SEARCH

SEARCH BY CITATION

Keywords:

  • urologic oncology;
  • disparities;
  • sex;
  • race;
  • insurance;
  • income

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

BACKGROUND:

Socioeconomic status represents an established barrier to health care access. Age, sex, and race may also play a role. The authors examined whether these affect the access to high-volume hospitals for uro-oncologic procedures in the United States.

METHODS:

Within the Nationwide Inpatient Sample (NIS), the authors focused on radical prostatectomy (RP), radical cystectomy, and nephrectomy (Nx) performed within the 5 most contemporary years (2003-2007). Logistic regression models were used to estimate the impact of the primary predictors on the likelihood of receiving care at a high-volume hospital.

RESULTS:

Between 2003 and 2007, 62,165 RP, 6557 radical cystectomy, and 28,062 Nx cases were recorded within the NIS. Patient age (P = .001), year of surgery (P = .001), Charlson Comorbidity Index (P ≤ .025), median Zip Code income (highest vs lowest quartile, P = .001), and insurance status (private vs Medicare, P = .008) were independent predictors of being treated at high-volume institutions. Moreover, black race was an independent predictor of decreased utilization of high-volume institutions for radical cystectomy (P = .012), and female sex was an independent predictor of decreased utilization of high-volume institutions for Nx (P = .016).

CONCLUSIONS:

On average, old, sick, poor, and Medicare patients were less likely to be treated at high-volume hospitals for uro-oncologic surgery. Similarly, black patients were less likely to have a radical cystectomy at a high-volume hospital, and female patients were less likely to have an Nx at a high-volume hospital. Selective referral of individuals who are less likely to receive care at such institutions may represent a health care priority intended to optimize outcomes across all population strata. Cancer 2012. © 2012 American Cancer Society.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

Numerous studies have shown that surgeon and hospital volume are inversely correlated with favorable outcomes after surgery.1 Specifically, investigators have reported lower in-hospital mortality and morbidity, shorter length of stay (LOS), and decreased hospital charges when patients are treated by high-volume providers.2-4 On the basis of these reports, patients are now seeking care at hospitals that meet certain volume thresholds, such as those suggested by the Leapfrog Group for Patient Safety.5 However, a recent study by Liu et al showed that there are substantial disparities in the characteristics of patients receiving care at high-volume hospitals for complex procedures, such as abdominal aortic aneurysm repair and coronary artery bypass grafting.6 These findings have not been validated in the context of uro-oncologic surgery.

Given the lack of available data, we sought to explore the effect of significant patient and socioeconomic determinants on access to high-volume hospitals for the uro-oncologic procedures radical prostatectomy (RP), radical cystectomy, and nephrectomy (Nx). Within a large contemporary (2003-2007) population-based cohort of individuals, we examine the influence of age, sex, year of surgery, comorbidity profile, insurance status, race, and income.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

Data Source

Data from the 5 most contemporary years (2003-2007) of the Nationwide Inpatient Sample (NIS) were abstracted. The NIS includes inpatient discharge data collected via federal-state partnerships, as part of the Agency for Healthcare Research and Quality's Healthcare Cost and Utilization Project. As of the year 2007, the NIS contained administrative data on 8,043,415 discharges from 1044 hospitals within 40 states, approximating 20% of community hospitals within the United States, including public hospitals and academic medical centers. The NIS is the sole hospital database in the United States with charge information on all patients regardless of payer, including persons covered by Medicare, Medicaid, and private insurance and the uninsured.

Sample Population and Surgical Procedures

Relying on discharge records, all patients with a primary diagnosis of prostate cancer (International Classification of Diseases, 9th Edition, Clinical Modification [ICD-9-CM] code 185), bladder cancer (188) or kidney cancer (189) were considered for the study. Secondary diagnostic codes (ICD-9-CM 197.0, 197.7, 198.x) were used to identify patients with metastases, who were excluded from any further analysis. The prostatectomy procedure code (60.5) resulted in the identification of 63,827 patients. The cystectomy procedure code (57.7) resulted in the identification of 6697 patients. The partial (55.4) and radical (55.5) Nx codes resulted in the identification of 28,728 patients.

