- Email: [email protected]
CONTEMPORARY TRENDS IN SURGICAL CORRECTION OF PEDIATRIC URETEROPELVIC JUNCTION OBSTRUCTION: DATA FROM THE NATIONWIDE INPATIENT SAMPLE
0022-5347/05/1731-0232/0 THE JOURNAL OF UROLOGY® Copyright © 2005 by AMERICAN UROLOGICAL ASSOCIATION
Vol. 173, 232–236, January 2005 Printed in U.S.A.
DOI: 10.1097/01.ju.0000148439.22885.b4
CONTEMPORARY TRENDS IN SURGICAL CORRECTION OF PEDIATRIC URETEROPELVIC JUNCTION OBSTRUCTION: DATA FROM THE NATIONWIDE INPATIENT SAMPLE CALEB P. NELSON, JOHN M. PARK, RODNEY L. DUNN
AND
JOHN T. WEI*
From the Department of Urology, University of Michigan, Ann Arbor, Michigan, and Brady Urological Institute, Johns Hopkins Hospital, Baltimore, Maryland (CPN)
ABSTRACT
Purpose: The presentation and diagnosis of ureteropelvic junction obstruction have changed in the last 20 years. We describe trends in pediatric surgical correction of ureteropelvic junction obstruction between 1988 and 2000. Materials and Methods: The Nationwide Inpatient Sample contains data on approximately 5 million to 7 million hospital inpatient stays per year, approximating a 20% sample of United States hospitals. We used International Classification of Disease-9 codes to identify pediatric pyeloplasty cases, and analyzed the data for practice patterns. Results: A total of 5,858 pediatric patients (mean age 62.8 months) underwent pyeloplasty. Males comprised 70.7% of the sample, and tended to undergo surgery at a younger age (60.1 vs 69.4 months, p ⬍0.0001). The proportion of procedures done during the first 6 months of life decreased from 34.2% (1988 to 1991) to 25.2% (1997 to 2000, p ⬍0.0001). Nonwhites underwent surgery in the first 6 months more often than whites (38.9% vs 25.0%, p ⬍0.0001) and had a lower mean age at surgery (44.4 vs 70.7 months, p ⬍0.0001). The percentage of procedures done at urban teaching hospitals increased from 48.9% (1988 to 1991) to 61.3% (1997 to 2000, p ⬍0.0001). Length of stay decreased significantly from 6.7 days (1988 to 1991) to 3.7 days (1997 to 2000, p ⬍0.0001). Conclusions: Practice patterns in pediatric pyeloplasty evolved between 1988 and 2000. Fewer procedures are being performed in newborns, suggesting that patients with prenatal hydronephrosis are increasingly being observed instead of undergoing early surgery. There was a substantial difference in timing of surgery between whites and nonwhites. To our knowledge this observation has not previously been reported. More procedures are being performed at teaching hospitals, and length of stay has decreased significantly. KEY WORDS: kidney pelvis, physician’s practice patterns, socioeconomic factors
Surgery for the correction of ureteropelvic junction (UPJ) obstruction has evolved with time. The use of routine prenatal sonographic examination has raised questions about the optimal timing of corrective surgery,1, 2 endoscopic and laparoscopic technologies have been introduced and used widely with uncertain risks and benefits,3 and developments in surgical and postoperative care have resulted in changes in length of hospital stay (LOS) and other elements of care. Pyeloplasty is a common procedure in pediatric patients but there has been little information published regarding nationwide practice patterns with respect to this procedure. Previous evaluations of pyeloplasty trends have included mailed surveys of endourologists but such efforts are limited by the small numbers and narrow perspective of the respondents.4 National administrative databases provide a source for investigation of such trends, and have been applied toward investigation of other procedures such as surgery for gastroesophageal reflux, providing a broad overview of practice patterns and how use of the procedure is changing.5 The purpose of this study was to use the nationally representative Healthcare Cost and Utilization Project Nationwide Inpatient Sample (NIS) to evaluate practice patterns in pe-
diatric pyeloplasty between 1988 and 2000, and to analyze the patient and hospital factors associated with these trends. MATERIALS AND METHODS
The Healthcare Cost and Utilization Project NIS is the largest all payer inpatient care database that is publicly available in the United States. Each year the NIS provides information on approximately 5 million to 7 million inpatient stays from about 1,000 hospitals, and is designed to approximate a 20% sample of United States nonfederal community hospitals, including specialty hospitals, public hospitals and academic medical centers. Excluded are short-term rehabilitation, long-term, psychiatric and chemical dependence treatment facilities. The data available include patient specific information, including age, diagnosis, procedures and disposition, while a separate data set includes information about each participating hospital. The year specific data sets for 1988 through 2000 were merged to create a single large data set. To identify all pyeloplasty cases, we extracted all hospital discharges of patients for whom the International Classification of Disease (ICD)-9 procedure code 5587 (“correction of ureteropelvic junction”) appeared in any of the 15 procedure code variables for the discharge. If the code was not present, then the patient was considered to have been discharged from the hospital without pyeloplasty and was not included in the analysis. We then selected those patients who were younger than 18 years at the time of hospitalization.
Submitted for publication May 26, 2004. Supported by Grant NIH-1-T-32 DK07782. * Correspondence: Department of Urology, Taubman Center 2916, Box 0330, 1500 E. Medical Center Dr., Ann Arbor, Michigan 481090330 (e-mail: [email protected]). 232
233
CONTEMPORARY TRENDS IN PEDIATRIC PYELOPLASTY
The primary outcome variables that we examined were age in months at the time of hospitalization, proportion of patients undergoing surgery before age 6 months, proportion of patients undergoing surgery at a teaching hospital and length of hospital stay in days. Hospital charges (adjusted for inflation to year 2000 dollars using published conversion factors) were also examined.6 A hospital is defined as a teaching hospital by the NIS if it has an American Medical Association approved residency program, is a member of the Council of Teaching Hospitals, or has a resident-to-bed ratio of 0.25 or higher. Because the definition of a teaching hospital by the NIS changed beginning in 1998, longitudinal analysis of teaching hospital status is not possible using the complete data set. Therefore, we used a subset of the data using a uniform definition of teaching hospitals. First, we identified all hospitals that were included in the NIS during 1 or more years between 1988 and 1997, and that had the same teaching status during every year in which the hospital was included. Next, we selected those patients (from the complete data set) who were discharged from one of these hospitals. The teaching hospital status of each discharge from the 1998 to 2000 period was defined using the 1988 to 1997 definition of that hospital. This approach decreased the number of pyeloplasty procedures in the analysis from 5,858 to 4,362. The analysis of this data set was not weighted due to the skewed sampling required to obtain the uniform teaching hospital definition. A variety of covariates were examined. Socioeconomic status was estimated by median income of patient home ZIP code, grouped as less than $25,000, $25,000 to $34,999, $35,000 to $44,999 and more than $45,000. Insurance status was categorized as insured (private/health maintenance organization [HMO]/Medicare), Medicaid/other (Medicaid, other coverage and no charge) or self-pay (no coverage). The level of comorbidity was determined by the total number of diagnoses at discharge from the hospital, including UPJ obstruction. Hospital data in the NIS include teaching status, urban vs rural status and region (West, South, Midwest and East). Hospital size is categorized as small, medium or large, as defined by the NIS. The definitions of these categories vary by region and rural or urban location. Due to the fact that several states did not provide race/ethnicity information in some years, for analytical purposes we assigned a discrete “race unavailable” value to all subjects for whom a race category was not provided. This approach allowed these subjects, who would otherwise be excluded from the multivariate models due to missing values, to be included, greatly increasing the power of the model. These individuals comprised 33.7% of the subjects in the sample. Surgical volume was determined by computing the number of pyeloplasty procedures done by each hospital in the sample for each given year of participation. These “hospital-volumeyear” statistics were then sorted in descending order and divided into quartiles according to the cumulative case volume.7 Pyeloplasties done at a hospital performing 16 or more procedures during that year were considered to have been done at a “high volume” center. Pyleoplasties done at a hospital in the lower 3 quartiles were considered to have been done at a “low volume” center. Bivariate and multivariate analyses were carried out using SAS statistical software (SAS Institute, Cary, North Carolina). To provide accurate national estimates of trends in outcomes, analyses were weighted. Comparison of means was achieved using t tests and ANOVA. Individual categorical variables were tested for associations using chi-square tests. The time (year) variable was analyzed as an “era” variable, with the study period divided into 3 eras—1988 to 1991, 1992 to 1996 and 1997 to 2000. Significance level of 0.05 was applied for hypothesis testing.
