After Total Hip Arthroplasty Comparison of a Traditional Disease-specific and a Quality-of-life Measurement of Outcome

The Journal of Arthroplasty Vol. 12 No. 6 1997 Outcome After Total Hip Arthroplasty Comparison of a Traditional Disease-specific and a Quality-of-life Measurement of Outcome Jay R. Lieberman, MD,* Frederick Dorey, PhD,* Paul Shekelle, MD, PhD,]- Lana Schumacher, BS,* Douglas J. Kilgus, MD,* Bert J. Thomas, MD,* and Gerald A. Finerman, MD,* Abstract: The purpose of this study was to examine the relationship between the Harris Hip Score (HHS), a traditional method of patient assessment of a total hip arthroplasty (THA), and the Medical Outcomes Study 36-Item Short-Form Health Survey (SF-36), a commonly used health-related quality-oi-life survey. One hundred forty patients returning for routine clinical follow-up evaluation of a primary THA were asked to fill out the SF-36 quality-of-life survey, as well as questions concerning their perceptions of their THA. The patient's surgeon assessed the THA with the traditional HHS. The correlations between the HHS and the SF-36 domains were highest in the physical component summary scores for male patients of all ages and female patients 65 years of age or older. The correlations were lower for the mental component summary scores of all patients, but particularly in female patients younger than 65. When the SF-36 scores were compared with age and sex-matched population norms, both age and sex were found to be important. Men younger than 65 had scores lower than norms in the physical function domains, but were comparable in the mental health domains. The older men had scores comparable to the norms in all domains. Female patients of all ages, however, had lower scores in the physical function domains. The greatest differences were noted in the female patients younger than 65. The HHS is commonly used to assess disease-specific pain and function in THA patients; however, the results of this study suggest that the SF-36 health survey can capture additional important quality-oflife domains that are influenced by a THA and that these domains are influenced by the age and sex of the patient. The combination of a disease-specific scoring system and a quality-of-life survey would allow a more global assessment of a THA in all patients. Studies evaluating the results of THAs should either assess the results of male and female patients separately when sample size is sufficiently large or use sex as a possible covariate in a multivariate analysis. Key words: Harris Hip Score, SF-36, quality of life, total hip arthroplasty. From the *Department of Orthopaedic Surgery, University of California at Los Angeles Medical Center, Los Angeles, and the ~-Departmerit of Medicine, University of California at Los Angeles Medical Center and West Los Angeles Veterans Administration Medical Center, Los Angeles, California. Reprint requests: Jay R. Lieberman, MD, Department of Orthopaedic Surgery, CHS 76-134, UCEA Medical Center, 10833 Le Conte Avenue, Los Angeles, CA 90095. 1997 Churchill Livingstone Inc. The effectiveness of a total hip arthroplasty (THA) in relieving pain and improving function has been well documented over the past 20 years [1-4]. It is apparent that quality of life is an important outcome valued by patients, and the influence of surgical procedures on quality of life can be measured [5]. Health-related quality of life encompasses not only pain and physical functioning, but other 639

640 The Journal of Arthroplasty Vol. 12 No. 6 September 1997 domains such as social functioning, mental health, vitality, and general health. In addition, orthopaedic surgeons are no longer interested only in evaluating patient function before and after THA; more subtle comparisons between 2 potentially efficacious treatments (ie, results of cement versus cementless hips) are now required. Therefore, the use of instruments that have increased sensitivity and specificity in evaluating quality of life compared with traditional scoring systems may enhance our ability to assess overall outcome in THA patients. Orthopaedic surgeons have traditionally used disease-specific scoring systems that focus on relief of pain and improvement in function when assessing a total joint arthroplasty; however, the assessment of the impact a medical or surgical intervention has on quality of life is of considerable importance in an age where resources are limited. The evaluation of a medical or surgical intervention should assess its effect on all aspects of health-related quality of life, as various aspects of quality of life may not be affected to the same degree or even in the same manner by a particular procedure. The Harris Hip Score (HI-IS) is a commonly used physician assessment of localized pain and physical functioning [6]. Patients are scored on a 0-100 scale based on the degree of pain, function, and range of motion. Although this scoring system is quite popular among orthopaedic surgeons, it has never been validated psychometrically. The Medical Outcomes Study 36-Item Short-Form Health Survey (SF-36) is a commonly used general health-related survey that measures 8 domains of quality of life [7]. In addition, the results of the 8 domains have been combined into 2 summary scales, 