Baseline Patient and Hospital Characteristics

For all patients, the following variables were available: age, sex, year of surgery, race (white vs black vs other vs unknown), Charlson Comorbidity Index (CCI), hospital volume, insurance status (private vs Medicaid vs Medicare vs other), and median Zip Code income. CCI was derived from ICD-9 codes according to previously established criteria7 and was stratified according to 4 levels, 0 to 3. The median household income for the patient's Zip Code was coded within the NIS and was stratified according to 5 levels: 1) $1 to $35,999; 2) $36,000 to $44,999; 3) $45,000 to $58,999; 4) $59,000 or more; and 5) unknown. Patients with unknown insurance status or median Zip Code income were not considered for analyses.

Hospital volume for a given surgical procedure was defined as the annual procedural volume for each year between January 1, 2003 and December 31, 2007. The dependent variable was created by stratifying hospital volume into patient quartile, with the highest 25% of patients by mean annual volume considered high volume. Specifically, the 75th percentile cutoffs for RP, radical cystectomy, and Nx were 148, 10, and 23 cases per year, respectively. On the basis of our methodology, a given hospital could move from 1 quartile to another from 1 year to the next.

Statistical Analysis

Descriptive statistics focused on frequencies and proportions for categorical variables. Means, medians, and ranges were reported for continuously coded variables. The chi-square and independent t test were used to compare the statistical significance of differences in proportions and means, respectively.

Subsequently, logistic regression models were used to estimate the impact of the primary predictors (age, sex, year of surgery, comorbidity profile, insurance, race, and median Zip Code income) on the likelihood of receiving care at a high-volume hospital. Models were fitted as a single cohort and separately for each procedure of interest. In addition, we relied on generalized estimating equations to adjust for clustering.8

All tests were 2-sided, with statistical significance set a P < .05. Analyses were conducted using the R statistical package (the R Foundation for Statistical Computing, version 2.13.1).

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

Between 2003 and 2007, 62,165 eligible RP, 6557 eligible radical cystectomy, and 28,062 eligible Nx cases were recorded within the NIS, for a total of 96,784 procedures. Of these, 15,298, 1623, and 6972 were treated at high-volume institutions, respectively. Baseline characteristics of patients undergoing uro-oncologic procedures in the NIS between 2003 and 2007 are listed in Table 1. Patients treated at high-volume centers were younger (median age, 61 vs 62 years), had fewer comorbidities (CCI of 0 in 77.2% vs 72.2%), and were more likely to be treated in the most recent years (36.8% vs 19.6% treated in 2007). Patients treated at high-volume institutions were also more likely to be white (61.9% vs 56.8%), more likely to be within the highest quartile of median Zip Code income (39.3% vs 27.6%), and more likely to hold private insurance (62.5% vs 55.5%; P < .001). Table 1 further describes patient characteristics for each of the 3 procedures.

Table 1. Demographic Characteristics of Patients Undergoing Radical Prostatectomy, Radical Cystectomy, and Nephrectomy, Stratified According to Institutional Volume, Nationwide Inpatient Sample, 2003-2007
 OverallRadical ProstatectomyRadical CystectomyNephrectomy
CharacteristicNonhigh VolumeHigh VolumePNonhigh VolumeHigh VolumePNonhigh VolumeHigh VolumePNonhigh VolumeHigh VolumeP
  • Abbreviation: CCI, Charlson Comorbidity Status.

  • a

    Includes Asian, Pacific Islander, Native American, and other unspecified.

  • b

    Based on comorbidity as developed by Charlson et al9 and adapted by Deyo et al.10