RESULTS
The NIS contained data on 5,858 hospital admissions of subjects younger than 18 years from 1988 through 2000 during which pyeloplasty was performed. The demographic characteristics of the sample are shown in table 1. Weighted mean patient age at hospitalization was 62.8 ⫾ 162.3 months. Of the patients 70.7% were male. This proportion did not vary by year (p ⫽ 0.5660), and varied only slightly by race (70.8% male among whites vs 72.7% male among nonwhites, p ⫽ 0.0003) and region (69.1% in the Northeast vs 71.8% male in the South, p ⫽ 0.0008). Most patients were generally healthy. Median number of diagnoses at discharge from the hospital was 2.0 (mean 2.2 ⫾ 3.8). Weighted inhospital mortality among the sample was low at 0.02%. Age at time of surgery. Weighted mean age at time of surgery decreased moderately during the study period from 64.0 months during 1988 to 1991 to 61.2 months during 1997 to 2000 (table 2). Males tended to undergo surgery almost a year earlier than females (60.4 vs 70.8 months, p ⬍0.0001). Similarly, nonwhite patients also tended to undergo surgery significantly earlier than white patients (44.7 vs 69.6 months, p ⬍0.0001). There were also dramatic differences in age at surgery between high and low volume centers, with high volume centers operating almost 30 months earlier on average (p ⬍0.0001). Surgery during infancy. We found that the weighted proportion of procedures performed in children younger than 6 months decreased during the study period from 34.2% (1988 to 1991) to 25.2% (1997 to 2000), even after adjusting for several patient and hospital factors associated with early
TABLE 1. Demographic characteristics Variables
Raw No. Pts (%)
Weighted %
Pt age: Less than 1 yr 2,422 (41.3) 42.0 1–6 Yrs 1,485 (25.3) 25.1 7–12 Yrs 1,110 (18.9) 18.8 13–18 Yrs 841 (14.4) 14.1 Gender: M 4,119 (70.3) 70.7 F 1,738 (29.7) 29.3 Geographic region: South 2,092 (35.7) 39.7 Midwest 1,316 (22.5) 22.4 Northeast 1,249 (21.3) 20.2 West 1,201 (20.5) 17.7 Socioeconomic status (median ZIP code income): $45,000⫹ 1,135 (21.6) 20.7 $35,000–$44,999 1,293 (24.7) 24.0 $25,000–$34,999 1,790 (34.1) 34.9 $1,000–$24,999 1,027 (19.6) 20.4 Insurance coverage: Private/HMO/Medicare 3,826 (65.9) 65.9 Medicaid/other 1,861 (32.1) 32.1 None (self-pay) 115 (2.0) 2.0 Race/ethnicity: White 2,808 (72.3) 72.5 Hispanic 483 (12.4) 12.2 Black 354 (9.1) 9.4 Other 153 (3.9) 3.9 Asian/Pacific 75 (1.9) 1.8 Native American 12 (0.3) 0.3 Hospital location: Rural 247 (4.2) 4.5 Urban nonteaching 2,060 (35.2) 32.3 Urban teaching 3,547 (60.5) 63.2 Hospital size: Large 3,533 (60.3) 58.8 Medium 1,395 (23.8) 21.7 Small 926 (15.8) 19.5 Data set consists of pediatric patients undergoing surgery for correction of ureteropelvic junction obstruction, including raw counts and percentages for study sample. Weighted percentages represent national population based estimates of proportion of pediatric patients undergoing surgery for UPJ obstruction in each category.
234
CONTEMPORARY TRENDS IN PEDIATRIC PYELOPLASTY
TABLE 2. Age at time of surgery for correction of UPJ obstruction Variables
Weighted Mean Age at Surgery (mos)
Weighted p Value*
Era: 1988–1991 64.0 0.0168 1992–1996 63.2 1997–2000 61.2 Gender: M 60.1 ⬍0.0001 F 69.4 Race/ethnicity: White 70.7 ⬍0.0001 Nonwhite 44.4 Not available 61.9 No. diagnoses at discharge: More than 2 65.7 0.0099 0–2 61.6 Insurance coverage: Private/HMO/Medicare 70.3 ⬍0.0001 Medicaid/other 46.1 Self-pay 78.1 Hospital case vol: High (top quartile) 40.6 ⬍0.0001 Low (bottom quartiles) 70.4 Hospital type and location: Rural 134.1 ⬍0.0001 Urban nonteaching 74.0 Urban teaching 52.1 Hospital size: Small 50.4 ⬍0.0001 Medium 73.4 Large 63.1 Patient and hospital factors associated with age at time of surgery for correction of ureteropelvic junction obstruction, weighted to national population (nonsignificant covariates excluded from multivariate model are socioeconomic status (median income of home ZIP code) and geographic region). * Adjusted for all other significant covariates (multivariate linear regression).
