed a significant four-point difference in postoperative Oxford hip scores (range, 12-60 points) between patients waiting more than 1 year compared with those with shorter waits.11
Less is known regarding the effect of surgical delays on outcome at the individual level. Previous research on waiting time frequently has ignored preoperative scores and has concentrated on group-level measurements such as the mean of postoperative scores.11,21,23,24 Such methods do not provide the best evidence for explaining the association between postoperative outcomes and waiting time.6 The dependency of postoperative scores on preoperative scores has been indicated but not elaborated. Previous studies have shown that preoperative HRQOL is a strong predictor of postoperative outcome,7,8,20 and that function and pain in patients with lower preoperative function did not improve after surgery to the level achieved by those with higher preoperative scores.7,8 In methodologic terms, measures of postoperative HRQOL outcome need to be adjusted by preoperative functional status.
We hypothesized that longer waiting times are detrimental to the achievement of the full benefit of surgery. We secondarily hypothesized that postoperative HRQOL scores are strongly linked to their preoperative values and the preoperative values would be predictive of postoperative function.
MATERIALS AND METHODS
We included all patients with OA of the hip who were waiting for THAs in Vancouver General Hospital between March 2001 and May 2003. We defined a long wait greater than 6 months and a short wait as 6 months or less. Ethical approval was issued by the University of British Columbia Clinical Ethics Review Board.
Two hundred thirteen patients with OA of the hip treated by four orthopaedic surgeons agreed to join the study. All patients completed a WOMAC questionnaire at the time of consultation.2 Twelve patients were excluded later because they had a previous hip arthroplasty of the index joint or were having both hips replaced simultaneously, leaving 201 preoperative assessments eligible in the study. One year after THA, 147 (73%) patients returned a postoperative WOMAC questionnaire by mail. The mean age of these patients was 65 years with the ages ranging from the third to ninth decades (Table 1). Eighty three (53%) were women. The mean and median waiting times of the study population were both 6 months.
The WOMAC is the tool recommended for disease-specific outcome measures of hip and knee arthroplasties.2 It is a self-administered multidimensional index containing dimensions for pain (five items), stiffness (two items), and function (17 items). Each item is represented by a Likert scale response from 0 (best health state) to 4 (worst health state). Dimensions are equally weighted and reported as sums, the higher number indicating the greater burden of OA. The WOMAC is valid and reliable in patients with OA of the hip and knee.2 It is the most frequently used measure of pain and functional disability among patients who have had arthroplasties. The WOMAC is proven responsive to changes with time, which makes it suitable for subtle discrimination during waiting.
A two-step analysis was used in this study. First, a linear regression analysis of log-transformed postoperative WOMAC scores was used to determine the expected outcome for a given individual preoperative WOMAC score. Postoperative WOMAC function scores follow a skewed distribution because postoperative outcome is nearly as good as normal function; that is, postoperative outcomes have, as a floor limit, a score of 0 (best function). Because distribution of the postoperative WOMAC score is skewed, we transformed the postoperative scores on a logarithmic scale. Therefore, the dependent variable follows a symmetric distribution similar to the predictor (preoperative score) in this regression, as shown in the quantile-quantile plot (Fig 1). The following model was studied:
where Post is the postoperative WOMAC score and Pre is the preoperative WOMAC score. The error term ε follows a normal distribution with a mean of 0 and a standard deviation of 1, and σ is a fixed constant that changes the location of the expected value. The median of the postoperative WOMAC score for a certain preoperative WOMAC score was estimated. Because of the small sample size, we took the variation of the estimate into consideration. The lower boundary of the 90% confidence interval (CI) of the estimated median then was defined as the classification criterion. Patients below the lower boundary of the 90% CI of the median postoperative score were considered to have achieved a better than expected outcome (Fig 2). The performance of the classification criterion was evaluated using a Monte-Carlo simulation.27 The variation of the estimated median for a certain preoperative WOMAC score became smaller with increasing sample size.27 When the sample size is rather large (≥ 500), the variation of the estimated median is minimal, and then one can use the estimated median as the classification criterion.
In the second step, logistic regression models were used to study the effects of treatment delay (Model 2). The dependent variable was defined as a binary outcome (better than expected or not better than expected) derived from the first model. This binary variable represents the individual HRQOL outcome adjusted by the individual preoperative score. Waiting time was defined as the interval from the mutual decision between the patient and surgeon to proceed with surgery to the time of the operation. This is the definition recommended in the Institute for Clinical Evaluative Sciences Report, in the Western Canada Waiting List,9 and it is widely accepted in the literature.19 Therefore, each patient in this study had his or her waiting time calculated as the number of months from registration on the wait list until surgery. Waiting time was analyzed in the model as a continuous variable and a binary variable.