No. of patients72,89123,89346,86715,2984934162321,0906972
Mean age (median), range, y62 (62), 18-9761 (61), 18-104<.00161 (62), 32-8960 (60), 28-88<.00169 (70), 40-9568 (68), 40-90<.00162 (63), 18-9761 (62), 18-104<.001
Sex  .015    .284  .001
 Men87.387.9 100.0100.0 82.984.0 60.162.2 
 Women12.712.1 0.00.0 17.116.0 39.937.8 
Year of surgery  <.001  <.001  <.001  <.001
 200320.912.9 21.710.9 20.818.0 19.116.1 
 200419.813.7 19.911.5 20.016.7 19.317.9 
 200519.214.4 18.812.5 19.415.7 20.018.1 
 200620.522.3 20.422.5 20.417.5 20.822.9 
 200719.636.8 19.242.6 19.532.1 20.825.0 
Race  <.001  <.001  <.001  <.001
 White56.861.9 55.960.9 63.374.2 57.261.1 
 Black7.77.3 8.57.5 3.72.6 6.87.9 
 Othera7.67.8 7.17.1 4.97.7 9.39.3 
 Unknown27.923.1 28.524.5 28.115.5 26.721.7 
CCIb  <.001  <.001  <.001  <.001
 072.277.2 78.283.7 64.073.8 60.763.8 
 121.817.9 18.814.2 25.218.0 27.626.1 
 23.63.0 2.31.6 6.05.4 6.05.4 
 ≥32.41.9 0.70.5 4.82.8 5.74.7 
Median Zip Code income  <.001  <.001  .006  <.001
 1st quartile20.016.2 18.813.3 20.120.5 22.521.8 
 2nd quartile25.019.7 24.517.7 25.826.1 25.822.5 
 3rd quartile27.524.8 27.724.5 28.224.2 26.725.6 
 4th quartile27.639.3 28.944.5 25.929.2 25.030.1 
Insurance status  <.001  <.001  <.001  <.001
 Private55.562.5 62.971.3 28.934.2 45.349.6 
 Medicaid2.62.5 1.71.4 3.73.9 4.34.4 
 Medicare37.431.6 31.024.5 63.657.4 45.441.1 
 Other4.53.5 4.32.8 3.84.6 5.04.9 

In multivariate analyses predicting the likelihood of being treated at a high-volume institution and adjusted for clustering, age, year of surgery, CCI, median Zip Code income (fourth vs first quartile), and insurance status (Medicare vs private) achieved independent predictor status (Table 2). Specifically, patients with a CCI of 1, 2 or ≥3 were 18%, 17%, and 19% less likely to be treated at a high-volume institution, respectively (P ≤ .025). Patients from the highest quartile of median Zip Code income were 63% more likely to be treated at a high-volume institution than their counterparts from the lowest quartile of median Zip Code income (P = .001). Moreover, patients with Medicare coverage were 14% less likely to be treated at a high-volume institution than their counterparts with private insurance (P = .008). In these models, lower age and more recent year of surgery also increased the odds of being treated at a high-volume institution (P = .001 and .001, respectively).

Table 2. Multivariate Logistic Regression Analyses With General Estimation Equation Adjustment Assessing Prediction of Being Treated by a High-Volume Institution in Patients Undergoing Radical Prostatectomy, Radical Cystectomy, and Nephrectomy, Nationwide Inpatient Sample, 2003-2007
 OverallRadical ProstatectomyRadical CystectomyNephrectomy
CharacteristicOdds Ratio (95% CI)POdds Ratio (95% CI)POdds Ratio (95% CI)POdds Ratio (95% CI)P
  • Abbreviations: Ref., reference.

  • a

    Includes Asian, Pacific Islander, Native American, and other unspecified.

  • b

    Based on comorbidity as developed by Charlson et al9 and adapted by Deyo et al.10

