Radical nephrectomy surgical outcomes in the University HealthSystem Consortium Data Base

Impact of Hospital Case Volume, Hospital Size, and Geographic Location on 40,000 Patients

Authors


Abstract

BACKGROUND:

We sought to determine the impact of radical nephrectomy case volume, hospital size, and geographic region on immediate surgical outcomes for patients undergoing radical nephrectomy in academic centers across the country.

METHODS:

The University HealthSystem Consortium (UHC) Clinical Data Base was queried for data corresponding to patients who underwent radical nephrectomy at 1 of 134 academic medical centers nationwide between 2003 and quarter 2 of 2007 (n = 42,988). Radical nephrectomy case volume (1-99, 100-499, and 500 + ), total discharges (1-49,999, 50,000-99,999, 100,000 + ), and geographic region (5 categories) were determined for each academic center. ANOVA and the Tukey statistic were used. Length of stay, intensive care unit (ICU) rate, complication (comp) rate, and in-hospital mortality were analyzed.

RESULTS:

Case volume was a significant predictor of length of stay, ICU, and comp. Mean length of stay was 6.88, 5.61, and 4.76 days, respectively, for centers from lowest to highest case volumes (P < .001). ICU rates for the 3 tiers were 30.77, 17.93, and 12.22 (P < .001). Comp rates were 24.50, 19.40, and 15.48 (P < .001). Tukey analysis revealed a ceiling effect: No differences were seen between the 2 higher case volume groups. Stratification by total discharges revealed differences in ICU rates (P = .001) and comp rates (P = .001). Region of the country had no significant impact on any of the outcome variables in this study.

CONCLUSIONS:

Radical nephrectomy case volume emerged as an important variable in predicting 3 of the 4 outcome parameters in this study. Results suggest that a “critical volume” of cases portends improved surgical outcomes. Cancer 2009. © 2009 American Cancer Society.

Radical nephrectomy is commonly performed across the United States by urologists across a spectrum of hospital centers, from small community hospitals to large referral centers. We have previously studied the relationship between hospital center characteristics and outcomes after cystectomy.1 For radical cystectomy and prostatectomy, multiple investigators have demonstrated an association between hospital center procedure volume and outcomes after surgery.2, 3 There is also evidence to suggest that overall hospital surgical volume alone has an impact of surgical outcomes.4 These findings have important and far-reaching implications for optimizing patient care and invite debate about how to most effectively export certain processes of care from high-volume centers to lower-volume centers. They also raise questions about the value of regionalization of certain procedures. The current study examines radical nephrectomy-specific hospital volume, hospital size, and geographic region of the country and their impact on outcomes after nephrectomy.

MATERIALS AND METHODS

The University HealthSystem Consortium (UHC) is a member-driven alliance of approximately 90% of the nonprofit academic medical centers in the United States. UHC offers tools and services that support member clinical resource management and quality improvement initiatives. For this study, UHC's Clinical Data Base was used. This electronic repository combines administrative, clinical, and financial data from participating UHC member institutions. Data are gathered from hospital discharge summaries and Uniform Billing-92 data.

The UHC Clinical Data Base was queried using the International Classification of Diseases ninth revision procedure code for nephrectomy (555) as a search criterion. This code captures nephrectomy and laparoscopic nephrectomy. Partial nephrectomy is not included. Data from 134 hospital centers that performed radical nephrectomy were used for this analysis. Data were available from 2003 through the second quarter of 2007. Hospital centers were stratified categorically by the total number of nephrectomy cases performed and hospital discharges (as a surrogate for hospital size) and by geographic region of the country. For case volume, 3 categories were defined, thus splitting the data into approximate tertiles by the number of hospitals in each category (0-99, 100-499, and 500 + ). For hospital discharges, limits were defined as 0-49,999, 50,000-99,999, and 100,000 + , again splitting the data into approximate tertiles. The country was divided into 5 geographic regions as defined by the United States Census Bureau (West, Midwest, Central, Southeast, and Northeast).

