G Model
ARTICLE IN PRESS
HEAP-3205; No. of Pages 6
Health Policy xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Health Policy journal homepage: www.elsevier.com/locate/healthpol
Relationship between peri-operative outcomes and hospital surgical volume of total hip arthroplasty in Japan Takeshi Kaneko a,1 , Kazuo Hirakawa b,2 , Kiyohide Fushimi a,∗ a Department of Health Policy and Informatics, Tokyo Medical Dental University, Graduate School, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8519, Japan b Shonan Kamakura Joint Reconstruction Center, 5-4-17 Dai Kamakura, Kanagawa 247-0061, Japan
a r t i c l e
i n f o
Article history: Received 26 June 2013 Received in revised form 10 February 2014 Accepted 24 March 2014
Keywords: Total hip arthroplasty Hospital surgical volume Peri-operative outcome Complication Quality of care
a b s t r a c t Background: This study aimed to evaluate the relationship between peri-operative complication of total hip arthroplasty (THA) and hospital surgical volume. Methods: We reviewed discharge administrative data from 8321 patients who underwent primary THA between July and December 2008. Relationships between complications and surgical volume were analyzed with multivariate logistic regression models including age, sex, and Charlson comorbidity index. Hospitals were categorized into four groups according to the 6-month volume of THA procedures. Results: The most frequent complication was dislocation (1.41%). Next was infection (1.24%). Fracture and pulmonary embolism occurred in less than 1% of procedures. Patients who underwent THA in hospitals with the highest surgical volume had lower risk of dislocation and infection than those treated in the hospitals with the lowest surgical volume (odds ratio [OR] 0.321, 95% confidence interval [CI] 0.167–0.572 and OR 0.123, 95% CI 0.020–0.421). Patients aged 65 years and older were associated with increased risk of dislocation (OR 2.342, CI 1.555–3.624) and fracture (OR 2.799, CI 1.372–6.301). Females demonstrated lower risk of dislocation (OR 0.558, CI 0.352–0.869) and infection (OR 0.560, CI 0.365–0.882). Conclusion: These results indicated that the increase in the risk of peri-operative dislocation of primary THA may be associated with low hospital surgical volume as well as age and male sex. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Osteoarthritis (OA) is the most common joint disorder affecting the elderly population worldwide and a leading cause of pain and disability. OA is estimated to affect 40% of people over 70 years of age [1], and 25.3% of peo-
∗ Corresponding author. Tel.: +81 358034025; fax: +81 358030357. E-mail addresses: [email protected] (T. Kaneko), [email protected] (K. Hirakawa), [email protected] (K. Fushimi). 1 Tel.: +81 358034025; fax: +81 358030357. 2 Tel.: +81 467472377; fax: +81 467472370.
ple develop symptomatic hip OA by age 85 years [2]. Pain caused by OA is a major factor affecting quality of life. Total hip arthroplasty (THA) is a highly effective treatment to improve mobility and eliminate pain [3,4]. As people live longer, the demand of joint replacement will be necessary in more cases. However, complications of THA can lead to poor functional outcome for patients; thus, reducing their occurrence is important. The relationship between the rate of postoperative complications and hospital surgical volume or socioeconomic factors has been unclear. Several studies have indicated that hospitals with a high surgical volume have a lower rate of overall mortality [5–8]. Among 76,627 Medicare patients
http://dx.doi.org/10.1016/j.healthpol.2014.03.013 0168-8510/© 2014 Elsevier Ireland Ltd. All rights reserved.
Please cite this article in press as: Kaneko T, et al. Relationship between peri-operative outcomes and hospital surgical volume of total hip arthroplasty in Japan. Health Policy (2014), http://dx.doi.org/10.1016/j.healthpol.2014.03.013
G Model HEAP-3205; No. of Pages 6
ARTICLE IN PRESS T. Kaneko et al. / Health Policy xxx (2014) xxx–xxx
2
who underwent primary or revision THA, a significant association was observed between higher hospital volume and lower 90-day mortality and dislocation rates [9]. Additionally, postoperative complications and hospital surgical volume were associated with total length of hospital stay in 1561 total joint arthroplasty patients [10]. By contrast, several studies have indicated no association between THA complication rate and surgical volume. Using data from 2965 patients, Yasunaga et al. reported that surgeon volume of over 500 procedures was inversely associated with postoperative THA complications, while hospital surgical volume was not significantly associated with postoperative THA complications [11]. Katz et al. also reported that race and income were associated with complication risk in the Medicare population [9]. The purpose of this study was to assess the relationship between peri-operative complications after THA and hospital surgical volume with large-scale nation-wide patient database and with minimal effects of socio-economic status and patient ethnicities. We have employed the DPC (Diagnosis Procedure Combination) database, which is a large-scale nationally representative administrative database of acute care inpatients in Japan and has successfully been used for many clinical epidemiological studies [12–14]. In Japanese setting, the influence of race and socioeconomic factors is negligible because of the universal medical insurance system, which provides almost unrestricted access to health care [15], and ethnic homogeneity [16]. 2. Patients and methods 2.1. Data source and study population The DPC code-based prospective payment scheme consists of 18 major diagnostic categories, 507 diagnostic groups, and 2658 case-mix groups in acute-care hospitals. This diagnosis system was instituted in 2003 by the Ministry of Health, Labour and Welfare of Japan. In this study, we used the DPC database consisting of patient level administrative data including age, sex, inhospital mortality, length of stay, and hospital reference cost calculated from fee-for-service tariff schedule. Preoperative comorbidities and postoperative complications that occurred during hospitalization were coded with the International Classification of Diseases, Tenth Revision (ICD-10) codes [17]. Additionally, surgical procedures were coded according to a Japanese fee-for-service tariff schedule. We reviewed administrative data from 8607 patients who underwent primary THA between July and December 2008 in the Japanese DPC database. The data were voluntarily offered to us by 855 hospitals that agreed to their use for research, which represent approximately 45% of inpatient admissions to acute-care hospitals in Japan. The data were anonymous and could not be linked to any other information to identify patients upon their collection by the research group. Prior approval by the ethics committee of Tokyo Medical and Dental University was granted to the study. We excluded 161 patients who underwent THA due to fracture (ICD-10, S72), infection (ICD-10, M00-03, M46, M86, M90.1, M90.2), and pathologic fracture (ICD-10
C40, C41, C76.3, C76.5, C79.5, D48.0, M48.5, M49.5, M80, M84.3, M84.4, M90.7) to select elective surgery cases to minimize case-mix heterogeneity. A total of 125 patients were excluded from one hospital where some inpatients can selectively be treated with non-prospective payment schema. Our final analysis included 8321 patients from 601 hospitals. A previously published coding algorithm was modified [9,18,19], and we defined outcome as the peri-operative complication rate. Complications were defined as dislocation (ICD-10, M24.3, M24.4, S73.0 and requiring repositioning of joint dislocation), infection (ICD-10, M00, M01, M86, T81.4, T84.5-9, T85.7 and surgical debridement for pyogenic arthritis, removal of the prosthesis), pulmonary embolism (ICD-10, I26 and filtering in inferior vena cava or vein thrombectomy), and fracture (ICD-10, S72, M96.6, T84.1 and requiring open reduction and internal fixation or closed reduction) diagnosed after an operation and analyzed as independent variables. To adjust for the impact of patient comorbidity status on in-hospital complications, the Charlson comorbidity index [20] was determined based on the Quan version [21]. 2.2. Hospital primary THA volume group Hospital surgical volume was calculated as the total number of patients who underwent THA. Hospital surgical volume per 6-month period was classified into four groups so that each group contains at least 20 hospitals for statistical stability as follows: 0–16 procedures (Group A, lowest), 17–33 procedures (Group B, lower), 34–54 procedures (Group C, higher), and over 55 procedures (Group D, highest). Hospitals with the lowest volume (0–16 procedures per 6 months) were defined as the reference group. 2.3. Statistical analysis Patient complication rates were compared across hospital surgical volume groups using the chisquare test for trends in categorical variables and the Wilcoxon/Kruskal–Wallis equality of populations rank test for continuous variables. An unadjusted logistic regression model was used to assess change in the risk of in-hospital complications compared with the reference group. Multivariate logistic regression models adjusting for patient sex, age, and Charlson index were then created to assess the association between in-hospital complications and hospital surgical volume. Risk-adjusted complication rate of hospitals was not used because of expected statistical instability. All analyses were carried out using JMP 9.0 (SAS Institute Inc, Cary, NC, USA) statistical software. Two-sided p-values 0 was associated with significantly elevated risk of infection (OR 2.420, 95% CI 1.612–3.605). The multivariate adjustment model indicated that patients treated at hospitals in the lower-, higher- and highest-volume groups had a significantly decreased risk of peri-operative dislocation (OR
Patients treated at hospitals with the highest surgical volume were found to have a significantly lower risk of in-hospital dislocation and infection after THA surgeries compared with that in those treated at hospitals with the lowest surgical volume after adjusting for patient case mix. Our study have confirmed hospital-volume effects on the quality of THA surgeries from a large-scale nation-wide patient database in Japanese setting where socio-economic and ethnic factors are negligible due to universal social health insurance system. Biedermann et al. indicated that the higher rate of dislocation with inexperienced surgeons is due to malpositioning of the acetabular component [22]. Physician and hospital personnel at high-volume hospitals generally have more experience treating a wide variety of patients, which consequently increases their level of individual and organizational skill. Our results also indicated a higher rate of dislocation and fracture among patients 65 y of age and older. Older age may increase the likelihood of a fall. However, the reasons for this difference remain unclear. Further studies are needed to understand the causes of postoperative dislocation and fracture. Female sex was associated with a lower risk of infection. Pedersen et al. also reported that males who underwent primary THA procedure were at higher risk of infection [23]. PE is one of the most important complications of THA. The occurrence of PE was significantly decreased in the lower-volume hospitals, but not in the highest-volume hospitals, compared to that in the lowestvolume hospitals.