Age0.994 (0.991-0.998).0010.989 (0.982-0.996).0030.995 (0.984-1.006).3720.996 (0.993-0.999).024
Year of surgery1.334 (1.127-1.580).0011.496 (1.187-1.886).0011.163 (0.964-1.404).1151.093 (0.988-1.210).084
Sex        
 Men    Ref.Ref.Ref.Ref.
 Women    0.941 (0.799-1.109).4700.932 (0.880-0.987).016
Race        
 WhiteRef.Ref.Ref.Ref.Ref.Ref.Ref.Ref.
 Black0.939 (0.757-1.165).5680.912 (0.691-1.203).5130.582 (0.381-0.889).0121.127 (0.891-1.425).319
 Othera0.947 (0.677-1.324).7500.953 (0.623-1.459).8261.235 (0.729-2.093).4320.922 (0.700-1.214).562
 Unknown0.746 (0.448-1.241).2590.774 (0.401-1.493).4450.462 (0.247-0.863).0150.758 (0.540-1.063).108
CCIb        
 0Ref.Ref.Ref.Ref.Ref.Ref.Ref.Ref.
 10.818 (0.766-0.874)<.0010.757 (0.700-0.817)<.0010.639 (0.574-0.712)<.0010.931 (0.868-0.998).043
 20.828 (0.733-0.936).0030.737 (0.624-0.870)<.0010.739 (0.521-1.047).0890.894 (0.781-1.024).105
 ≥30.812 (0.678-0.974).0250.749 (0.575-0.975).0310.499 (0.365-0.681)<.0010.828 (0.723-0.948).006
Median Zip Code income        
 1st quartileRef.Ref.Ref.Ref.Ref.Ref.Ref.Ref.
 2nd quartile0.970 (0.848-1.108).6501.016 (0.853-1.211).8561.012 (0.775-1.322).9270.918 (0.800-1.053).221
 3rd quartile1.083 (0.908-1.291).3771.194 (0.954-1.494).1210.835 (0.636-1.096).1931.001 (0.849-1.179).992
 4th quartile1.627 (1.222-2.168).0011.984 (1.388-2.836)<.0011.012 (0.716-1.503).8451.223 (0.971-1.540).087
Insurance status        
 PrivateRef.Ref.Ref.Ref.Ref.Ref.Ref.Ref.
 Medicaid0.944 (0.641-1.388).7680.843 (0.397-1.789).6570.852 (0.554-1.310).4650.984 (0.799-1.211).876
 Medicare0.863 (0.774-0.962).0080.823 (0.726-0.933).0020.839 (0.636-1.108).2160.914 (0.837-0.997).043
 Other0.740 (0.517-1.057).0980.600 (0.414-0.868).0070.999 (0.573-1.741).9980.909 (0.642-1.286).589

Table 2 also describes separate multivariate models for each of the 3 procedures. For example, with regard to radical cystectomy, black patients were less likely to be treated at high-volume institutions than their white counterparts (odds ratio [OR], 0.582; P = .012). Moreover, with regard to Nx, female patients were less likely to be treated at high-volume institutions than their male counterparts (OR, 0.932; P = .016).

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

Although the importance of the volume-outcome relation is still controversial, evidence of this relation has led many organizations, such as the Leapfrog Group for Patient Safety5 and the National Quality Forum,11 to suggest hospital volume as a marker of quality in the performance of complex surgical procedures. High-volume hospitals are assumed to have structural characteristics associated with better quality of care, and providers in these hospitals are thought to improve their processes of care through experience in providing complex care. However, certain barriers exist in the access to these preferred centers. A recent study by Liu et al demonstrated consistent and robust disparities in the use of high-volume hospitals across 10 complex inpatient procedures, with minorities and lower-income patients less likely to receive care at high-volume centers.6 To date, the extent of these disparities has not been examined in the context of uro-oncologic surgery. In the current article, we assess the effect of important patient and socioeconomic determinants on access to high-volume hospitals for 3 uro-oncologic procedures, namely RP, radical cystectomy, and Nx.

Our results demonstrated several important points. First, a higher proportion of patients were treated at high-volume institutions in the most recent study years. These findings are validated in multivariate analyses, where year of surgery was an independent predictor of access to high-volume centers. Specifically, patients treated in the late period were more likely to be treated at a high-volume center than those in the early period. Taken together, these findings suggest that increased regionalization or selective referral occurred over the study period. Indeed, we corroborate the findings of previous reports on the regionalization of uro-oncologic care.12, 13

Second, direct comparisons of patient characteristics showed important differences between high-volume institutions and other institutions. Patients treated at high-volume centers were younger, had fewer comorbidities, were more likely to be treated in the most recent years, were more likely to be of white race, were more likely to be within the highest quartile of median Zip Code income, and were more likely to hold private insurance. To further elucidate the association between these parameters and access to high-volume institutions, we performed a multivariate adjustment. Generalized estimating equations were used to further adjust for clustering. In these models, we confirmed the existence of important disparities in access to high-volume hospitals. Specifically, age, CCI, income, and insurance status were shown to be independent predictors of being treated at high-volume institutions. Similarly, black race was an independent predictor of decreased utilization of high-volume institutions for radical cystectomy, and female sex was an independent predictor of decreased utilization of high-volume institutions for Nx. Taken together, our study demonstrates that several patient characteristics prevent some individuals from receiving care at centers where optimal outcomes may be expected.