1 for physical functioning and 1 for mental health [7]. The SF-36 has been shown to discriminate between patients at different stages of disease for a variety of different disease conditions including hypertension [8], coronary artery disease [9], and acquired immune deficiency syndrome [7]. There are also data comparing the results ot the SF-36 tests with normal individuals of different age and sex groups [10]. This allows for the comparison of THA patients with population norms [11,12]. The purpose of this study was to examine the relationship between the commonly used disease-spedfic HI-IS and a commonly used general healthrelated quality-of-life survey, the SF-36, in assessing the results of a THA [10-14]. In addition, the SF-36 scores were compared with published data on age and sex-matched norms [10,11,14]. Materials and Methods One hundred forty-four patients returning for routine clinical follow-up evaluation for their primary THA were asked to fill out a questionnaire that contained the SF-36 quality-of-life survey and specific questions regarding their perceptions about their THA. The results of the surgery were evaluated by the patient's surgeon using a more traditional method of assessment of a THA, the HI-IS [6]. The patients filled out the questionnaire in the waiting room prior to seeing the physician. The patients were informed that their physician would not be reviewing the questionnaire. The SF-36 health survey is a psychometrically validated questionnaire that measures 8 domains of quality of life: physical functioning, role physical, role emotional, social functioning, bodily pain, mental health, vitality, and general health. All domains are scored separately on a 0-100 scale, with higher numbers representing a more favorable health status. Recently, the 8 domains of the SF-36 have been combined in a linear fashion to form a physical and mental health summary score [14]. First, each of the 8 SF-36 domain scores is standardized by subtracting the mean of the general population and dividing by the standard deviation. This transtorms the scores for the general population to have a mean of 0 and a standard deviation of 1. Second, each standardized score is multiplied by an appropriate number or weight, based on a multivariate factor analysis (coefficients equal to factor loadings), to obtain an aggregate physical component score and a mental health component score. The physical component summary score consists of a weighted average of the 8 domains, with positive weights assigned to physical function, role physical, bodily pain, general health, and vitality, and negative weights assigned to the social function, role emotional, and mental health transformed scores. The mental health component summary score assigns positive weights to vitality, social function, role emotion, and mental health scores and negative weights to physical function, role physical, bodily pain, and general health. Finally, each of the summary scores is multiplied by 10 and added to 50. This results in a physical health and mental health summary score with an expected mean of 50 and a standard deviation of 10 in the general population. In addition, the patients responded to a series of Likert questions regarding their level of function, satisfaction with the procedure, willingness to have the surgery again, effect of the THA

Measurement of Outcome AfterTHA Lieberman et al. 641 surgery on quality of life, and the need for pain medication. Statistical Analysis Correlations between the HHS and the SE-36 dimensions were evaluated using the nonparamettic Kendall Tau correlation. Multivariate linear regression was used to evaluate the relationships between the outcome scores, with age, sex, number of comorbidities, and length of follow-up period considered as possible covariates. The relationship between patient responses to subjective evaluations and the outcome scores was made using t-tests or nonparametric tests as appropriate. Each patient was matched to his or her expected U.S. population norm (n = 2474) based on age, decade, and sex. The differences between patient scores and population-based norms were analyzed as above using linear regression analysis. In addition, comorbid conditions were coded as cardiac, vascular, pulmonary, renal, gastrointestinal, and diabetes. Patients were stratified based on the number of comorbidities (none, 1, more than 1). Results The mean age at the time of surgery was 58 years (range, 17-86) and the mean follow-up period was 57 months (range, 6-261 months). There were 59 men and 85 women. The primary diagnosis was osteoarthritis in 83 patients (57%), avascular necrosis in 25 patients (17%), rheumatoid arthritis in 18 patients (12%), post-traumatic arthritis in 6 patients (4%), developmental dysplasia of the hip in 4 patients (3%), slipped capital femoral epiphysis in four patients (3 %), and Legg-Calv~-Perthes disease, Gaucher's disease, Paget's disease, and a fractured femoral head in 1 patient each. When asked to indicate the 2 major reasons for having the surgery, pain was listed 89% of the time, difficulty with ambulation 39%, and limited hip mobility 35%. Eighty-one percent of patients indicated they would be very likely to have the surgery again, and 73% considered their