surgery (table 3). Mirroring the findings for overall age at surgery, nonwhite patients were significantly more likely to undergo surgery during the first 6 months (38.9% vs 25.0%, p ⬍0.0001). Regionalization of care. The proportion of procedures done at teaching hospitals increased significantly with time from 48.9% in 1988 to 1991 to 61.3% in 1997 to 2000 (table 4). This trend was present after adjusting for several other factors that were also associated with increased odds of a procedure being done at a teaching hospital, including nonwhite race, early surgery (before 6 months) and Northeast geographic region. Length of stay. Weighted mean length of hospital stay decreased from 6.6 days in 1988 to 1991 to 3.7 days in 1997 to 2000, and this change was significant (p ⬍0.0001) even when adjusting for other patient and hospital factors (see figure). As one would expect, patients with greater co-morbidity (more than 2 diagnoses) had longer LOS than healthier patients (7.6 vs 3.9 days, p ⬍0.0001). Longer hospital stays were also seen for nonwhite compared to white patients (6.1 vs 4.6 days, p ⬍0.0001) as well as for Medicaid patients compared to private/HMO patients (6.0 vs 4.5 days, p ⬍0.0001). Other factors associated with longer LOS included urban teaching hospital and lower socioeconomic status. Gender, hospital volume and geographic region were not significantly associated with LOS. Interestingly, weighted mean inflation adjusted total hospital charges did not change significantly during the study period (p ⫽ 0.8451) despite the dramatic decrease in length of stay (see figure). DISCUSSION
A key finding from these data is the dramatic decrease in the proportion of patients undergoing pyeloplasty as infants. In fact, the odds of undergoing surgery for UPJ obstruction during the first 6 months of life decreased by more than 50%
between 1988 and 2000. This finding suggests that, increasingly, surgical intervention for prenatally detected UPJ obstruction is being delayed in favor of observation. The debate over the optimal management of prenatal hydronephrosis has been ongoing since the introduction of practical prenatal sonographic screening in the 1980s.8 Early in the debate a number of authors advocated early intervention to preserve renal function.1 However, increasingly, observation has been recommended for most infants, as many seem to do well without aggressive surgical intervention, and many urologists prefer to avoid procedures in these young patients.2 This evolution in the literature is consistent with the trend that we observed in our analysis, as a gradual but definite shift away from pyeloplasty in newborns and infants. Another key finding of this study with respect to timing of surgery is that significant differences exist between white and nonwhite patients. Nonwhite patients underwent surgery more than 2 years earlier than whites, and nonwhite patients were much more likely to undergo surgery during the first 6 months of life. To our knowledge this is the first report of such a marked difference along racial or ethnic lines in the management of UPJ obstruction. Vesicoureteral reflux is reported to be rare among black children, particularly those with prenatal hydronephrosis, but there is little in the literature to support the existence of racial difference in the incidence or natural history of UPJ obstruction.9 It is important to note that the racial difference in timing of surgery persisted after adjustment for socioeconomic status and insurance status. This finding raises questions with respect to other factors that may be driving the differential timing of intervention in white and nonwhite children. Racial disparity in medical treatment has been demonstrated for a number of conditions, and our findings represent further evidence of such disparity.10, 11 The data reveal a significant shift of caseload away from rural and nonteaching hospitals to teaching institutions. The reasons for this trend are likely multifactorial. Changes in the financial reimbursement environment may have made it less cost-efficient for private practitioners to perform highly subspecialized procedures such as pediatric pyeloplasty. Another possible factor is the emergence of laparoscopic pyeloplasty. While the number of urologists who perform laparoscopic pyeloplasty is growing, they still represent a minority, which may mean that more patients migrate to teaching institutions in search of this or other minimally invasive techniques.12 The decrease in length of hospital stay after pyeloplasty reflects a wider health care trend. Decreasing length of stay has been documented for a variety of medical and surgical conditions, including those in pediatric urology.13 A wide range of factors have been reported to have an impact on length of inpatient stay, including use of care pathways and protocols, specialty consultation, level of use of imaging technology, legislation and evolution of surgical technique.13–17 Our study shows that race and socioeconomic status also are associated with LOS. The growing use of laparoscopic and other minimally invasive techniques may also have had an impact on length of hospital stay after pyeloplasty. Laparoscopic procedures are commonly promoted as resulting in shorter hospital stays, and this has been demonstrated to be the case with pyeloplasty.18 However, since the ICD-9 codes for correction of ureteropelvic junction obstruction do not differentiate between open and laparoscopic procedures, we were unable to examine the effect of laparoscopic techniques on length of stay. Increasing use of laparoscopy and other technology intensive methods may also be responsible for our finding that despite the substantial decrease in length of hospital stay, total inflation adjusted hospital charges essentially remained constant during the study period. This finding is particularly
CONTEMPORARY TRENDS IN PEDIATRIC PYELOPLASTY
235
TABLE 3. Surgery during newborn period and infancy Variables
Weighted % Pts Younger Than 6 Mos
Weighted OR (95% CI)*
Weighted p Value*
Era: 1988–1991 34.2 1.0 ⬍0.0001 1992–1996 29.8 0.78 (0.73–0.84) 1997–2000 25.2 0.48 (0.45–0.52) Gender: F 24.6 1.0 ⬍0.0001 M 32.0 1.48 (1.39–1.57) Race/ethnicity: White 25.0 1.0 ⬍0.0001 Nonwhite 38.9 1.57 (1.46–1.70) Not available 31.3 1.08 (1.00–1.16) No. diagnoses at discharge: 0–4 28.7 1.0 ⬍0.0001 More than 4 41.6 1.79 (1.63–1.96) Insurance coverage: Medicaid/other 36.6 1.0 ⬍0.0001 Private/HMO/Medicare 26.7 0.71 (0.67–0.76) Self-pay 22.9 0.59 (0.48–0.74) Socioeconomic status (median ZIP code income): $1,000–$24,999 30.5 1.0 ⬍0.0001 $25,000–$34,999 30.9 1.14 (1.06–1.23) $35,000–$44,999 26.9 1.02 (0.93–1.11) $45,000⫹ 30.7 1.36 (1.24–1.50) Geographic region: Midwest 24.4 1.0 0.0042 South 32.0 1.07 (0.99–1.16) Northeast 31.7 1.15 (1.05–1.26) West 29.8 1.16 (1.06–1.27) Hospital vol: Low 26.5 1.0 ⬍0.0001 High 39.7 1.84 (1.72–1.96) Hospital type and location: Rural 3.7 1.0 ⬍0.0001 Urban nonteaching 26.1 5.59 (4.21–7.42) Urban teaching 33.7 8.98 (6.78–11.9) Hospital size: Large 29.4 1.0 Small 36.2 0.85 (0.78–0.91) ⬍0.0001 Medium 24.9 0.70 (0.65–0.75) Patient and hospital factors associated with surgery for correction of ureteropelvic junction obstruction before age 6 months, weighted to national population. * Adjusted for all other significant covariates.
TABLE 4. Surgery at teaching hospitals Variables
Teaching Hospital %
OR (95% CI)*
p Value*
Era: 1988–1991 48.9 1.0 ⬍0.0001 1992–1996 57.4 1.24 (1.05–1.46) 1997–2000 61.3 1.65 (1.38–1.97) Race/ethnicity: White 55.3 1.0 ⬍0.0001 Nonwhite 66.6 1.81 (1.49–2.18) Not available 52.0 0.98 (0.85–1.14) Age at surgery: 6 Mos or older 51.9 1.0 ⬍0.0001 Younger than 6 mos 65.4 1.80 (1.56–2.08) Geographic region: West 35.5 1.0 ⬍0.0001 South 56.0 2.32 (1.94–2.77) Midwest 54.4 2.72 (2.24–3.31) Northeast 79.9 7.87 (6.35–9.74) Patient factors associated with surgery for correction of ureteropelvic junction obstruction being done at teaching hospital (nonsignificant covariates excluded from multivariate model are gender, comorbidity, insurance status and socioeconomic status). * Adjusted for all other significant covariates.