Age, gender, and comorbidity were entered in the model as independent variables to obtain adjusted effects. Comorbidity was characterized by the modified Charnley classification: A, single hip with OA; B1, bilateral hip arthritis; B2, previous THA on the contralateral hip; or C, multiple-joint arthritis or a chronic disease that affects HRQOL (specifically walking).
All data analyses were performed by SAS software 8.1 (SAS institute Inc, Cary, NC). A p value less than 0.05 was considered statistically significant.
Longer waiting times are detrimental to achieving full benefit on functional outcome of surgery. Forty-three percent of patients in the shorter waiting group achieved a better than expected functional outcome, whereas only 31% of patients in the longer waiting group attained a better than expected result. The power to detect such a difference among 147 subjects is 0.26. Waiting longer than 6 months results in a 50% decrease in the odds of achieving a better than expected outcome compared with those who waited less than 6 months. Each additional month spent waiting was associated with an 8% decrease in the odds of a better than expected functional outcome. The adjusted odds ratio of achieving a better than expected outcome is 0.92 for each month of waiting time (p = 0.05). Therefore, waiting time is negatively associated with the probability of better than expected outcome measured by the WOMAC functional subscale. No evidence of the negative effect of waiting time on outcome was found for WOMAC pain and stiffness. The long and short waiting groups had similar numbers of better than expected outcomes in the WOMAC pain and stiffness subscales (3% fewer in the long waiting group). Waiting longer did not impair the probability of achieving a better than expected pain or stiffness outcome. The power to detect such a difference among 147 subjects is only 0.05.
We also found that the preoperative WOMAC score was strongly associated with the postoperative WOMAC score in all subscales. The preoperative values are predictive of postoperative function. The postoperative WOMAC functional score will increase by 35% when the preoperative WOMAC functional score increases by 10 points (functional score range, 0-68; p = 0.0005). This predictive model is well fitted as attested by the goodness of fit statistics (p = 0.03). The mean (standard deviation) of preoperative WOMAC functional scores was 39 (13), whereas that of the postoperative WOMAC functional score was 14 (14). Thirty-seven percent of study patients achieved a better than expected functional outcome. Similarly, the postoperative WOMAC pain score will increase by 30% when the preoperative WOMAC pain score increases by three points (pain score range, 0-20; p = 0.0036). The postoperative WOMAC stiffness score will increase by 13% when the preoperative WOMAC stiffness score increases by one point (stiffness score range, 0-8; p = 0.0004). Age, gender, and comorbidity did not change the predictive effect of the preoperative score.
Prolonged waiting time may impair patient outcomes. This is biologically plausible because prolonging the arthritic process in these joints may result in muscle atrophy, tissue contractures, and deterioration of the general medical condition that may not be recoverable after surgery. There is some evidence of deterioration in quality of life during waiting,15 and there is clearer evidence showing that preoperative quality of life status is a strong predictor of postoperative outcome.7,8,20 It may be supposed that THA performed too late during the natural history of OA may lead to worse long-term outcomes. Our primary hypothesis is that prolonged waiting is a significant independent risk factor for reduced benefit after surgery.
There are several potential limitations in our study. Collection of the questionnaire is part of the routine administrative work in a busy clinical environment. The medical office assistant as much as possible collected surveys (213/668) from patients who arrived for first-time consultations. We compared the study patients with the patients with other disorders having 176 primary THAs with respect to age, gender, and waiting time and found similar distributions. Therefore, we think that there is no systematic selection bias existing in this setting. Another limitation of this study is that the response rate is rather low (147/201 or 73%). We compared the nonresponders with the responders with respect to age, gender, waiting time, and comorbidity and found similar distributions. A response bias might be supposed, considering that the responders and nonresponders had different preoperative scores (5.0 points worse in the nonresponders). But assuming that patients who do not respond tend to have poorer outcomes,16 the net effect would tend to underestimate the number of failures among the long-waiters. Therefore, we think that the potential response bias would not alter the direction of our result.
This study was performed at the Vancouver General Hospital, Vancouver, Canada. The demographics of our patients having arthoplasties are not different from elsewhere in western Canada (Table 1). According to statistics published by British Columbia's Ministry of Health, waiting times for patients at Vancouver General Hospital are comparable with those of other hospitals in British Columbia.4 Therefore we think the results are generalizable to other clinical settings.
Our results confirm previous reports that THA is an effective treatment for OA of the hip. Postoperative WOMAC scores showed significant clinical and statistical improvement. When the aim of comparing treatments is to show a marginal but significant difference, the research must define in advance a meaningful clinical change in patient status. A consensus on the design of clinical drug trials determined a 9.3-point change in normalized WOMAC functional scores as a minimal clinical improvement.5 But this change is too small to be applied to the outcomes of THAs, which typically show a 60-100% improvement over preoperative WOMAC scores.1,8 The expected change in WOMAC functional scores after THA is four times larger than the minimal clinically important difference derived from drug trials in OA. For treatment as successful as THA, this marginal difference is inadequate for documenting the positive impact of treatment.
Our results agree with previously published results in that postoperative HRQOL scores were found to be strongly associated with their preoperative values.7,8 By using the expected postoperative WOMAC score for any given preoperative score, we were able to differentiate patients who achieve the maximal benefit from THA from those who do not. Using this method, clinicians and researchers would have an intuitive standard to assess whether patients had achieved an undeniably satisfactory result. This approach also could be used to investigate other predictors of outcome after THA.
In previous studies concerned with access time to THA, authors have reported the effects of waiting in terms of change in HRQOL score at the group level.23,24 The evidence supported by various prospective cohort studies is contradictory (Table 2).7,8,11,20,23,26 For example, the Swedish study showed that long wait times are associated with failure to reach full recovery,23 whereas the Dutch study found no direct effect of waiting time on postoperative outcome.24 Patterns of individual change in HRQOL before and after THA have not been described before.
We classified patients according to the magnitude of individual change in HRQOL to compare interindividual changes with waiting times. Measurement of individual changes is an increasingly attractive method of quantifying HRQOL outcomes because it can objectively document the impact of treatment. A relevant methodologic paper by Hageman and Arrindell draws attention to the distinction between individual and group perspectives on assessments of change.10 In a review on determining the clinical significance of change in health status, Samsa et al state that “even groups with negligible mean changes in HRQOL scores are likely to contain individual patients whose improvement is noteworthy”.25 Our method of determining expected results is somewhat complicated in that it involves left-censoring and a log transformation. It is a way of coping with the ceiling effect and skewed distribution characteristic of WOMAC measurements. If there were a disease-specific HRQOL instrument designed for THA that did not display the ceiling effect and skewing, the determination of expected results could be achieved easily through a general linear regression or by simply stratifying the preoperative scores and classifying the postoperative scores by strata.
Our results provide evidence that there is an association between waiting time and the probability of a better than expected postoperative functional outcome. The odds of achieving a better than expected postoperative functional outcome decrease by 8% for each month on the waiting list. Patients who waited more than 6 months had 50% lower odds of a better than expected functional outcome than patients with shorter waits. The decrease in the proportion of patients with better than expected functional outcomes among patients with longer waits as compared with patients with shorter waits indicates that longer waiting may reduce the chance of maximizing surgical benefits.
We found that prolonged waiting time for an elective THA reduces the benefit in HRQOL gained after the operation as measured by WOMAC score. We examined postoperative hip scores adjusted by the preoperative score and explored factors affecting outcomes after THA at the individual level rather than at the group level. Results of our study support those of previous research that waiting for a THA can adversely affect final outcome after THA. Findings of our study are based on the probability of individual improvement as estimated by a regression model developed within the study population. Future studies should take HRQOL immediately before the surgery as an end point. The waiting time between referral from a family doctor until the orthopaedic consultation rarely is described. Deterioration during this waiting period needs to be included in future studies.
We thank Francisco Luna for data collection, James Latteier for data collection and proofreading the manuscript, and Dr. Nelson Greidanus who participated in recruiting patients.
1. Bachmeier CJ, March LM, Cross MJ, Lapsley HM, Tribe KL, Courtenay BG, Brooks PM. Arthritis Cost and Outcome Project Group. A comparison of outcomes in osteoarthritis patients undergoing total hip and knee replacement surgery. Osteoarthritis Cartilage
2. Bellamy N, Kirwan J, Boers M, Brooks P, Strand V, Tugwell P, Altman R, Brandt K, Dougadoa M, Lequesne M. Recommendations for a core set of outcome measures for future phase III clinical trials in knee, hip, and hand osteoarthritis. Consensus development at OMERACT III. J Rheumatol
3. Braeken AM, Lochhaas-Gerlach JA, Gollish JD, Myles JD, Mackenzie TA. Determinants of 6-12 month postoperative functional status and pain after elective total hip replacement. Int J Qual Health Care
4. British Columbia Ministry of Health. Surgical Wait List Registry for BC
. Vancouver, BC, Canada: British Columbia Ministry of Health; 2005.