Hospital length of stay, intensive care unit admission rate, in-hospital mortality, and complication rate were the outcome variables measured in the study. Complications were defined as those that triggered a separate International Classification of Diseases ninth revision designation postoperatively. Only those complications that occurred during the same hospitalization during which surgery was performed were captured in the analysis. Outcome variables were available as mean values for each medical center. These means were used in the subsequent analyses. Data analysis was conducted using the ANOVA and Tukey statistics. The ANOVA statistic was used to determine whether an overall difference existed between volume categories for a given outcome variable. Where differences existed, the Tukey analysis was used to compare each volume category with every other category. All p value calculations were 2-tailed, and a p value of 0.05 or less was used to determine statistical significance for each analysis.

RESULTS

The study population comprised 42,988 patients who underwent radical nephrectomy at 1 of 134 hospital centers during the period of time encompassed by this study. Hospital center characteristics are displayed in Table 1. It is important to note that that nephrectomy case volume and total hospital discharges are cumulative over the study period. Therefore, a case volume of 100 in this study equates to approximately 20 cases per year, and a case volume of 500 equates to about 110 cases per year, etc.

Table 1. Hospital Center Characteristics
CharacteristicsNo. of Hospitals (%)No. of Cases (%)
Geographic region
 Central18 (13.4)4159 (9.7)
 Midwest33 (24.6)12,330 (28.7)
 Northeast45 (33.6)12,739 (29.6)
 Southeast17 (12.7)6708 (15.6)
 West21 (15.7)7052 (16.4)
Total hospital discharges
 0-49,99936 (26.8)1666 (3.9)
 50,000-99,99940 (29.8)9220 (21.4)
 ≥100,00058 (43.3)32,102 (74.7)
Radical nephrectomy case volume
 0-9948 (35.8)1761 (4.1)
 100-49952 (38.8)15,479 (36.0)
 ≥50034 (25.3)25,748 (59.9)

Results of ANOVA statistics are displayed in Table 2. Nephrectomy case volume was significantly related to length of stay, intensive care unit (ICU) admission rate, and comp rate, but not to in-hospital mortality. Hospital length of stay ranged from 4.76 days at the highest case volume centers, to 6.88 days at the lowest volume centers. Interestingly, ICU admission rates were lowest for the intermediate case volume tier. In increasing order of case volume, ICU rates were 30.77%, 17.93%, and 21.22%. Comp rates were also disparate when stratified by case volume, ranging from 24.5% to 15.48% from lowest to highest case volume. Hospital size (as measured by total hospital discharges) was related both to ICU admission rates and comp rates. As hospital size increased, ICU admission rates and comp rates decreased. Geographic region of the country was not significantly related to any of the outcome variables studied.

Table 2. Outcome Variable Means and Standard Deviations With ANOVA P Values
Case VolumeA. Mean Values By Case VolumeP
0-99100-499500+
  1. SD indicates standard error of the mean (standard deviation); ICU, intensive care unit.

  2. ANOVA for each outcome variable categorized by case volume, hospital size, and geographic region.

Length of stay, Days+/- SD6.88+/-2.965.61+/-1.254.76+/-0.74<.001  
ICU rates, %+/-SD30.77+/-24.3817.93+/-11.6412.22+/-7.57<.001  
Complication rates, %+/-SD24.50+/-13.7419.40+/-5.5315.48+/-5.08<.001  
In-hospital mortality, %+/-SD1.79+/-4.031.75+/-3.561.03+/-1.00.532  
B. Mean Values By Hospital Size
Hospital Size0-49,99950,000-99,999100,000+P  
Length of stay, Days+/-SD6.26+/-2.686.25+/-2.435.32+/-1.29.042  
ICU rates, %+/-SD25.81+/-24.3626.43+/-18.3914.50+/-10.25.001  
Complication rates, %+/-SD25.99+/-14.9919.75+/-7.3917.55+/-5.45.001  
In-hospital mortality, %+/-SD1.54+/-4.501.69+/-1.721.54+/-3.34.972  
C. Mean Values By Region
RegionWestCentralMidwestSoutheastNortheastP
Length of stay, Days+/-SD6.30+/-2.645.83+/-1.745.30+/-1.876.43+/-2.655.83+/-1.96.360
ICU rates, %+/-SD17.21+/-10.7631.97+/-22.1921.92+/-22.3017.17+/-12.5618.59+/-15.91.060
Complication rates, %+/-SD19.43+/-7.2220.49+/-5.6920.89+/-13.4518.14+/-8.0520.81+/-10.02.879
In-hospital mortality, %+/-SD3.40+/-5.591.60+/-1.601.57+/-4.341.47+/-1.400.78+/-1.13.060