Table 2 Rate of complications according to hospital surgical volume of primary total hip arthroplasty in DPC data from July to December 2008. Complications
Hospital number LOS (range) Age (years) Number of patients Dislocation Infection Fracture Pulmonary embolism Charlson index ≥ 1 In hospital death
Total
8321 117 (1.31%) 103 (1.24%) 39 (0.47%) 62 (0.75%) 10 (0.12%)
p-Valuea
Hospital surgical volume A (0–16)
B (17–33)
C (34–54)
D (55≤)
433 44.0 (9–215) 67.2 (22–91) 2219 55 (0.66%) 22 (0.26%) 16 (0.19%) 22 (0.26%) 483 (21.8%) 5 (0.06%)
101 35.5 (3–263) 65.6 (16–97) 2287 30 (0.36%) 23 (0.28%) 7 (0.08%) 11 (0.13%) 558 (24.4%) 4 (0.05%)
46 32.7 (4–210) 64.2 (18–91) 1921 19 (0.23%) 56 (0.67%) 9 (0.11%) 9 (0.11%) 417 (21.7%) 0 (0%)
21 28.1 (5–151) 63.3 (24–91) 1894 13 (0.16%) 2 (0.02%) 7 (0.08%) 20 (0.24%) 263 (13.9%) 1 (0.01%)
ARTICLE IN PRESS
HEAP-3205; No. of Pages 6
Health Policy xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Health Policy journal homepage: www.elsevier.com/locate/healthpol
Relationship between peri-operative outcomes and hospital surgical volume of total hip arthroplasty in Japan Takeshi Kaneko a,1 , Kazuo Hirakawa b,2 , Kiyohide Fushimi a,∗ a Department of Health Policy and Informatics, Tokyo Medical Dental University, Graduate School, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8519, Japan b Shonan Kamakura Joint Reconstruction Center, 5-4-17 Dai Kamakura, Kanagawa 247-0061, Japan
a r t i c l e
i n f o
Article history: Received 26 June 2013 Received in revised form 10 February 2014 Accepted 24 March 2014
Keywords: Total hip arthroplasty Hospital surgical volume Peri-operative outcome Complication Quality of care
a b s t r a c t Background: This study aimed to evaluate the relationship between peri-operative complication of total hip arthroplasty (THA) and hospital surgical volume. Methods: We reviewed discharge administrative data from 8321 patients who underwent primary THA between July and December 2008. Relationships between complications and surgical volume were analyzed with multivariate logistic regression models including age, sex, and Charlson comorbidity index. Hospitals were categorized into four groups according to the 6-month volume of THA procedures. Results: The most frequent complication was dislocation (1.41%). Next was infection (1.24%). Fracture and pulmonary embolism occurred in less than 1% of procedures. Patients who underwent THA in hospitals with the highest surgical volume had lower risk of dislocation and infection than those treated in the hospitals with the lowest surgical volume (odds ratio [OR] 0.321, 95% confidence interval [CI] 0.167–0.572 and OR 0.123, 95% CI 0.020–0.421). Patients aged 65 years and older were associated with increased risk of dislocation (OR 2.342, CI 1.555–3.624) and fracture (OR 2.799, CI 1.372–6.301). Females demonstrated lower risk of dislocation (OR 0.558, CI 0.352–0.869) and infection (OR 0.560, CI 0.365–0.882). Conclusion: These results indicated that the increase in the risk of peri-operative dislocation of primary THA may be associated with low hospital surgical volume as well as age and male sex. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Osteoarthritis (OA) is the most common joint disorder affecting the elderly population worldwide and a leading cause of pain and disability. OA is estimated to affect 40% of people over 70 years of age [1], and 25.3% of peo-
∗ Corresponding author. Tel.: +81 358034025; fax: +81 358030357. E-mail addresses: [email protected] (T. Kaneko), [email protected] (K. Hirakawa), [email protected] (K. Fushimi). 1 Tel.: +81 358034025; fax: +81 358030357. 2 Tel.: +81 467472377; fax: +81 467472370.
ple develop symptomatic hip OA by age 85 years [2]. Pain caused by OA is a major factor affecting quality of life. Total hip arthroplasty (THA) is a highly effective treatment to improve mobility and eliminate pain [3,4]. As people live longer, the demand of joint replacement will be necessary in more cases. However, complications of THA can lead to poor functional outcome for patients; thus, reducing their occurrence is important. The relationship between the rate of postoperative complications and hospital surgical volume or socioeconomic factors has been unclear. Several studies have indicated that hospitals with a high surgical volume have a lower rate of overall mortality [5–8]. Among 76,627 Medicare patients
http://dx.doi.org/10.1016/j.healthpol.2014.03.013 0168-8510/© 2014 Elsevier Ireland Ltd. All rights reserved.
Please cite this article in press as: Kaneko T, et al. Relationship between peri-operative outcomes and hospital surgical volume of total hip arthroplasty in Japan. Health Policy (2014), http://dx.doi.org/10.1016/j.healthpol.2014.03.013
G Model HEAP-3205; No. of Pages 6
ARTICLE IN PRESS T. Kaneko et al. / Health Policy xxx (2014) xxx–xxx
2
who underwent primary or revision THA, a significant association was observed between higher hospital volume and lower 90-day mortality and dislocation rates [9]. Additionally, postoperative complications and hospital surgical volume were associated with total length of hospital stay in 1561 total joint arthroplasty patients [10]. By contrast, several studies have indicated no association between THA complication rate and surgical volume. Using data from 2965 patients, Yasunaga et al. reported that surgeon volume of over 500 procedures was inversely associated with postoperative THA complications, while hospital surgical volume was not significantly associated with postoperative THA complications [11]. Katz et al. also reported that race and income were associated with complication risk in the Medicare population [9]. The purpose of this study was to assess the relationship between peri-operative complications after THA and hospital surgical volume with large-scale nation-wide patient database and with minimal effects of socio-economic status and patient ethnicities. We have employed the DPC (Diagnosis Procedure Combination) database, which is a large-scale nationally representative administrative database of acute care inpatients in Japan and has successfully been used for many clinical epidemiological studies [12–14]. In Japanese setting, the influence of race and socioeconomic factors is negligible because of the universal medical insurance system, which provides almost unrestricted access to health care [15], and ethnic homogeneity [16]. 2. Patients and methods 2.1. Data source and study population The DPC code-based prospective payment scheme consists of 18 major diagnostic categories, 507 diagnostic groups, and 2658 case-mix groups in acute-care hospitals. This diagnosis system was instituted in 2003 by the Ministry of Health, Labour and Welfare of Japan. In this study, we used the DPC database consisting of patient level administrative data including age, sex, inhospital mortality, length of stay, and hospital reference cost calculated from fee-for-service tariff schedule. Preoperative comorbidities and postoperative complications that occurred during hospitalization were coded with the International Classification of Diseases, Tenth Revision (ICD-10) codes [17]. Additionally, surgical procedures were coded according to a Japanese fee-for-service tariff schedule. We reviewed administrative data from 8607 patients who underwent primary THA between July and December 2008 in the Japanese DPC database. The data were voluntarily offered to us by 855 hospitals that agreed to their use for research, which represent approximately 45% of inpatient admissions to acute-care hospitals in Japan. The data were anonymous and could not be linked to any other information to identify patients upon their collection by the research group. Prior approval by the ethics committee of Tokyo Medical and Dental University was granted to the study. We excluded 161 patients who underwent THA due to fracture (ICD-10, S72), infection (ICD-10, M00-03, M46, M86, M90.1, M90.2), and pathologic fracture (ICD-10
C40, C41, C76.3, C76.5, C79.5, D48.0, M48.5, M49.5, M80, M84.3, M84.4, M90.7) to select elective surgery cases to minimize case-mix heterogeneity. A total of 125 patients were excluded from one hospital where some inpatients can selectively be treated with non-prospective payment schema. Our final analysis included 8321 patients from 601 hospitals. A previously published coding algorithm was modified [9,18,19], and we defined outcome as the peri-operative complication rate. Complications were defined as dislocation (ICD-10, M24.3, M24.4, S73.0 and requiring repositioning of joint dislocation), infection (ICD-10, M00, M01, M86, T81.4, T84.5-9, T85.7 and surgical debridement for pyogenic arthritis, removal of the prosthesis), pulmonary embolism (ICD-10, I26 and filtering in inferior vena cava or vein thrombectomy), and fracture (ICD-10, S72, M96.6, T84.1 and requiring open reduction and internal fixation or closed reduction) diagnosed after an operation and analyzed as independent variables. To adjust for the impact of patient comorbidity status on in-hospital complications, the Charlson comorbidity index [20] was determined based on the Quan version [21]. 2.2. Hospital primary THA volume group Hospital surgical volume was calculated as the total number of patients who underwent THA. Hospital surgical volume per 6-month period was classified into four groups so that each group contains at least 20 hospitals for statistical stability as follows: 0–16 procedures (Group A, lowest), 17–33 procedures (Group B, lower), 34–54 procedures (Group C, higher), and over 55 procedures (Group D, highest). Hospitals with the lowest volume (0–16 procedures per 6 months) were defined as the reference group. 2.3. Statistical analysis Patient complication rates were compared across hospital surgical volume groups using the chisquare test for trends in categorical variables and the Wilcoxon/Kruskal–Wallis equality of populations rank test for continuous variables. An unadjusted logistic regression model was used to assess change in the risk of in-hospital complications compared with the reference group. Multivariate logistic regression models adjusting for patient sex, age, and Charlson index were then created to assess the association between in-hospital complications and hospital surgical volume. Risk-adjusted complication rate of hospitals was not used because of expected statistical instability. All analyses were carried out using JMP 9.0 (SAS Institute Inc, Cary, NC, USA) statistical software. Two-sided p-values 0 was associated with significantly elevated risk of infection (OR 2.420, 95% CI 1.612–3.605). The multivariate adjustment model indicated that patients treated at hospitals in the lower-, higher- and highest-volume groups had a significantly decreased risk of peri-operative dislocation (OR
Patients treated at hospitals with the highest surgical volume were found to have a significantly lower risk of in-hospital dislocation and infection after THA surgeries compared with that in those treated at hospitals with the lowest surgical volume after adjusting for patient case mix. Our study have confirmed hospital-volume effects on the quality of THA surgeries from a large-scale nation-wide patient database in Japanese setting where socio-economic and ethnic factors are negligible due to universal social health insurance system. Biedermann et al. indicated that the higher rate of dislocation with inexperienced surgeons is due to malpositioning of the acetabular component [22]. Physician and hospital personnel at high-volume hospitals generally have more experience treating a wide variety of patients, which consequently increases their level of individual and organizational skill. Our results also indicated a higher rate of dislocation and fracture among patients 65 y of age and older. Older age may increase the likelihood of a fall. However, the reasons for this difference remain unclear. Further studies are needed to understand the causes of postoperative dislocation and fracture. Female sex was associated with a lower risk of infection. Pedersen et al. also reported that males who underwent primary THA procedure were at higher risk of infection [23]. PE is one of the most important complications of THA. The occurrence of PE was significantly decreased in the lower-volume hospitals, but not in the highest-volume hospitals, compared to that in the lowestvolume hospitals.
Table 2 Rate of complications according to hospital surgical volume of primary total hip arthroplasty in DPC data from July to December 2008. Complications
Hospital number LOS (range) Age (years) Number of patients Dislocation Infection Fracture Pulmonary embolism Charlson index ≥ 1 In hospital death
Total
8321 117 (1.31%) 103 (1.24%) 39 (0.47%) 62 (0.75%) 10 (0.12%)
p-Valuea
Hospital surgical volume A (0–16)
B (17–33)
C (34–54)
D (55≤)
433 44.0 (9–215) 67.2 (22–91) 2219 55 (0.66%) 22 (0.26%) 16 (0.19%) 22 (0.26%) 483 (21.8%) 5 (0.06%)
101 35.5 (3–263) 65.6 (16–97) 2287 30 (0.36%) 23 (0.28%) 7 (0.08%) 11 (0.13%) 558 (24.4%) 4 (0.05%)
46 32.7 (4–210) 64.2 (18–91) 1921 19 (0.23%) 56 (0.67%) 9 (0.11%) 9 (0.11%) 417 (21.7%) 0 (0%)
21 28.1 (5–151) 63.3 (24–91) 1894 13 (0.16%) 2 (0.02%) 7 (0.08%) 20 (0.24%) 263 (13.9%) 1 (0.01%)