5. Ehrich EW, Davies GM, Watson DJ, Bolognese JA, Seidenberg BC, Bellamy N. Minimal perceptible clinical improvement with the Western Ontario and McMaster Universities osteoarthritis index questionnaire and global assessments in patients with osteoarthritis. J Rheumatol
6. Fitzpatrick R, Norquist JM, Dawson J, Jenkinson C. Rasch scoring of outcomes of total hip replacement. J Clin Epidemiol
. 2003;56: 68-74.
7. Fortin PR, Clarke AE, Joseph L, Liang MH, Tanzer M, Ferland D, Phillips C, Partridge AJ, Belisle P, Fossel AH, Mahomed N, Sledge CB, Katz JN. Outcomes of total hip and knee replacement: preoperative functional status predicts outcomes at six months after surgery. Arthritis Rheum
8. Fortin PR, Penrod JR, Clarke AE, St-Pierre Y, Joseph L, Belisle P, Liang MH, Ferland D, Phillips CB, Mahomed N, Tanzer M, Sledge C, Fossel AH, Katz JN. Timing of total joint replacement affects clinical outcomes among patients with osteoarthritis of the hip or knee. Arthritis Rheum
9. Hadorn DC. Setting priorities for waiting lists: defining our terms. Steering Committee of the Western Canada Waiting List Project. CMAJ
10. Hageman WJ, Arrindell WA. Establishing clinically significant change: increment of precision and the distinction between individual and group level of analysis. Behav Res Ther
11. Hajat S, Fitzpatrick R, Morris R, Reeves B, Rigge M, Williams O, Murray D, Gregg P. Does waiting for total hip replacement matter? Prospective cohort study. J Health Serv Res Policy
12. Harris WH, Sledge CB. Total hip and total knee replacement (1). N Engl J Med
13. Harris WH, Sledge CB. Total hip and total knee replacement (2). N Engl J Med
14. Jones CA, Voaklander DC, Johnston DW, Suarez-Almazor ME. Health related quality of life outcomes after total hip and knee arthroplasties in a community based population. J Rheumatol
15. Kelly KD, Voaklander DC, Johnston DW, Newman SC, Suarez-Almazor ME. Change in pain and function while waiting for major joint arthroplasty. J Arthroplasty
16. Kim J, Lonner JH, Nelson CL, Lotke PA. Response bias: effect on outcomes evaluation by mail surveys after total knee arthroplasty. J Bone Joint Surg
17. Laupacis A, Bourne R, Rorabeck C, Feeny D, Wong C, Tugwell P, Leslie K, Bullas R. The effect of elective total hip replacement on health-related quality of life. J Bone Joint Surg
18. Liang MH, Cullen KE, Poss R. Primary total hip or knee replacement: evaluation of patients. Ann Intern Med
19. Llewellyn-Thomas HA, Arshinoff R, Bell M, Williams JI, Naylor CD. In the queue for total joint replacement: patients' perspectives on waiting times. Ontario Hip and Knee Replacement Project Team. J Eval Clin Pract
20. MacWilliam CH, Yood MU, Verner JJ, McCarthy BD, Ward RE. Patient-related risk factors that predict poor outcome after total hip replacement. Health Serv Res
21. Mahon JL, Bourne RB, Rorabeck CH, Feeny DH, Stitt L, Webster-Bogaert S. Health-related quality of life and mobility of patients awaiting elective total hip arthroplasty: a prospective study. CMAJ
22. NIH Consensus Conference. Total hip replacement. NIH Consensus Development Panel on Total Hip Replacement. JAMA
. 1995;273: 1950-1956.
23. Nilsdotter AK, Lohmander LS. Age and waiting time as predictors of outcome after total hip replacement for osteoarthritis. Rheumatology (Oxford)
24. Ostendorf M, Buskens E, van Stel H, Schrijvers A, Marting L, Dhert W, Verbout A. Waiting for total hip arthroplasty: avoidable loss in quality time and preventable deterioration. J Arthroplasty
. 2004;19: 302-309.
25. Samsa G, Edelman D, Rothman ML, Williams GR, Lipscomb J, Matchar D. Determining clinically important differences in health status measures: a general approach with illustration to the Health Utilities Index Mark II. Pharmacoeconomics
26. Williams JI, Llewellyn TH, Arshinoff R, Young N, Naylor CD. The burden of waiting for hip and knee replacements in Ontario. Ontario Hip and Knee Replacement Project Team. J Eval Clin Pract
27. Xu M, Garbuz DS, Kuramoto L, Sobolev B. Classifying health-related quality of life outcomes of total hip arthroplasty. BMC Musculoskelet Disord
28. Zelder M, Wilson G. Waiting Your Turn: Hospital Waiting Lists in Canada
ed. Vancouver, BC, Canada: The Fraser Institute; 2000.