Access to care at high-volume institutions has been directly tied to decreased morbidity and mortality after complex surgical procedures. Begg et al demonstrated a significant reduction in operative mortality after major cancer procedures, namely pancreatectomy, esophagectomy, liver resection, and pelvic exenteration.2 Similar findings were demonstrated in procedures as widely varied as hip fracture treatment,14 knee replacement,15 coronary artery bypass grafting,16 and angioplasty.17 In a landmark 2002 study, Birkmeyer et al found that patients undergoing 6 different types of cardiovascular procedures and 8 types of major cancer resections between 1994 and 1999 could significantly reduce their risk of operative death simply by selecting a high-volume hospital.3 Finally, the same investigators also showed that better cancer-specific survival rates were expected in patients treated at high-volume hospitals.4

The volume-outcomes relation has also been validated in uro-oncologic literature. Konety et al demonstrated a significant association between hospital volume and in-hospital mortality, LOS, and hospital charges after radical cystectomy.18 Hollenbeck et al found that patients treated at low-volume hospitals were nearly 50% more likely to die in the postoperative period (4.9% vs 3.5%; adjusted OR, 1.48) after radical cystectomy.19 Importantly, they identified several processes of care that were more broadly utilized at high-volume institutions; these processes, including preoperative cardiac testing, intraoperative arterial monitoring, the use of a continent diversion, and others, explained 23% of the volume-mortality effect. Although the overall rate of in-hospital mortality is low after prostatectomy, patients at low-volume centers have been found to be 78% more likely to have in-hospital mortality than those at high-volume centers, as well as having increased hospital charges and increased LOS.20

As the body of literature pertaining to the volume-quality relation for uro-oncology procedures develops, it becomes increasingly important to understand the potential barriers to receipt of care at high-volume institutions. These barriers have been demonstrated in other surgical disciplines, but there is a paucity of data concerning urologic procedures. Liu et al demonstrated Medicaid insurance status, low income, and nonwhite race to be predictive of receipt of care at low-volume institutions for 10 complex operations.6 This is largely consistent with findings in the orthopedic surgery literature demonstrating the influence of nonwhite race and Medicaid insurance status on receipt of care at low-volume institutions for hip and knee replacement21, 22 and the impact of race on receipt of cardiovascular care at high-volume hospitals.23 In this study, black patients had a 36% higher risk-adjusted mortality after abdominal aortic aneurysm repair. It was estimated that 26% of the disparity was attributable to socioeconomic factors, but 25% was simply because of receipt of care in lower-quality hospitals.

The consistency of the findings across surgical disciplines would appear to point to broad and important patient- and physician-related factors underpinning these disparities. In 2004, Bach et al reported on the findings of the 2000-2001 Community Tracking Study Physician Survey, which demonstrated that black patients tended to receive lower-quality health care.24 Fewer of the physicians seen by black patients were board-certified, when compared with white patients. Moreover, physicians treating black patients reported facing greater difficulties in obtaining access for their patients to high-quality subspecialists, high-quality diagnostic imaging, and nonemergency admission to the hospital.

Despite its strengths, our study is not devoid of weaknesses. Several differences in patient variables (disease characteristics, personal preferences, education) and socioeconomic determinants may be advanced for the recorded differences. Specifically, patient selection can represent an important confounder. The complexity of the surgery is not captured in this data set. Patients with more aggressive disease may be diverted toward certain types of institutions. More evidence is urgently needed to support the regionalization of complex urologic procedures that is already under way,10 burdening patients from rural areas to travel long distances to obtain medical care.25

To summarize, old, sick, poor, and Medicare patients were less likely, on average, to be treated at high-volume hospitals for uro-oncologic surgery. Similarly, black patients were less likely to have a radical cystectomy at a high-volume hospital, and female patients were less likely to have an Nx at a high-volume hospital. Selective referral of individuals who are less likely to receive care at such institutions may represent a health care priority intended to optimize outcomes across all population strata.

FUNDING SOURCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

P.I.K. is partially supported by the University of Montreal Health Centre Urology Specialists, Fonds de la Recherche en Sante du Quebec, University of Montreal Department of Surgery, and University of Montreal Health Centre Foundation.

CONFLICT OF INTEREST DISCLOSURES

The authors made no disclosures.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES