SEARCH

SEARCH BY CITATION

Keywords:

  • Total hip replacement;
  • Rural health care;
  • Hospital choice;
  • Medicare

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES

Objective

To identify factors associated with utilization of low-volume hospitals for total hip replacement (THR) and to estimate differences in the distances that Medicare beneficiaries had to travel to reach low- or high-volume hospitals.

Methods

We studied a population-based sample of 1,146 Medicare beneficiaries who underwent elective THR in 1995. Using multiple data sources including medical record review, Medicare claims data, 1990 Census data, and a patient survey, we examined factors independently associated with utilization of low-volume hospitals for elective THR. We estimated the magnitude of difference in distances for patients undergoing THR in low- and high-volume hospitals. We determined the distance between each patient's residence and the treating hospital using MapQuest.

Results

Rural residency, low income, and low educational attainment, as well as belief in the importance of convenient location in the choice of hospital, were associated with higher utilization of low-volume hospitals. Rural and suburban patients who went to low-volume hospitals traveled much less than patients operated upon in high-volume centers.

Conclusion

Policies aimed at restricting THR to high-volume centers would differentially affect poor, less educated, and rural patients. Voluntary efforts to shift THR to high-volume centers should involve educating these patients and their referring physicians about differences in both short-term and longer-term outcomes between high- and low-volume centers.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES

A large body of evidence indicates that for certain surgical procedures and types of medical admissions, hospitals with a lower annual caseload have higher complication and mortality rates. These associations have been documented in cardiac revascularization procedures, cataract, peripheral and cancer surgery, human immunodeficiency virsu care, and acute myocardial infarction, among others (1–6). This body of evidence has led the Institute of Medicine and the Leapfrog Group of major US employers to endorse referral to high-volume centers as a key strategy for improving the quality and outcomes of care, particularly for high-risk procedures (7). The Centers for Medicare and Medicaid Services (formerly the Health Care Financing Administration) has launched centers of excellence pilot programs for cardiac surgery and, most recently, for total hip and total knee replacement (8). Hospital volume is a key criterion for selection into these programs.

However, there is also emerging evidence that some patients have strong preferences for receiving care locally rather than traveling to a referral center. In one study that involved responses to a hypothetical scenario involving the Whipple procedure for pancreatic cancer, some patients stated they would prefer local care, even if mortality in local centers were as much as 6 times the mortality rates in referral hospitals (9). Previous research showed that hospital factors, such as teaching status, hospital quality, overall bed size, and convenient location influence patient choice of hospitals (10). Patient age, rural residency, and complexity or severity of illness also play a role (11, 12), as do distance or travel time to the point of service and waiting time (13, 14). However, most prior research on factors influencing hospital choice has focused exclusively on rural patients or high-risk interventions for life-threatening morbid conditions (9, 15).

Total hip replacement (THR) is an elective procedure performed frequently to relieve pain and improve function in patients with advanced hip joint destruction. Hospital volume of THR is inversely associated with postoperative mortality and complications (16–18). The issues of hospital choice are particularly relevant to THR because ∼25% of elective THRs are performed in low-volume centers (18). Thus, regionalization of this procedure may potentially result in underutilization of THR for some patients. However, there has been little research on why some patients choose low-volume hospitals for elective THR nor has there been investigation of who would be most affected by policies that shift care from low- to high-volume centers.

To clarify these issues, we performed a secondary analysis of data from a population-based cohort of THR recipients to identify demographic, geographic, and clinical factors associated with patient use of low-volume hospitals for elective THR. We also estimated differences in distance from the patient's residence to the hospital for patients undergoing THR in high- and low-volume hospitals. We hypothesized that rural residence and lower socioeconomic status would predict utilization of low-volume hospitals and that patients who stated that convenience influenced their hospital choice would more likely have THR in a low-volume center. We further hypothesized that the association between low socioeconomic status and use of low-volume hospitals would be especially striking in urban patients.

SUBJECTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES

Study sample.