Trends through time in length of stay (LOS) and hospital charges for surgical correction of UPJ obstruction. Note decreasing LOS during study period, while inflation adjusted total hospital charges (Charges) remain statistically unchanged.
revealing, given the strong correlation of charges with length of stay (p ⬍0.0001). The fact that inflation adjusted charges failed to decrease in an environment of markedly decreasing length of stay suggests that other powerful cost pressures are driving up the overall cost of surgical care. These factors may include increased hospital overhead and personnel costs, increased equipment and pharmaceutical costs, and increasing malpractice insurance costs.19, 20
Although this is one of the first known studies to examine surgical correction of UPJ obstruction at a population level, several limitations need to be considered. Our definition of a patient who underwent pyeloplasty relies on the ICD-9 procedure coding system. It is possible that coding may have been incomplete or inaccurate. In addition, as noted previously, we are limited in our ability to differentiate the type of surgical procedure performed, specifically regarding the use
236
CONTEMPORARY TRENDS IN PEDIATRIC PYELOPLASTY
of minimally invasive techniques such as laparoscopy. We also are limited in our ability to evaluate complication rates, particularly delayed presentations, since the data can only provide diagnoses during the operative hospitalization itself. Length of hospital stay is a good proxy measure for initial surgical outcome, and we have used it as such, but this variable does not provide information about the subsequent clinical course. Finally, there was a relatively high proportion of subjects with missing race/ethnicity data. To accommodate for this problem, we elected to apply a separate categorical value for observations with missing race data. This approach allowed these subjects to be included in the final multivariate models, while acknowledging that they did not fit into a defined race category. CONCLUSIONS
Pediatric pyeloplasty has evolved in significant ways between 1988 and 2000, and these changes are reflected in national practice patterns. Although mean age at surgery decreased slightly during the study period, a significantly smaller proportion of procedures are being performed during the first 6 months of life. This finding suggests that patients with a diagnosis of prenatal hydronephrosis are increasingly being observed instead of undergoing surgery in the newborn period and infancy. There are marked racial differences in the timing of surgery, with white patients being less likely to undergo surgery during the first 6 months of life, and having a higher mean age at surgery. More cases are being done at teaching hospitals, suggesting that regionalization of care is occurring. Length of hospital stay decreased markedly during the study period, which is consistent with broader trends in medicine, although inflation adjusted hospital charges failed to decrease significantly as would be expected from shorter hospital stays. REFERENCES
1. Koyle, M. A. and Ehrlich, R. M.: Management of ureteropelvic junction obstruction in neonate. Urology, 31: 496, 1988 2. Onen, A., Jayanthi, V. R. and Koff, S. A.: Long-term followup of prenatally detected severe bilateral newborn hydronephrosis initially managed nonoperatively. J Urol, 168: 1118, 2002 3. Figenshau, R. S., Clayman, R. V., Colberg, J. W., Coplen, D. E., Soble, J. J. and Manley, C. B.: Pediatric endopyelotomy: the Washington University experience. J Urol, 156: 2025, 1996 4. Marcovich, R., Jacobson, A. I., Aldana, J. P., Lee, B. R. and Smith, A. D.: Practice trends in contemporary management of adult ureteropelvic junction obstruction. Urology, 62: 22, 2003
5. Finlayson, S. R., Laycock, W. S. and Birkmeyer, J. D.: National trends in utilization and outcomes of antireflux surgery. Surg Endosc, 17: 864, 2003 6. Sahr, R.: Inflation Conversion Factors for Dollars 1665 to Estimated 2013. Available at http://oregonstate.edu/dept/pol_sci/ fac/sahr/sahr.htm. Accessed October 26, 2003 7. Taub, D. A., Miller, D. C., Cowan, J. A., Dimick, J. B., Montie, J. E. and Wei, J. T.: Impact of surgical volume on mortality and length of stay after nephrectomy. Urology, 63: 862, 2004 8. Pocock, R. D., Witcombe, J. B., Andrews, H. S., Berry, P. J. and Frank, J. D.: The outcome of antenatally diagnosed urological abnormalities. Br J Urol, 57: 788, 1985 9. Horowitz, M., Gershbein, A. B. and Glassberg, K. I.: Vesicoureteral reflux in infants with prenatal hydronephrosis confirmed at birth: racial differences. J Urol, 161: 248, 1999 10. Underwood, W., DeMonner, S., Ubel, P., Fagerlin, A., Sanda, M. G. and Wei, J. T.: Racial/ethnic disparities in the treatment of localized/regional prostate cancer. J Urol, 171: 1504, 2004 11. Ferguson, J. A., Weinberger, M., Westmoreland, G. R., Mamlin, L. A., Segar, D. S., Greene, J. Y. et al: Racial disparity in cardiac decision making: results from patient focus groups. Arch Intern Med, 158: 1450, 1998 12. Gerber, G. S.: Trends in endourologic practice. J Endourol, 16: 347, 2002 13. McMullin, N., Khor, T. and King, P.: Internal ureteric stenting following pyeloplasty reduces length of hospital stay in children. Br J Urol, 72: 370, 1993 14. Dy, S. M., Garg, P. P., Nyberg, D., Dawson, P. B., Pronovost, P. J., Morlock, L. et al: Are critical pathways effective for reducing postoperative length of stay? Med Care, 41: 637, 2003 15. Wagner, A. K., Fabio, T., Zafonte, R. D., Goldberg, G., Marion, D. W. and Peitzman, A. B.: Physical medicine and rehabilitation consultation: relationships with acute functional outcome, length of stay, and discharge planning after traumatic brain injury. Am J Phys Med Rehabil, 82: 526, 2003 16. Fleszler, F., Friedenberg, F., Krevsky, B., Friedel, D. and Braitman, L. E.: Abdominal computed tomography prolongs length of stay and is frequently unnecessary in the evaluation of acute pancreatitis. Am J Med Sci, 325: 251, 2003 17. Madlon-Kay, D. J., DeFor, T. A. and Egerter, S.: Newborn length of stay, health care utilization, and the effect of Minnesota legislation. Arch Pediatr Adolesc Med, 157: 579, 2003 18. Baldwin, D. D., Dunbar, J. A., Wells, N. and McDougall, E. M.: Single-center comparison of laparoscopic pyeloplasty, Acucise endopyelotomy, and open pyeloplasty. J Endourol, 17: 155, 2003 19. Studdert, D. M., Mello, M. M. and Brennan, T. A.: Medical malpractice. N Engl J Med, 350: 283, 2004 20. Goetghebeur, M. M., Forrest, S. and Hay, J. W.: Understanding the underlying drivers of inpatient cost growth: a literature review. Am J Manag Care, 9: SP3, 2003
Vol. 173, 232–236, January 2005 Printed in U.S.A.