quality of life to be substantially improved since having the surgery. Sixty-eight percent of patients considered their work capacity or activity level to be improved since the surgery. Thirty-seven percent of the patients participated in some type of sports activity at least once per week. Thirty-two percent of patients were limited in their daily activity and 7% were unemployed because ot health problems. The mean HHS was 85 (range, 26-100). The mean score for men was 89 (range, 44-100) and the mean score for women was 84 (range, 26-100). The following were the overall mean scores (of a possible 100) for the 8 domains of the SF-36: physical functioning, 56; role physical, 53; role emotional, 79; social functioning, 78; bodily pain, 54; mental health, 73; vitality, 57; general health, 74. The summary scores (of a possible 80) were 39 for physical function and 53 for mental health. Thus, the study population is approximately one standard deviation below the national average in physical health and the same as the national average in mental health. The willingness of the patient to have surgery again was significantly associated with high scores in social functioning (P <.003), general health (P <.007), and role emotional (P <.001); however, bodily pain (P <.253), physical functioning (P <.062), and role physical (P <.181) were not significantly associated with willingness to have surgery again. Correlations between the Harris Hip Score and Medical Outcomes Study 36-Item Short-Form Health Survey The Kendall Tau correlations for the entire group (n = 144) as well as patients with no comorbidities (n = 62) are listed in Table 1. The correlation patterns are similar in the entire population, as well as the population with no comorbidities. The relationship between the HHS and the SF-36 is strongest in the physical domains, including physical function and bodily pain, and weakest in the mental health domains, including role emotional and mental health. Thus, in THA patients, the disease-spedfic HHS captures much of the information contained in a general quality-of-life evaluation with respect to pain and physical functioning, but very little of the information contained in other aspects of quality of life. A comparison of the mean scores for THA patients for the 8 domains of the SF-36 with the scores for published norms by sex and age [12] revealed that men over the age of 65 had scores comparable to published norms (adjusted for the age distribution of the patients) in all dimensions of the SF-36 (Table 2). Men younger than 65, however, scored lower than published norms with respect to physical functioning, role physical, and bodily pain. In contrast, female patients in all age groups consistently scored lower than the published norms in physical functioning, role physical, and bodily pain (Table 3). The linear regression analysis of the differences between patient scores and their corresponding

642 The Journal of Arthroplasty Vol. 12 No. 6 September 1997 Table 1. Kendall Tau Correlation With Harris Hip Score No Comorbidities (n = 62) Entire Group (n = 144) Physical functioning 0.446 0.516 Role physical 0.253 0.336 Role emotional 0.101 0.207 Social functioning 0.245 0.375 Bodily pain 0.458 0.385 Mental health 0.087 0.233 Vitality 0.386 0.423 General health 0.200 0.263 Summary scores Physical function 0.489 0.506 Mental health 0.037 0.185 Table 2. Comparison of SF-36 Scores With Published Norms Matched by Sex and Age: Men Age < 65 y (n = 31) Age_> 65 y (n = 29) Patients Norms Patients Norms Physical functioning 68 (52-75)* 85 63 (57-78) 66 Role physical 64 (56-90) 84 71 (46-80) 60 Role emotional 89 (76-100) 84 89 (78-98) 77 Social functioning 86 (76-95) 85 85 (76-96) 80 Bodily pain 63 (57-80) 74 67 (54-74) 69 Mental health 79 (73-87) 76 80 (72-85) 77 Vitality 67 (56-72) 64 64 (57-76) 58 General health 79 (67-87) 72 76 (72-86) 59 Summary scores Physical functioning 44 (39-49) 50 44 (39-49) 42 Mental health 56 (53-59) 51 57 (52-59) 53 *The 95% confidence intervals are listed in parentheses. SP-36, Medical Outcomes Study 36-Item Short- Form Health Survey. Table 3. Comparison of SF-36 Scores With Published Norms Matched by Sex and Age: Women Age<65y (n=30) Age _> 65 y (n = 54) Patients Norms Patients Norms Physical functioning 50 (41-58)* 81 48 (39-60) 62 Role physical 45 (31-55) 78 41 (27-62) 56 Role emotional 68 (64-86) 80 73 (52-86) 73 Social functioning 74 (62-80) 81 69 (60-86) 77 Bodily pain 45 (38-55) 71 46 (33-57) 63 Mental heatth 70 (63-75) 73 69 (63-78) 75 Vitality 51 (44-58) 59 50 (41-60) 55 General health 70 (65-77) 69 71 (63-78) 62 Summary scores Physical functioning 36 (32-39) 48 35 (30-41) 41 Mental health 51 (48-55) 49 50 (46-56) 51 *The 95% confidence intervals are listed in parentheses. SF-36, Medical Outcomes Study 36-Item Short- Form Health Survey. population norms' physical and mental health summary scores revealed that female and male patients younger than 65 had physical component summary scores that were significantly lower (P =.005) than the expected values according to the data for U.S. population norms (Tables 2, 3). Evaluation of the mental health summary scores revealed that there was no significant difference in mental health summary scores for male and female patients of all ages.