After significant differences were detected by the ANOVA statistic, the Tukey statistic was used to probe differences between tiers. Results are displayed qualitatively in Figure 1, and confidence intervals are displayed in Table 3. For length of stay, ICU admission rates, and comp rates, differences were not significant between the higher 2 tiers of hospitals based on case volume. A significant difference was found for each of these 3 variables when comparing outcomes between the lowest and intermediate tier of hospital centers.

Figure 1.

Outcome variables are stratified by length of stay.

Table 3. Tukey 95% Confidence Intervals for Outcomes on the Basis of Case Volume
VariableMeanDifference in Mean95% CI
  • Tukey statistic results and 95% confidence intervals (CI).

  • ICU indicates intensive care unit.

  • *

    Denotes significance.

Length of stay, d   
 Tier 1 (0-99 cases)6.88(1-2) −1.27−2.20 to −.33*
 Tier 2 (100-499 cases)5.61(1-3) −2.11−3.16 to −1.06*
 Tier 3 (≥500 cases)4.76(2-3) −.84−1.87 to .19
ICU rates   
 Tier 130.77(1-2) −12.84−20.93 to −4.76*
 Tier 217.93(1-3) −18.55−27.50 to −9.61*
 Tier 321.22(2-3) −5.71−14.46 to 3.04
Complication rates   
 Tier 124.50(1-2) −5.10−9.51 to −.71*
 Tier 219.37(1-3) −9.02−13.95 to −4.09*
 Tier 315.48(2-3) −3.92−8.77 to .93
In-hospital mortality rates   
 Tier 11.79(1-2) −.04−1.61 to 1.54
 Tier 21.75(1-3) −.76−2.52 to 1.00
 Tier 31.02(2-3) −.72−2.45 to 1.01

DISCUSSION

The current study's findings confirm the positive relationship between case-specific hospital surgical volume and short term outcomes after radical nephrectomy. Specifically, these findings demonstrate an association between nephrectomy case-specific volume and hospital length of stay, comp rates, and ICU admission rates.

With the emergence of large, nationwide databases such as the UHC Clinical Data Base, it has become feasible to study the relationship between hospitals and outcomes on a very large scale. It is important to note that major studies do not always agree on the importance of hospital characteristics in determining surgical outcomes. Depending on the type of surgery and the outcome being studied, data vary widely. Birkmeyer and colleagues studied mortality after a total of 2.5 million major surgical procedures, including 6 types of cardiovascular procedures and 8 major cancer surgeries, including nephrectomy and cystectomy. They found that 30-day or in-hospital mortality after each procedure decreased with increasing surgical volume, but the effect varied depending on the type of surgery.5 Interestingly, they did not find any systemic relationship between length of stay and hospital volume. By using the same 14 high-risk surgical cases and the same nationwide databases, Goodney and colleagues studied the impact of hospital case volume on length of stay and readmission rates within 30 days. They found no relationship between length of stay and hospital volume, nor did they discover any association between volume and readmission rates.6

Focusing specifically on cancer surgery, Finlayson and colleagues studied operative mortality after 8 different procedures, using the Nationwide Inpatient Sample database, encompassing almost 200,000 cases. Their findings were variable: Mortality decreased only for esophagectomy and pancreatic resection. They found no difference for cystectomy and nephrectomy.7 In a unique study of nephrectomy, partial nephrectomy, and nephroureterectomy, Taub and colleagues have found that for all types of nephrectomy combined, mortality rates are lower for high-volume centers. For radical nephrectomy and nephroureterectomy alone, however, there was no detectable difference in mortality across all hospital centers. Length of stay was significantly different only for partial nephrectomy when stratified by hospital volume.8 Konety and colleagues studied urologic surgery exclusively and examined the relationship between procedure-specific volume and outcomes after radical cystectomy, prostatectomy, and nephrectomy. They found a significant decrease in mortality for cystectomy and prostatectomy for patients undergoing surgery in hospitals that perform large numbers of those procedures. The authors did not find any relationship between mortality and specialized urology status or high scoring centers by Leapfrog criteria.9

The current study is unique in that the data being studied are compiled exclusively from academic medical centers, and is more contemporary than virtually any other study relating outcomes to hospital characteristics. It is, to our knowledge, the first such study to be drawn from the UHC Clinical Data Base. It, along with the aforementioned studies, suggests a small but definite association between case specific volume and short-term outcomes after radical nephrectomy.