DOI: 10.1097/01.ju.0000148439.22885.b4
CONTEMPORARY TRENDS IN SURGICAL CORRECTION OF PEDIATRIC URETEROPELVIC JUNCTION OBSTRUCTION: DATA FROM THE NATIONWIDE INPATIENT SAMPLE CALEB P. NELSON, JOHN M. PARK, RODNEY L. DUNN
AND
JOHN T. WEI*
From the Department of Urology, University of Michigan, Ann Arbor, Michigan, and Brady Urological Institute, Johns Hopkins Hospital, Baltimore, Maryland (CPN)
ABSTRACT
Purpose: The presentation and diagnosis of ureteropelvic junction obstruction have changed in the last 20 years. We describe trends in pediatric surgical correction of ureteropelvic junction obstruction between 1988 and 2000. Materials and Methods: The Nationwide Inpatient Sample contains data on approximately 5 million to 7 million hospital inpatient stays per year, approximating a 20% sample of United States hospitals. We used International Classification of Disease-9 codes to identify pediatric pyeloplasty cases, and analyzed the data for practice patterns. Results: A total of 5,858 pediatric patients (mean age 62.8 months) underwent pyeloplasty. Males comprised 70.7% of the sample, and tended to undergo surgery at a younger age (60.1 vs 69.4 months, p ⬍0.0001). The proportion of procedures done during the first 6 months of life decreased from 34.2% (1988 to 1991) to 25.2% (1997 to 2000, p ⬍0.0001). Nonwhites underwent surgery in the first 6 months more often than whites (38.9% vs 25.0%, p ⬍0.0001) and had a lower mean age at surgery (44.4 vs 70.7 months, p ⬍0.0001). The percentage of procedures done at urban teaching hospitals increased from 48.9% (1988 to 1991) to 61.3% (1997 to 2000, p ⬍0.0001). Length of stay decreased significantly from 6.7 days (1988 to 1991) to 3.7 days (1997 to 2000, p ⬍0.0001). Conclusions: Practice patterns in pediatric pyeloplasty evolved between 1988 and 2000. Fewer procedures are being performed in newborns, suggesting that patients with prenatal hydronephrosis are increasingly being observed instead of undergoing early surgery. There was a substantial difference in timing of surgery between whites and nonwhites. To our knowledge this observation has not previously been reported. More procedures are being performed at teaching hospitals, and length of stay has decreased significantly. KEY WORDS: kidney pelvis, physician’s practice patterns, socioeconomic factors
Surgery for the correction of ureteropelvic junction (UPJ) obstruction has evolved with time. The use of routine prenatal sonographic examination has raised questions about the optimal timing of corrective surgery,1, 2 endoscopic and laparoscopic technologies have been introduced and used widely with uncertain risks and benefits,3 and developments in surgical and postoperative care have resulted in changes in length of hospital stay (LOS) and other elements of care. Pyeloplasty is a common procedure in pediatric patients but there has been little information published regarding nationwide practice patterns with respect to this procedure. Previous evaluations of pyeloplasty trends have included mailed surveys of endourologists but such efforts are limited by the small numbers and narrow perspective of the respondents.4 National administrative databases provide a source for investigation of such trends, and have been applied toward investigation of other procedures such as surgery for gastroesophageal reflux, providing a broad overview of practice patterns and how use of the procedure is changing.5 The purpose of this study was to use the nationally representative Healthcare Cost and Utilization Project Nationwide Inpatient Sample (NIS) to evaluate practice patterns in pe-
diatric pyeloplasty between 1988 and 2000, and to analyze the patient and hospital factors associated with these trends. MATERIALS AND METHODS
The Healthcare Cost and Utilization Project NIS is the largest all payer inpatient care database that is publicly available in the United States. Each year the NIS provides information on approximately 5 million to 7 million inpatient stays from about 1,000 hospitals, and is designed to approximate a 20% sample of United States nonfederal community hospitals, including specialty hospitals, public hospitals and academic medical centers. Excluded are short-term rehabilitation, long-term, psychiatric and chemical dependence treatment facilities. The data available include patient specific information, including age, diagnosis, procedures and disposition, while a separate data set includes information about each participating hospital. The year specific data sets for 1988 through 2000 were merged to create a single large data set. To identify all pyeloplasty cases, we extracted all hospital discharges of patients for whom the International Classification of Disease (ICD)-9 procedure code 5587 (“correction of ureteropelvic junction”) appeared in any of the 15 procedure code variables for the discharge. If the code was not present, then the patient was considered to have been discharged from the hospital without pyeloplasty and was not included in the analysis. We then selected those patients who were younger than 18 years at the time of hospitalization.
Submitted for publication May 26, 2004. Supported by Grant NIH-1-T-32 DK07782. * Correspondence: Department of Urology, Taubman Center 2916, Box 0330, 1500 E. Medical Center Dr., Ann Arbor, Michigan 481090330 (e-mail: [email protected]). 232
233
CONTEMPORARY TRENDS IN PEDIATRIC PYELOPLASTY
The primary outcome variables that we examined were age in months at the time of hospitalization, proportion of patients undergoing surgery before age 6 months, proportion of patients undergoing surgery at a teaching hospital and length of hospital stay in days. Hospital charges (adjusted for inflation to year 2000 dollars using published conversion factors) were also examined.6 A hospital is defined as a teaching hospital by the NIS if it has an American Medical Association approved residency program, is a member of the Council of Teaching Hospitals, or has a resident-to-bed ratio of 0.25 or higher. Because the definition of a teaching hospital by the NIS changed beginning in 1998, longitudinal analysis of teaching hospital status is not possible using the complete data set. Therefore, we used a subset of the data using a uniform definition of teaching hospitals. First, we identified all hospitals that were included in the NIS during 1 or more years between 1988 and 1997, and that had the same teaching status during every year in which the hospital was included. Next, we selected those patients (from the complete data set) who were discharged from one of these hospitals. The teaching hospital status of each discharge from the 1998 to 2000 period was defined using the 1988 to 1997 definition of that hospital. This approach decreased the number of pyeloplasty procedures in the analysis from 5,858 to 4,362. The analysis of this data set was not weighted due to the skewed sampling required to obtain the uniform teaching hospital definition. A variety of covariates were examined. Socioeconomic status was estimated by median income of patient home ZIP code, grouped as less than $25,000, $25,000 to $34,999, $35,000 to $44,999 and more than $45,000. Insurance status was categorized as insured (private/health maintenance organization [HMO]/Medicare), Medicaid/other (Medicaid, other coverage and no charge) or self-pay (no coverage). The level of comorbidity was determined by the total number of diagnoses at discharge from the hospital, including UPJ obstruction. Hospital data in the NIS include teaching status, urban vs rural status and region (West, South, Midwest and East). Hospital size is categorized as small, medium or large, as defined by the NIS. The definitions of these categories vary by region and rural or urban location. Due to the fact that several states did not provide race/ethnicity information in some years, for analytical purposes we assigned a discrete “race unavailable” value to all subjects for whom a race category was not provided. This approach allowed these subjects, who would otherwise be excluded from the multivariate models due to missing values, to be included, greatly increasing the power of the model. These individuals comprised 33.7% of the subjects in the sample. Surgical volume was determined by computing the number of pyeloplasty procedures done by each hospital in the sample for each given year of participation. These “hospital-volumeyear” statistics were then sorted in descending order and divided into quartiles according to the cumulative case volume.7 Pyeloplasties done at a hospital performing 16 or more procedures during that year were considered to have been done at a “high volume” center. Pyleoplasties done at a hospital in the lower 3 quartiles were considered to have been done at a “low volume” center. Bivariate and multivariate analyses were carried out using SAS statistical software (SAS Institute, Cary, North Carolina). To provide accurate national estimates of trends in outcomes, analyses were weighted. Comparison of means was achieved using t tests and ANOVA. Individual categorical variables were tested for associations using chi-square tests. The time (year) variable was analyzed as an “era” variable, with the study period divided into 3 eras—1988 to 1991, 1992 to 1996 and 1997 to 2000. Significance level of 0.05 was applied for hypothesis testing.
RESULTS
The NIS contained data on 5,858 hospital admissions of subjects younger than 18 years from 1988 through 2000 during which pyeloplasty was performed. The demographic characteristics of the sample are shown in table 1. Weighted mean patient age at hospitalization was 62.8 ⫾ 162.3 months. Of the patients 70.7% were male. This proportion did not vary by year (p ⫽ 0.5660), and varied only slightly by race (70.8% male among whites vs 72.7% male among nonwhites, p ⫽ 0.0003) and region (69.1% in the Northeast vs 71.8% male in the South, p ⫽ 0.0008). Most patients were generally healthy. Median number of diagnoses at discharge from the hospital was 2.0 (mean 2.2 ⫾ 3.8). Weighted inhospital mortality among the sample was low at 0.02%. Age at time of surgery. Weighted mean age at time of surgery decreased moderately during the study period from 64.0 months during 1988 to 1991 to 61.2 months during 1997 to 2000 (table 2). Males tended to undergo surgery almost a year earlier than females (60.4 vs 70.8 months, p ⬍0.0001). Similarly, nonwhite patients also tended to undergo surgery significantly earlier than white patients (44.7 vs 69.6 months, p ⬍0.0001). There were also dramatic differences in age at surgery between high and low volume centers, with high volume centers operating almost 30 months earlier on average (p ⬍0.0001). Surgery during infancy. We found that the weighted proportion of procedures performed in children younger than 6 months decreased during the study period from 34.2% (1988 to 1991) to 25.2% (1997 to 2000), even after adjusting for several patient and hospital factors associated with early
TABLE 1. Demographic characteristics Variables
Raw No. Pts (%)
Weighted %
Pt age: Less than 1 yr 2,422 (41.3) 42.0 1–6 Yrs 1,485 (25.3) 25.1 7–12 Yrs 1,110 (18.9) 18.8 13–18 Yrs 841 (14.4) 14.1 Gender: M 4,119 (70.3) 70.7 F 1,738 (29.7) 29.3 Geographic region: South 2,092 (35.7) 39.7 Midwest 1,316 (22.5) 22.4 Northeast 1,249 (21.3) 20.2 West 1,201 (20.5) 17.7 Socioeconomic status (median ZIP code income): $45,000⫹ 1,135 (21.6) 20.7 $35,000–$44,999 1,293 (24.7) 24.0 $25,000–$34,999 1,790 (34.1) 34.9 $1,000–$24,999 1,027 (19.6) 20.4 Insurance coverage: Private/HMO/Medicare 3,826 (65.9) 65.9 Medicaid/other 1,861 (32.1) 32.1 None (self-pay) 115 (2.0) 2.0 Race/ethnicity: White 2,808 (72.3) 72.5 Hispanic 483 (12.4) 12.2 Black 354 (9.1) 9.4 Other 153 (3.9) 3.9 Asian/Pacific 75 (1.9) 1.8 Native American 12 (0.3) 0.3 Hospital location: Rural 247 (4.2) 4.5 Urban nonteaching 2,060 (35.2) 32.3 Urban teaching 3,547 (60.5) 63.2 Hospital size: Large 3,533 (60.3) 58.8 Medium 1,395 (23.8) 21.7 Small 926 (15.8) 19.5 Data set consists of pediatric patients undergoing surgery for correction of ureteropelvic junction obstruction, including raw counts and percentages for study sample. Weighted percentages represent national population based estimates of proportion of pediatric patients undergoing surgery for UPJ obstruction in each category.
234
CONTEMPORARY TRENDS IN PEDIATRIC PYELOPLASTY
TABLE 2. Age at time of surgery for correction of UPJ obstruction Variables
Weighted Mean Age at Surgery (mos)
Weighted p Value*
Era: 1988–1991 64.0 0.0168 1992–1996 63.2 1997–2000 61.2 Gender: M 60.1 ⬍0.0001 F 69.4 Race/ethnicity: White 70.7 ⬍0.0001 Nonwhite 44.4 Not available 61.9 No. diagnoses at discharge: More than 2 65.7 0.0099 0–2 61.6 Insurance coverage: Private/HMO/Medicare 70.3 ⬍0.0001 Medicaid/other 46.1 Self-pay 78.1 Hospital case vol: High (top quartile) 40.6 ⬍0.0001 Low (bottom quartiles) 70.4 Hospital type and location: Rural 134.1 ⬍0.0001 Urban nonteaching 74.0 Urban teaching 52.1 Hospital size: Small 50.4 ⬍0.0001 Medium 73.4 Large 63.1 Patient and hospital factors associated with age at time of surgery for correction of ureteropelvic junction obstruction, weighted to national population (nonsignificant covariates excluded from multivariate model are socioeconomic status (median income of home ZIP code) and geographic region). * Adjusted for all other significant covariates (multivariate linear regression).
surgery (table 3). Mirroring the findings for overall age at surgery, nonwhite patients were significantly more likely to undergo surgery during the first 6 months (38.9% vs 25.0%, p ⬍0.0001). Regionalization of care. The proportion of procedures done at teaching hospitals increased significantly with time from 48.9% in 1988 to 1991 to 61.3% in 1997 to 2000 (table 4). This trend was present after adjusting for several other factors that were also associated with increased odds of a procedure being done at a teaching hospital, including nonwhite race, early surgery (before 6 months) and Northeast geographic region. Length of stay. Weighted mean length of hospital stay decreased from 6.6 days in 1988 to 1991 to 3.7 days in 1997 to 2000, and this change was significant (p ⬍0.0001) even when adjusting for other patient and hospital factors (see figure). As one would expect, patients with greater co-morbidity (more than 2 diagnoses) had longer LOS than healthier patients (7.6 vs 3.9 days, p ⬍0.0001). Longer hospital stays were also seen for nonwhite compared to white patients (6.1 vs 4.6 days, p ⬍0.0001) as well as for Medicaid patients compared to private/HMO patients (6.0 vs 4.5 days, p ⬍0.0001). Other factors associated with longer LOS included urban teaching hospital and lower socioeconomic status. Gender, hospital volume and geographic region were not significantly associated with LOS. Interestingly, weighted mean inflation adjusted total hospital charges did not change significantly during the study period (p ⫽ 0.8451) despite the dramatic decrease in length of stay (see figure). DISCUSSION
A key finding from these data is the dramatic decrease in the proportion of patients undergoing pyeloplasty as infants. In fact, the odds of undergoing surgery for UPJ obstruction during the first 6 months of life decreased by more than 50%
between 1988 and 2000. This finding suggests that, increasingly, surgical intervention for prenatally detected UPJ obstruction is being delayed in favor of observation. The debate over the optimal management of prenatal hydronephrosis has been ongoing since the introduction of practical prenatal sonographic screening in the 1980s.8 Early in the debate a number of authors advocated early intervention to preserve renal function.1 However, increasingly, observation has been recommended for most infants, as many seem to do well without aggressive surgical intervention, and many urologists prefer to avoid procedures in these young patients.2 This evolution in the literature is consistent with the trend that we observed in our analysis, as a gradual but definite shift away from pyeloplasty in newborns and infants. Another key finding of this study with respect to timing of surgery is that significant differences exist between white and nonwhite patients. Nonwhite patients underwent surgery more than 2 years earlier than whites, and nonwhite patients were much more likely to undergo surgery during the first 6 months of life. To our knowledge this is the first report of such a marked difference along racial or ethnic lines in the management of UPJ obstruction. Vesicoureteral reflux is reported to be rare among black children, particularly those with prenatal hydronephrosis, but there is little in the literature to support the existence of racial difference in the incidence or natural history of UPJ obstruction.9 It is important to note that the racial difference in timing of surgery persisted after adjustment for socioeconomic status and insurance status. This finding raises questions with respect to other factors that may be driving the differential timing of intervention in white and nonwhite children. Racial disparity in medical treatment has been demonstrated for a number of conditions, and our findings represent further evidence of such disparity.10, 11 The data reveal a significant shift of caseload away from rural and nonteaching hospitals to teaching institutions. The reasons for this trend are likely multifactorial. Changes in the financial reimbursement environment may have made it less cost-efficient for private practitioners to perform highly subspecialized procedures such as pediatric pyeloplasty. Another possible factor is the emergence of laparoscopic pyeloplasty. While the number of urologists who perform laparoscopic pyeloplasty is growing, they still represent a minority, which may mean that more patients migrate to teaching institutions in search of this or other minimally invasive techniques.12 The decrease in length of hospital stay after pyeloplasty reflects a wider health care trend. Decreasing length of stay has been documented for a variety of medical and surgical conditions, including those in pediatric urology.13 A wide range of factors have been reported to have an impact on length of inpatient stay, including use of care pathways and protocols, specialty consultation, level of use of imaging technology, legislation and evolution of surgical technique.13–17 Our study shows that race and socioeconomic status also are associated with LOS. The growing use of laparoscopic and other minimally invasive techniques may also have had an impact on length of hospital stay after pyeloplasty. Laparoscopic procedures are commonly promoted as resulting in shorter hospital stays, and this has been demonstrated to be the case with pyeloplasty.18 However, since the ICD-9 codes for correction of ureteropelvic junction obstruction do not differentiate between open and laparoscopic procedures, we were unable to examine the effect of laparoscopic techniques on length of stay. Increasing use of laparoscopy and other technology intensive methods may also be responsible for our finding that despite the substantial decrease in length of hospital stay, total inflation adjusted hospital charges essentially remained constant during the study period. This finding is particularly
CONTEMPORARY TRENDS IN PEDIATRIC PYELOPLASTY
235
TABLE 3. Surgery during newborn period and infancy Variables
Weighted % Pts Younger Than 6 Mos
Weighted OR (95% CI)*
Weighted p Value*
Era: 1988–1991 34.2 1.0 ⬍0.0001 1992–1996 29.8 0.78 (0.73–0.84) 1997–2000 25.2 0.48 (0.45–0.52) Gender: F 24.6 1.0 ⬍0.0001 M 32.0 1.48 (1.39–1.57) Race/ethnicity: White 25.0 1.0 ⬍0.0001 Nonwhite 38.9 1.57 (1.46–1.70) Not available 31.3 1.08 (1.00–1.16) No. diagnoses at discharge: 0–4 28.7 1.0 ⬍0.0001 More than 4 41.6 1.79 (1.63–1.96) Insurance coverage: Medicaid/other 36.6 1.0 ⬍0.0001 Private/HMO/Medicare 26.7 0.71 (0.67–0.76) Self-pay 22.9 0.59 (0.48–0.74) Socioeconomic status (median ZIP code income): $1,000–$24,999 30.5 1.0 ⬍0.0001 $25,000–$34,999 30.9 1.14 (1.06–1.23) $35,000–$44,999 26.9 1.02 (0.93–1.11) $45,000⫹ 30.7 1.36 (1.24–1.50) Geographic region: Midwest 24.4 1.0 0.0042 South 32.0 1.07 (0.99–1.16) Northeast 31.7 1.15 (1.05–1.26) West 29.8 1.16 (1.06–1.27) Hospital vol: Low 26.5 1.0 ⬍0.0001 High 39.7 1.84 (1.72–1.96) Hospital type and location: Rural 3.7 1.0 ⬍0.0001 Urban nonteaching 26.1 5.59 (4.21–7.42) Urban teaching 33.7 8.98 (6.78–11.9) Hospital size: Large 29.4 1.0 Small 36.2 0.85 (0.78–0.91) ⬍0.0001 Medium 24.9 0.70 (0.65–0.75) Patient and hospital factors associated with surgery for correction of ureteropelvic junction obstruction before age 6 months, weighted to national population. * Adjusted for all other significant covariates.
TABLE 4. Surgery at teaching hospitals Variables
Teaching Hospital %
OR (95% CI)*
p Value*
Era: 1988–1991 48.9 1.0 ⬍0.0001 1992–1996 57.4 1.24 (1.05–1.46) 1997–2000 61.3 1.65 (1.38–1.97) Race/ethnicity: White 55.3 1.0 ⬍0.0001 Nonwhite 66.6 1.81 (1.49–2.18) Not available 52.0 0.98 (0.85–1.14) Age at surgery: 6 Mos or older 51.9 1.0 ⬍0.0001 Younger than 6 mos 65.4 1.80 (1.56–2.08) Geographic region: West 35.5 1.0 ⬍0.0001 South 56.0 2.32 (1.94–2.77) Midwest 54.4 2.72 (2.24–3.31) Northeast 79.9 7.87 (6.35–9.74) Patient factors associated with surgery for correction of ureteropelvic junction obstruction being done at teaching hospital (nonsignificant covariates excluded from multivariate model are gender, comorbidity, insurance status and socioeconomic status). * Adjusted for all other significant covariates.
Trends through time in length of stay (LOS) and hospital charges for surgical correction of UPJ obstruction. Note decreasing LOS during study period, while inflation adjusted total hospital charges (Charges) remain statistically unchanged.
revealing, given the strong correlation of charges with length of stay (p ⬍0.0001). The fact that inflation adjusted charges failed to decrease in an environment of markedly decreasing length of stay suggests that other powerful cost pressures are driving up the overall cost of surgical care. These factors may include increased hospital overhead and personnel costs, increased equipment and pharmaceutical costs, and increasing malpractice insurance costs.19, 20
Although this is one of the first known studies to examine surgical correction of UPJ obstruction at a population level, several limitations need to be considered. Our definition of a patient who underwent pyeloplasty relies on the ICD-9 procedure coding system. It is possible that coding may have been incomplete or inaccurate. In addition, as noted previously, we are limited in our ability to differentiate the type of surgical procedure performed, specifically regarding the use
236
CONTEMPORARY TRENDS IN PEDIATRIC PYELOPLASTY
of minimally invasive techniques such as laparoscopy. We also are limited in our ability to evaluate complication rates, particularly delayed presentations, since the data can only provide diagnoses during the operative hospitalization itself. Length of hospital stay is a good proxy measure for initial surgical outcome, and we have used it as such, but this variable does not provide information about the subsequent clinical course. Finally, there was a relatively high proportion of subjects with missing race/ethnicity data. To accommodate for this problem, we elected to apply a separate categorical value for observations with missing race data. This approach allowed these subjects to be included in the final multivariate models, while acknowledging that they did not fit into a defined race category. CONCLUSIONS
Pediatric pyeloplasty has evolved in significant ways between 1988 and 2000, and these changes are reflected in national practice patterns. Although mean age at surgery decreased slightly during the study period, a significantly smaller proportion of procedures are being performed during the first 6 months of life. This finding suggests that patients with a diagnosis of prenatal hydronephrosis are increasingly being observed instead of undergoing surgery in the newborn period and infancy. There are marked racial differences in the timing of surgery, with white patients being less likely to undergo surgery during the first 6 months of life, and having a higher mean age at surgery. More cases are being done at teaching hospitals, suggesting that regionalization of care is occurring. Length of hospital stay decreased markedly during the study period, which is consistent with broader trends in medicine, although inflation adjusted hospital charges failed to decrease significantly as would be expected from shorter hospital stays. REFERENCES
1. Koyle, M. A. and Ehrlich, R. M.: Management of ureteropelvic junction obstruction in neonate. Urology, 31: 496, 1988 2. Onen, A., Jayanthi, V. R. and Koff, S. A.: Long-term followup of prenatally detected severe bilateral newborn hydronephrosis initially managed nonoperatively. J Urol, 168: 1118, 2002 3. Figenshau, R. S., Clayman, R. V., Colberg, J. W., Coplen, D. E., Soble, J. J. and Manley, C. B.: Pediatric endopyelotomy: the Washington University experience. J Urol, 156: 2025, 1996 4. Marcovich, R., Jacobson, A. I., Aldana, J. P., Lee, B. R. and Smith, A. D.: Practice trends in contemporary management of adult ureteropelvic junction obstruction. Urology, 62: 22, 2003
5. Finlayson, S. R., Laycock, W. S. and Birkmeyer, J. D.: National trends in utilization and outcomes of antireflux surgery. Surg Endosc, 17: 864, 2003 6. Sahr, R.: Inflation Conversion Factors for Dollars 1665 to Estimated 2013. Available at http://oregonstate.edu/dept/pol_sci/ fac/sahr/sahr.htm. Accessed October 26, 2003 7. Taub, D. A., Miller, D. C., Cowan, J. A., Dimick, J. B., Montie, J. E. and Wei, J. T.: Impact of surgical volume on mortality and length of stay after nephrectomy. Urology, 63: 862, 2004 8. Pocock, R. D., Witcombe, J. B., Andrews, H. S., Berry, P. J. and Frank, J. D.: The outcome of antenatally diagnosed urological abnormalities. Br J Urol, 57: 788, 1985 9. Horowitz, M., Gershbein, A. B. and Glassberg, K. I.: Vesicoureteral reflux in infants with prenatal hydronephrosis confirmed at birth: racial differences. J Urol, 161: 248, 1999 10. Underwood, W., DeMonner, S., Ubel, P., Fagerlin, A., Sanda, M. G. and Wei, J. T.: Racial/ethnic disparities in the treatment of localized/regional prostate cancer. J Urol, 171: 1504, 2004 11. Ferguson, J. A., Weinberger, M., Westmoreland, G. R., Mamlin, L. A., Segar, D. S., Greene, J. Y. et al: Racial disparity in cardiac decision making: results from patient focus groups. Arch Intern Med, 158: 1450, 1998 12. Gerber, G. S.: Trends in endourologic practice. J Endourol, 16: 347, 2002 13. McMullin, N., Khor, T. and King, P.: Internal ureteric stenting following pyeloplasty reduces length of hospital stay in children. Br J Urol, 72: 370, 1993 14. Dy, S. M., Garg, P. P., Nyberg, D., Dawson, P. B., Pronovost, P. J., Morlock, L. et al: Are critical pathways effective for reducing postoperative length of stay? Med Care, 41: 637, 2003 15. Wagner, A. K., Fabio, T., Zafonte, R. D., Goldberg, G., Marion, D. W. and Peitzman, A. B.: Physical medicine and rehabilitation consultation: relationships with acute functional outcome, length of stay, and discharge planning after traumatic brain injury. Am J Phys Med Rehabil, 82: 526, 2003 16. Fleszler, F., Friedenberg, F., Krevsky, B., Friedel, D. and Braitman, L. E.: Abdominal computed tomography prolongs length of stay and is frequently unnecessary in the evaluation of acute pancreatitis. Am J Med Sci, 325: 251, 2003 17. Madlon-Kay, D. J., DeFor, T. A. and Egerter, S.: Newborn length of stay, health care utilization, and the effect of Minnesota legislation. Arch Pediatr Adolesc Med, 157: 579, 2003 18. Baldwin, D. D., Dunbar, J. A., Wells, N. and McDougall, E. M.: Single-center comparison of laparoscopic pyeloplasty, Acucise endopyelotomy, and open pyeloplasty. J Endourol, 17: 155, 2003 19. Studdert, D. M., Mello, M. M. and Brennan, T. A.: Medical malpractice. N Engl J Med, 350: 283, 2004 20. Goetghebeur, M. M., Forrest, S. and Hay, J. W.: Understanding the underlying drivers of inpatient cost growth: a literature review. Am J Manag Care, 9: SP3, 2003