Measurement of Outcome After THA Lieberman et al. 643 Comorbidities The following comorbidities were observed: cardiac, 22; pulmonary, 6; vascular, 11; gastrointestinal, 4; diabetes, 5; renal, 2. Patients were categorized as having none (n = 62), 1 (n = 70), or more than 1 (n = 15) comorbidity. Both male and female patients of all ages with multiple comorbidities had lower HHS scores (P <.002) and lower SF-36 scores (P <.05). Both the physical (P =.001} and mental health (P =.005) summary scores were significantly lower in patients with multiple comorbidities. Discussion Earlier studies have demonstrated a dramatic short-term increase in most dimensions of the SF- 36 over presurgery values [15]. The SF-36 allows one to examine 8 different domains related to quality of life rather than just pain and function. This study relates the SF-36 scores to the diseasespecific HHS at various times after surgery. Overall, there were good correlations between the HHS and the SF-36 with respect to the physical component summary score in both men and women (Figs. 1, 2); however, the correlations were.9 Fig. I. Correlations between the Harris Hip Score and the.8 8 Medical Outcomes Study 36-Item Short-Form Health.7 Survey (SF-36) domains and the physical (PCS) and men-.6 tal health (MCS) component P.5 summary scores for the male patients. PE physical func-.4 tioning; RE role physical; RE, role emotional; SE social.3 functioning; BE bodily pain; MH, mental health; VT, vital-.2 ity; GH, general health..1 i 0 PF RP RE SF BP MH VT GH PCS MCS Fig. 2, Correlations between the Harris Hip Scores and the 8 Medical Outcomes Study 36-Item Short-Form Health Survey (SF-36) domains and the physical (PCS) and mental health (MCS) component summary scores for the female patients. PF, physical functioning; RE role physical; RE, role emotional; SF, social functioning; BE bodily pain; MH, mental health; VT, vitality; GH, general health. P i.9.8.7.6.5.4.3.2.i 0 PF RP RE SF BP MH VT GH PCS MCS

644 The Journal of Arthroplasty Vol. 12 No. 6 September 1997 mixed between the HHS and mental health component summary scores, particularly in patients younger than 65. In these cases, all the patients were satisfied with their overall surgical results of the THA, but there were still limitations in their overall quality of life. Thus, a high HHS may not accurately reflect excellent quality of life in all patients. Additionally, it is interesting to note that the patient's willingness to have surgery again was not highly correlated with either a decrease in bodily pain or an improvement in physical function. There was, however, a significant correlation between willingness to have surgery again and the mental health domains of the SF-36 (role emotional, mental health, vitality, and general health). This finding suggests that there are a variety of factors that are important to THA patients and that a general health survey should be combined with a disease-specific scoring system to enhance the assessment of the outcome of a THA. All patients with a minimum follow-up period of 6 months were included in this study. We recognize that studies evaluating THA patients traditionally have a minimum 2-year follow-up period, but we did not believe this was necessary for this study because we were assessing the tools used to assess a THA rather than the success of the procedure itself. The data revealed no association between length of follow-up period and the relationship between SF-36 and HHSs. Furthermore, as this is a cross-sectional study, we did not attempt to assess the efficacy of ditferent implants. An analysis based on patient etiology revealed that patients with osteoarthritis had higher HHS and SF-36 summary scores than patients with other etiologies, even after adjusting for sex and age. The correlations (r =.60) between the HHS and SF-36 summary scores were similar for the other etiologies (r = 0.60), except for patients with rheumatoid arthritis (r =.42). The rheumatoid arthritis patients had HHS scores similar to those of the nonosteoarthritis patients, but the SF-36 functional summary scores were substantially lower. A long-term prospective study is necessary to determine definitively the relationship between etiology or type of prosthesis and changes in quality-of-life outcomes. Generally, normal women tend to score lower than normal men on the SF-36, and we also noted this trend in our patient data. The reasons for this trend are not clear. The male patients 65 years of age and older had scores comparable to the published norms in all dimensions, whereas men younger