Which other hospital characteristics are important determinants of outcome? Apart from the aforementioned variables, a potential candidate is specialized hospital status such as a National Cancer Institute (NCI)-designated hospital or a Leapfrog nominee, as mentioned above. Birkmeyer and colleagues studied surgical mortality for 6 major cancer procedures at NCI-designated centers compared with control hospitals. They found significant decreases in mortality for 4 of the 6 major cases. No difference was detected for cystectomy.10 To identify another determinant of outcome, Gilbert and colleagues investigated the impact of nonindex case volume on the outcome of major urologic surgeries. The authors detected an attenuation in mortality rates after cystectomy and prostatectomy when adjusting for a hospital's volume of other major urologic surgeries.11 Their conclusion was that process of care may be superior at centers that specialize in multiple urologic surgeries than those that specialize in only 1 or none. These characteristics were not captured in the current study.

As Gilbert and others have noted, identifying factors that are associated with improved outcomes is only the first step in the process of quality improvement. Regionalization of care and care process exportation have been proposed and studied as vehicles by which to translate findings from studies such as these into strides toward better patient care.

It must be recognized, however, that the outcome metrics in this and other studies are not synonymous with quality. Finlayson clarifies that what physicians see as quality and what patients see as quality may be quite disparate. Moreover, he reminds us that quality improvement should be the “process of delivering quality care to patients, not the other way around.”12 Summarizing the potential benefits and pitfalls of regionalization, Birkmeyer notes that especially for rural-dwelling patients, regionalization may adversely impact continuity of care, and could potentially lead to underutilization of surgery in rural areas with a decrease in the ability of rural hospitals and surgeons to handle emergency situations.13

Birkmeyer and colleagues point out that there are no definite blueprints for the ideal process of care, but that patients at high-volume centers with relatively good outcomes are more likely to undergo preoperative stress testing and to have consulted with an oncologist before surgery.14 Although Birkmeyer did not find any association between the use of specific processes of care and postoperative outcomes, other evidences suggests that processes of care can be successfully instituted with measurable improvements in postoperative outcome metrics. We have previously described the successful institution of a clinical care pathway for radical prostatectomy. The study suggested patients can be safely transitioned to a 2-day length of stay from a 3-day pathway. The use of standardized diet and pain control orders, with education by nurse specialists, allowed for a safe transition. Since the time of that publication, the pathway has been further transitioned to a 1-day length of stay for most patients.15 Although length of stay is not a surrogate for quality of care after nephrectomy, such examples of the implementation and success of new processes of care should be modeled in the effort to optimize the perioperative care of nephrectomy patients.

This study has several important limitations that must be recognized. It was not possible to characterize outcomes by individual surgeon or patient health status. A patient's long-term outcome is not captured, making it difficult to control for the comorbidity profiles across hospital centers. Finally, data were available only as summaries (mean values) for each hospital center. Individual data points for each patient were not available. Each dependent variable was analyzed as a mean value for each medical center, and values were not weighted by case volume, hospital volume, or standard deviation.

Conclusion

This study strengthens a growing body of evidence to suggest that hospital case volume for major surgical procedures has a significant impact on postoperative outcomes. This is only one of the studies in the literature that examines the volume-outcome relationship for specifically for radical nephrectomy. Length of stay, comp rates, and ICU admission rates are more favorable for patients who have surgery performed at relatively high-volume centers. These endpoints are not surrogates for quality, however, and the findings of this study do not provide any evidence that regionalization of radical nephrectomy care would necessarily improve patient care. These results do necessitate further investigation into the attributes of high-volume centers that allow for shorter length of stay and avoidance of complications postoperatively. Those processes that are unique to high-volume centers should be identified and, when feasible, instituted at lower-volume hospitals.

Conflict of Interest Disclosures

The authors made no disclosures.

Ancillary