Our analysis was based on a sample of subjects from a study of the relationship between outcomes and hospital volume of THR. In the parent study, we used Medicare claims data to identify beneficiaries who underwent elective primary or revision total hip replacement in 1995. Our algorithms for selection of THR cases have been published elsewhere (18). We selected a random sample of THR recipients from Ohio, Pennsylvania, and Colorado for detailed investigation, using a stratified sampling technique (with strata reflecting hospital volume) to assure adequate representation of patients from both low- and high-volume hospitals. We invited these patients to participate in a written or phone survey. The response rate was ∼50%, with no difference between respondents and nonrespondents in utilization of low-volume hospitals. We reviewed the medical records of patients agreeing to participate in the study and mailed them a survey to solicit additional demographic characteristics, reasons for hospital choice, as well as the patient's reported functional outcomes 3 years following surgery.

The current analyses were restricted to the subsample of patients who completed the survey and did not change residence between 1995 and 1998. This additional exclusion criterion permitted us to accurately assess the distance between the patient's residence and the hospital where the elective THR was performed. This restriction was necessary because Health Care Financing Administration provided us with 1998 addresses, but our hospital zip codes were from 1995. Primary THR recipients who changed residence had rates of low-volume hospital utilization that were similar to those for THR patients who did not move beyond their zip code. However, revision THR recipients who changed residency used low-volume hospitals less often than those who did not move beyond their zip code. Neither for primary nor for revision cohorts did such restrictions introduce selection bias because those who moved did not differ from patients used in the analyses with respect to all factors under consideration beyond hospital volume.

Data elements.

We divided data elements into 2 domains: hospital and patient-based characteristics. Patient factors were further grouped into the following: demographic characteristics, geographic factors, clinical factors, and patients' reported reasons for choice of hospital. We used multiple sources to elicit these data elements, including Medicare claims, the 1995 American Hospital Association survey, 1990 US Census data, the patient survey, and medical records.

Hospital characteristics.

We obtained hospital volume data from Medicare claims. Hospital volume was defined as a total number of primary plus revision THRs performed in the hospital during 1995 in the Medicare population. We defined low-volume hospitals (LVHs) for primary THR to be those hospitals in the lowest quartile of the distribution of hospital volume among patients having primary THR. For primary THR, the cutoff was 12 THRs per year. Low-volume hospitals for revision THR were defined analogously (those hospitals in the lowest quartile of hospital volume for patients having revision surgery) and resulted in a cutoff of 24 THR per year. We obtained hospital addresses from the 1995 American Hospital Association survey data (19).

Patient characteristics.

Demographic factors.

We used Medicare denominator files to obtain data on patient age and sex. We obtained data on education and income from the patient survey. We considered people to have low socioeconomic status (SES) if they reported annual income of less than $20,000, they did not graduate from high school, or both.

Geographic factors.

Patient 1995 zip codes came from the Medicare denominator file. Centers for Medicare and Medicaid Services provided us with the complete patient addresses in 1998 from their enrollment file. We used 1990 Census data to obtain information on the population size of each zip code, and the proportion of each zip code considered rural (20). We defined residence as rural if the zip code population size was <10,000 and if >20% of the zip code area was considered rural by census data. Urban residences were defined as those with zip code population size >30,000 and 0% of the area was rural. Zip codes that did not meet definition or either urban or rural were defined as suburban.

Clinical factors.

We obtained data on underlying diagnosis, comorbidities, weight, and height from medical records. Recalled preoperative functional status, history of prior orthopedic surgery, weight, and height were obtained from the patient survey. Preoperative functional status was assessed in the patient survey by recall with several items from the self-administered Harris Hip Score (21, 22). We considered patients to have poor preoperative status if they had preoperative functional status scores in the lowest quartile of patients undergoing primary or revision THR. We calculated body mass index (BMI) based on data from both medical records and the patient survey, and found the correlation between those 2 measures to be 0.97. Because medical records were more likely to omit data on weight and height, we used BMI based on patient self report in all our analyses. We considered patients to be overweight if their BMI exceeded 30 kg/m2. We summarized comorbidities using the medical-record based Charlson comorbidity index (23). Based on medical record review, we also obtained information on the number of prior orthopedic surgeries and the complexity of the revision procedure. Revision was classified as complex if a patient either had a prior revision on that hip or needed bone or structural graft.

Hospital choice preference factors.

The survey asked patients to indicate the factor or factors that influenced their choice of hospital for THR. The possible factors included insurance coverage, convenient location, hospital reputation, and recommendation of their personal care physician or/and family and friends. These factors were suggested by published literature or expert opinion (10).

Outcomes.

We focused our analysis on 2 outcomes: utilization of a low-volume hospital and average distance between the patient's residence and the treating hospital. Using patient and hospital addresses we employed MapQuest—an Internet-based mapping tool (URL: www.mapquest.com)—to estimate distances between the patients' residences and the treating hospitals (24). We restricted our analysis to patients who traveled <200 miles to their hospitals to eliminate possible data errors and seasonal moves.

Statistical analysis.

We included in the multivariate analyses those factors that we suspected strongly a priori would be associated with hospital choice and those whose bivariate association with utilization of LVHs had a P value < 0.1 or a risk ratio (RR) > 1.5 or < 0.67. We used multivariate generalized linear models to estimate adjusted risk ratios. We also constructed models to evaluate interactions between residency type and convenience, as well as between SES, residency type, and utilization of LVHs.

Average distances (with 95% confidence intervals) between patient's residences and hospitals stratified by residency type and hospital volume were estimated from general linear models (25). All analyses were performed using SAS 8.2 software (SAS Institute, Cary, NC). Statistical significance was based on P values < 0.05.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES

Description of the cohorts.

About one-third of patients in our primary THR cohort and about half of the revision THR cohort were >75 years old; about two-thirds of both cohorts were women. About half of patients reported having income less than $20,000/per year and ∼20% did not graduate from high school (Table 1). The majority of patients lived in suburban areas, whereas 24% of primary THR recipients and 15% of the revision cohort resided in rural areas. About half of each primary and revision patients identified physician recommendation as a factor influencing hospital choice (Table 2); about one-third mentioned hospital reputation and convenient location.

Table 1. Select demographic, geographic, and clinical characteristics of primary and revision THR cohorts: utilization of low-volume hospitals*
FeaturePrimary THR n = 714Revision THR n = 435
No. (%)LVH utilization m (%)No. (%)LVH utilization m (%)
  • *

    THR = total hip replacement; LVH = low-volume hospital; OA = osteoarthritis; RA = rheumatoid arthritis; AVN = avascular necrosis; N/A = not applicable.

  • Denotes number and percentage of patients who used low volume hospital for THR within each patient subgroup described in the left column.

  • P < 0.05.

  • §

    Diagnosis not reliably reported (or relevant) for revision THR.

  • Preoperative function (recalled) assessed by self-administered Harris Hip Score.

  • #

    Revision surgery was considered to be complex if a patient already had at least 1 revision and/or bone or structural graft was used.

Demographic characteristics    
 Age, years    
  ≥75270 (38)77 (29)206 (47)49 (24)
  <75444 (62)104 (23)229 (53)56 (24)
 Sex    
  Female447 (63)115 (26)267 (61)59 (22)
  Male267 (37)66 (25)168 (39)46 (27)
 Income, $    
  <20,000335 (48)113 (34)214 (50)62 (29)
  ≥20,000368 (52)62 (17)212 (50)41 (19)
 Education    
  <high school147 (21)55 (37)71 (17)26 (37)
  ≥high school548 (79)119 (22)352 (83)76 (22)
Geographic characteristics    
 Type of residency    
  Rural173 (24)74 (43)67 (15)24 (36)
  Suburban451 (63)96 (21)316 (73)69 (22)
  Urban90 (13)11 (12)52 (12)12 (23)
Clinical characteristics    
 Diagnosis§    
  OA566 (88)147 (26)  
  RA27 (4)6 (22)  
  AVN53 (8)12 (23)  
 Preoperative function, lowest quartile    
  Yes189 (27)57 (30)93 (22)30 (32)
  No498 (73)117 (23)312 (77)67 (21)
 Comorbidity    
  >2138 (19)37 (27)89 (21)18 (20)
  ≤2576 (81)144 (25)343 (79)86 (25)
 Number of prior orthopedic surgeries    
  0455 (64)114 (25)00
  1215 (30)55 (26)192 (44)52 (27)
  >144 (6)12 (27)243 (56)53 (22)
 Complex#    
  YesN/AN/A224 (51)46 (21)
  No  211 (49)59 (28)
 Body mass index, kg/m2    
  >30164 (24)43 (26)85 (20)26 (31)
  ≤30527 (76)130 (25)337 (80)75 (22)
Table 2. Factors reported by patients as influencing utilization of low-volume hospital for primary and revision THR cohorts*
FeaturePrimary THR n = 714Revision THR n = 435
No. (%)LVH utilization m (%)No. (%)LVH utilization m (%)
  • *

    THR = total hip replacement; LVH = low-volume hospital.

  • Denotes number and percentage of patients who used low volume hospital for THR within each patient subgroup described in the left column.

Convenient location283 (40)111 (39)128 (29)51 (40)
Insurance purposes76 (11)22 (29)32 (7)7 (22)
Physician's recommendation366 (51)83 (23)242 (56)53 (22)
Family/friends recommendation141 (20)29 (21)58 (13)10 (17)
Hospital reputation262 (37)52 (20)144 (33)32 (22)

Utilization of low-volume hospitals.

Primary THR.

Preoperative functional status, SES, residency type, convenient location, and hospital reputation exhibited crude associations with utilization of an LVH and therefore advanced into the multivariate model (Table 3). In multivariate analyses, patients who resided in a rural area were 3.47 (95% confidence interval [95% CI] 1.63–7.39) times more likely to utilize an LVH than patients who lived in urban areas and 2.10 (95% CI 1.55–2.85) times more likely than those who lived in suburban areas.

Table 3. Crude and adjusted risk ratios of association between low-volume hospital utilization and select demographic characteristics, geographic characteristics, and factors influencing hospital choice for primary THR*
 Crude RR (95% CI)Adjusted RR (95% CI)P
  • *

    Factors exhibited association with LVH utilization in bivariate analyses by RR > 1.5 or RR < 0.67 or P < 0.1: convenient location, hospital reputation, type of residency, and preoperative functional status. THR = total hip replacement; RR = relative risk; 95% CI = 95% confidence interval; SES = socioeconomic status; LVH = low-volume hospital.

  • Preoperative function (recalled) assessed by self-administered Harris Hip Score.

  • Interaction terms in the model (values correspond to RR of relationship between the specific factor and the choice of low-volume hospital within each residency strata).

Preoperative function, lowest quartile   
 Yes1.28 (0.98–1.68)1.13 (0.87–1.46) 
 No1.00  
Type of residency  < 0.0001
 Rural versus urban3.50 (1.96–6.25)3.47 (1.63–7.39) 
 Rural versus suburban2.00 (1.57–2.57)2.10 (1.55–2.85) 
 Suburban versus urban1.74 (0.97–3.11)1.65 (0.77–3.52) 
Low SES (yes versus no)1.92 (1.45–2.54) 0.0249
 Rural 1.13 (0.78–1.65) 
 Suburban 2.08 (1.38–3.12) 
 Urban 2.68 (0.77–9.28) 
Hospital reputation influenced choice   
 Yes0.70 (0.52–0.92)0.73 (0.55–0.97) 
 No1.001.00 
Convenient location influenced hospital choice (yes versus no)2.42 (1.86–3.13) 0.0851
 Rural 1.69 (1.16–2.45) 
 Suburban 2.27 (1.57–3.30) 
 Urban 5.82 (1.63–20.82) 

The association between low SES and LVH utilization was modified by area of residence. For rural patients, there was no association between SES and use of low-volume hospitals (RR 1.13, 95% CI 0.78–1.65). In contrast, urban patients with low SES were more than twice as likely to use LVHs than their peers with higher SES (RR 2.68, 95% CI 0.78–9.28). Patients stating that convenient location was a factor influencing hospital choice were also more likely to use LVHs. This association was modified by residency type, with convenience particularly strongly associated with LVH use in urban patients (RR 5.82, 95% CI 1.63–20.82; Table 3). On the other hand, patients who cited hospital reputation among factors relevant to hospital choice had about a 25% lower chance of going to an LVH than those who did not cite hospital reputation (adjusted RR 0.73, 95% CI 0.55–0.97).

Revision THR.

For the revision THR cohort, we included in the multivariate model SES, residency type, preoperative functional status, procedure complexity, being overweight, and reporting that convenient location of the hospital influenced hospital choice (Table 4). Because urban and suburban residents undergoing revision THR had a similar likelihood of going to LVHs, we combined these 2 groups and contrasted them with rural patients. In the multivariate analysis, none of the clinical characteristics included retained statistical significance or effects of any magnitude. However, patients with low SES were ∼50% more likely to use LVHs for their revision THR than patients with higher SES. The patient's report that convenient location influenced hospital choice was strongly associated with LVH utilization after adjustment for residence, SES, and clinical factors (RR 2.20, 95% CI 1.64–3.13).

Table 4. Crude and adjusted risk ratios of association between LVH utilization and select demographic, geographic, and clinical characteristics and factors influencing hospital choice for revision THR*
 Crude RR (95% CI)Adjusted RR (95% CI)
  • *

    Factors exhibited association with low-volume hospital (LVH) utilization in bivariate analyses by relative risk (RR) > 1.5 or RR < 0.67 or P < 0.1: convenient location, body mass index (BMI), complexity, preoperative functional status, residency, socioeconomic status (SES). THR = total hip replacement; 95% CI = 95% confidence interval.

  • Preoperative function (recalled) assessed by self-administered Harris Hip Score.

  • Revision surgery was considered to be complex if a patient already had at least 1 revision and/or bone or structural graft was used.

Low SES  
 Yes1.65 (1.14–2.37)1.57 (1.07–2.31)
 No1.01.0
Type of residency  
 Rural1.63 (1.12–2.37)1.63 (1.10–2.40)
 Suburban or urban1.01.00
Preoperative function, lowest quartile  
 Yes1.50 (1.04–2.16)1.21 (0.82–1.77)
 No1.01.00
Complex  
 Yes0.73 (0.52–1.03)0.88 (0.63–1.22)
 No1.01.00
BMI  
 >301.37 (0.94–2.00)0.98 (0.64–1.48)
 ≤301.01.00
Convenient location  
 Yes2.27 (1.64–3.13)2.20 (1.55–3.13)
 No1.01.0

Distance

Patients from rural areas who went to LVHs for primary THR traveled an average of 19 miles (95% CI 13–24) compared with 35 miles (95% CI 30–40) for rural patients who went to high-volume hospitals (Figure 1). On average, suburban residents traveled 10 miles to an LVH (95% CI 5–15) and 18 miles to a high-volume hospital (95% CI 16–21). Urban residents traveled on average 6 miles to either low- or high-volume hospitals. A similar pattern held for revision THR (Figure 1).

thumbnail image

Figure 1. Average distance between patient residence and hospital for elective total hip replacement (THR) stratified by residence type and hospital volume. Top: Illustrates distances between patient residence and hospital for primary THR. Bottom: Illustrates distances between patient residence and hospital for revision THR. Squares correspond to mean distances and error bars correspond to 95% confidence intervals. Low-volume categories: for primaries, low = 1–12 (combined primary and revision THR in Medicare population); for revisions, low = 1–24.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES

We studied a population-based sample of Medicare beneficiaries who underwent elective primary or revision THR in 1995 to better understand the reasons patients choose low-volume hospitals for THR. We found that rural residency, low SES, and convenient hospital location independently predicted use of LVHs. Rural and suburban patients who went to low-volume hospitals traveled much less than patients who used high-volume centers, whereas there was no difference in travel distance to low- versus high-volume centers for urban residents.

Our results are consistent with research in pancreatic cancer (9) and breast cancer treatment (15), in which some patients stated that they would prefer to receive care in a local, small-volume hospital. The major distinction from these studies is that we studied a low-risk procedure for a nonfatal disease (hip arthritis), which may make preferences for local care even stronger.

The association of LVH use and importance of convenient location was most prominent among urban patients despite the fact that these patients had to travel very little to reach either low- or high-volume centers. This observation indicates that the concept of convenience is much broader than simply geographic proximity. Previous research has shown that convenience, defined as a multidimensional concept including convenient location and general familiarity with the structure and personnel of the referring facility, is one of the primary factors affecting the general practitioner's choice of hospital when referring patients for elective surgery (26). Because many patients indicated that their primary care physicians influenced their choice of hospital, voluntary efforts to shift THR to high-volume centers (8) should include discussion among patients and their referring physicians about hospital choice.

The differential effect of low SES on utilization of LVHs for rural and urban patients may have several explanations. For rural patients, the absence of an association between SES and utilization of low-volume hospitals may simply reflect the limited available choices of hospitals regardless of socioeconomic situation. The strong positive association between low SES and the likelihood of having THR in an LVH may reflect disparities in referral patterns for economically disadvantaged groups.

Centers with 1–10 Medicare cases per year have 2-fold greater relative risk of perioperative mortality after THR than centers performing >100 procedures annually. However, the corresponding absolute risks of mortality are minimal and do not exceed 1.3% for primary THR, even in the lowest-volume hospitals (16–18). These differences suggest that regionalization may help to avert a few perioperative deaths—but may also potentially cause more patients to decide not to undergo this elective procedure. Furthermore our data suggest that the volume–outcomes relationship may not persist beyond the postoperative period. Specifically, patients undergoing THR in high- and low-volume hospitals have similar functional status and pain relief 3 years postoperatively (27).

The strengths of our study include a population-based sample and access to a wide range of data sources. However, our study had several limitations. The survey was done 3 years after THR, and so required postoperative recall of some preoperative factors, such as functional status. Also, factors potentially influencing hospital choice, such as marital status, were documented in 1998 and may have differed from patient status at the time of the surgery. It is unclear if data from 1 year in 3 states can be generalized to the entire United States. Another limitation of our study is the lack of information about environmental factors, such as whether patients had any other hospitals closer to their residence than the one that they used for THR. Results of our analyses showing that rural patients had to travel an average of 19 miles to reach even a local (low volume) center indicate that there are no low- or high-volume hospitals in close proximity to residence for rural patients.

Results of our study indicate that policies aimed at restricting THR to high-volume centers would differentially affect older, poor, less-educated, rural patients. These restrictions may widen existing economic disparities in utilization of total hip replacement. More studies need to be undertaken to examine whether rural poor patients with advanced hip arthritis will be willing to travel to regional centers of excellence to have total hip replacement.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES
  • 1
    Bach PB, Cramer LD, Schrag D, Downey RJ, Gelfand SE, Begg CB. The influence of hospital volume on survival after resection for lung cancer. N Engl J Med 2001; 345: 1818.
  • 2
    Comarow A. Higher volume, fewer deaths: for a risky operation, head for a hospital that does it regularly. US News World Rep 2000; 129: 6870.
  • 3
    Birkmeyer JD, Finlayson SR, Tosteson AN, Sharp SM, Warshaw AL, Fisher ES. Effect of hospital volume on in-hospital mortality with pancreaticoduodenectomy. Surgery 1999; 125: 2506.
  • 4
    Glasgow RE, Showstack JA, Katz PP, Corvera CU, Warren RS, Mulvihill SJ. The relationship between hospital volume and outcomes of hepatic resection for hepatocellular carcinoma. Arch Surg 1999; 134: 305.
  • 5
    Thiemann DR, Coresh J, Oetgen WJ, Powe NR. The association between hospital volume and survival after acute myocardial infarction in elderly patients. N Engl J Med 1999; 340: 16408.
  • 6
    Dudley RA, Johansen KL, Brand R, Rennie DJ, Milstein A. Selective referral to high volume hospitals: estimating potentially avoidable deaths. JAMA 2000; 283: 115966.
  • 7
    The Leapfrog Group Web site. Accessed August 2, 2004. URL: www.leapfroggroup.org.
  • 8
    HCFA invites 84 hospitals to apply for joint project. Orthopedics Today 1997; 17: 1, 9.
  • 9
    Finlayson SR, Birkmeyer JD, Tosteson AN, Nease RF Jr. Patient preferences for location of care: implications for regionalization. Med Care 1999; 37: 2049.
  • 10
    Taylor SL, Capella LM. Hospital outshopping: determinant attributes and hospital choice. Health Care Manage Rev 1996; 21: 3344.
  • 11
    Skinner TJ, Price BS, Scott DW, Gorry GA. Factors affecting the choice of hospital-based ambulatory care by the urban poor. Am J Public Health 1977; 67: 43945.
  • 12
    Adams EK, Houchens R, Wright GE, Robbins J. Predicting hospital choice for rural Medicare beneficiaries: the role of severity of illness. Health Serv Res 1991; 26: 583612.
  • 13
    Luft HS, Garnick DW, Mark DH, Peltzman DJ, Phibbs CS, Lichtenberg E, et al. Does quality influence choice of hospital? JAMA 1990; 263: 2899906.
  • 14
    Llewellyn-Thomas HA, Arshinoff R, Bell M, Williams JI, Naylor CD. Ontario Hip and Knee Replacement Project Team. In the queue for total joint replacement: patients' perspectives on waiting times. J Eval Clin Pract 1998; 4: 6374.
  • 15
    Palda VA, Llewellyn-Thomas HA, Mackenzie RG, Pritchard KI, Naylor CD. Breast cancer patients' attitudes about rationing postlumpectomy radiation therapy: applicability of trade-off methods to policy-making. J Clin Oncol 1997; 15: 3192200.
  • 16
    Kreder HJ, Losina E, Barrett J, Phillips CB, Mahomed NN, Lew RA, et al. Relationship between the volume of total hip replacements performed by providers and the rates of postoperative complications in the state of Washington. J Bone Joint Surg Am 1997; 79: 48594.
  • 17
    Taylor HD, Dennis DA, Crane HS. Relationship between mortality rates and hospital patient volume for Medicare patients undergoing major orthopaedic surgery of the hip, knee, spine, and femur. J Arthroplasty 1997; 12: 23542.
  • 18
    Katz JN, Losina E, Bartlett J, Phillips CB, Mahomed NN, Lew RA, et al. Association between hospital and surgeon procedure volume and outcomes of total hip replacement in the United States Medicare population. J Bone Joint Surg Am 2001; 83-A: 16229.
  • 19
    American Hospital Association. American Hospitals Association 1995 Survey of Hospitals; 1996.
  • 20
    1990 United States Census. Accessed November 1, 2003. Available at URL: http://www.census.gov/geo/www/gazetteer/places.html.
  • 21
    Mahomed NN, Arndt DC, McGrory BJ, Harris WH. The Harris Hip Score: comparison of patient self-report with surgeon assessment. J Arthroplasty 2001; 16: 57580.
  • 22
    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 Am 1969; 51: 73755.
  • 23
    Charlson M, Szatrowski TP, Peterson J, Gold J. Validation of a combined comorbidity index. J Clin Epidemiol 1994; 47: 124551.
  • 24
    Mapquest Web site. URL: www.mapquest.com.
  • 25
    Freud RJ, Little RC. SAS system for regression. 3rd ed. New York: John Wiley & Sons; 2000.
  • 26
    Mahon A, Whitehouse C, Wilkin D, Nocon A. Factors that influence general practitioners' choice of hospital when referring patients for elective surgery. Br J Gen Pract 1993; 43: 2726.
  • 27
    Katz JN, Phillips CB, Baron JA, Fossel AH, Mahomed NN, Barrett J, et al. Association of hospital and surgeon volume of total hip replacement with functional status and satisfaction three years following surgery. Arthritis Rheum 2003; 48: 5608.