than 65 scored lower on physical functioning, role lunctioning, and bodily pain than agematched norms. This finding is not surprising, as younger individuals are usually more active than older individuals; however, the male patients younger than 65 had scores comparable to the published norms with respect to the mental health domains (vitality, role emotional, mental health, and social functioning). The female patients scored lower than the norms in all categories. In general, orthopaedic surgeons have assumed that a high HHS is associated with excellent quality of life. The HHS was designed to assess improvements in hip pain and function, and our data suggest that it does adequately assess general pain and tunction in patients over 65 years of age. Our data also suggest that the HHS may not adequately address a variety of quality-of-life issues that may be important to our patients. The use of a patient self-administered questionnaire such as the SF-36 can allow one to evaluate important quality-of-life domains that may be influenced by a THA; however, it is important to recognize that the SF-36 alone is not sufficient to properly analyze the results of a THA. The SF-36 is a general health survey and is not specifically designed to assess patients with arthritic conditions. McGuigan et al. noted that the SF-36 may not evaluate important quality-of-life issues for an individual patient [ 16]. The results of this study demonstrate that a variety of issues must be considered if one is to accurately assess the results of a THA. When multivariate statistical modeling is performed, age, sex, and comorbidity status should be considered as possible covariates. The effect that a THA has on a patient's overall quality of life is critical to assess. The use of a variety of instruments may enable orthopaedic surgeons to have a better understanding of the impact that medical (ie, disease type and severity) and social (ie, social supports and social functioning) issues have on the overall outcome of a THA. References 1. Kavanaugh BE, Dewitz /VIA, Illstrup DM et al: Charnley total hip arthroplasty with cement: fifteen Year Result. J Bone Joint Surg 71A:1496, 1989 2. Mulroy RD Jr, Harris WH: The effect of improved cementing techniques on component loosening in total hip replacement: an eleven year radiographic review. J Bone Joint Surg 72B:757, 1960 3. Wiklund I, Romanus B: A comparison of quality of life before and after arthroplasty in patients who

Measurement of Outcome AfterTHA Lieberman et al. 645 had arthrosis of the hip joint. J Bone Joint Surg 73A:765, 1991 4. Wroblewski BM: 15-21 year results of the Charnley low-friction arthroplasty. Clin Orthop 211:30, 1986 5. O'Boyle CA, McGee H, Hickey A et al: Individual quality of life in patients undergoing hip replacement. Lancet 339:1088, 1992 6. Harris WH: Traumatic arthritis of the hip after dislocation and acetabular fractures: treatment by mold arthroplasty. An end-result study using a new method of result evaluation. J Bone Joint Surg 51A:737, 1969 7. Wu AW, Rubin HR, Mathews WC et al: A health status questionnaire using 30 items from the Medical Outcomes Study: preliminary validation in persons with HIV infection. Med Care 29:786, 1991 8. Croog SH, Levine S, Testa MA et al: The effects of antihypertensive therapy on the quality of life. N Engl J Med 314:1657, 1986 9. Cleary PD, Epstein AM, Oster G et al: Healthrelated quality of life among patients undergoing percutaneous transluminal coronary angioplasty. Med Care 29:939, 1991 10. Ware JE Jr: SF-36 health survey manual and interpretation guide. The Health Institute, New England Medical Center, Boston, 1993 i 1. Jenkinson C, Coulter A, Wright L: The short form 36 (SF-36) health survey questionnaire: Normative data for adults of working age. Br Med J 306:1437, 1993 12. Liang MMH, Fossel AH, Larson MG: Comparisons of five health status instruments for orthopaedic evaluation. Med Care 28:632, 1990 i3. McHomey CA, Ware JE, Lu JFR, Sherbourne CD: The Mose 36-item short form health survey (SF- 36): III. Tests for data quality, scaling assumptions, and reliability across diverse patient groups. Med Care 32:40, 1994 14. Ware JE Jr: SF-36 physical and mental health summary scales: a user's manual. The Health Institute, New England Medical Center, Boston, 1994 15. Katz J, Larson M, Phillips C et al: Comparative measurement sensitivity of short and longer health status instruments. Med Care 30:991, 1992 16. McGuigan FX, Hozack WJ, Moriarty Let al: Predicting quality of life outcomes following total joint arthroplasty. J Arthroplasty 10:742, 1995