Boyd, Leslie R. MD; Novetsky, Akiva P. MD; Curtin, John P. MD, MBA
Hysterectomy, or removal of the uterus, is one of the most common surgical procedures performed in the United States. In 2004 an estimated 610,000 U.S. women had a hysterectomy, with more than 90% of these performed for benign disease.1,2 The options for treating benign uterine conditions have broadened in the past decade. Several procedures have been introduced that allow for symptomatic relief of benign conditions without requiring a hysterectomy. These techniques, including uterine artery embolization and endometrial ablation, are gaining in popularity.3,4 Recent authors have noted a decline5 or plateau2 in the hysterectomy rate since the late 1990s. The decrease in the total number of hysterectomies, combined with an expansion of the available surgical approaches (abdominal compared with laparoscopic-assisted compared with vaginal), may translate into a reduction in the number of hysterectomies performed by any one practitioner.
There is growing literature on improved outcomes when the procedures are performed in high-volume hospitals6–8 or by surgeons. This concept was first investigated with complex cardiovascular and cancer surgeries.8–13 However, there are little data on whether women are clustering to high-volume hospitals or gynecologists for hysterectomies. In fact, there are scant data on how many hysterectomies should be performed for one to qualify as a high-volume surgeon.
A large component of the short-term morbidity associated with hysterectomy is related to the manner in which the procedure is performed. By avoiding an abdominal incision, there are multiple short-term benefits, including decreased postoperative pain and decreased length of stay.14 Vaginal hysterectomy is associated with a clear cost-savings when compared with either abdominal or laparoscopic approaches15; however, the latter may be cost-effective overall when length of stay and reduced convalescence are factored in.16 Despite these benefits, reports on the hysterectomy trends in this decade show a continued predominance of the abdominal technique.1,2 Our goal was to estimate trends in performance of hysterectomy in New York State, focusing on surgical volume and its effects on route of hysterectomy and short-term morbidity.
MATERIALS AND METHODS
The study was submitted to the New York University School of Medicine Institutional Review Board and was granted exempt status. This was a cross-sectional analysis using the New York State Department of Health Statewide Planning and Research Cooperative System (SPARCS) to identify cases. SPARCS is a comprehensive data reporting system, established in 1979, which collects patient and treatment information for every hospital discharge, ambulatory surgery, and emergency department admission. New York State mandates reporting by every hospital.
In the SPARCS database, each hospital discharge contains demographic data, including age, gender, race and ethnicity, type of insurance (private, Medicare, Medicaid, self-pay, uninsured), zip code of residence, reason for admission by diagnosis-related group codes, primary and secondary diagnoses (up to 24), primary and secondary procedures (up to 20), dates of procedure, and length of stay. The version of the database obtained by the researchers was de-identified in regard to the individual, thereby omitting the individual's name, date of birth, and medical record number.
Cases were identified in the years 2001 to 2006 using International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) procedure codes 68.3 to 68.7 (hysterectomy). Each admission that resulted in hysterectomy was included in the study. Hysterectomies performed for gynecologic malignancies (ICD-9-CM diagnosis codes 179–185) and obstetric indications (ICD-9-CM diagnosis codes 640–679) were excluded.
An acute morbidity index was created using the SPARCS database to search for ICD-9-CM codes for acute myocardial infarction (410.x and 411.x), acute pulmonary embolus (415.1 and 415.2), acute pneumonia (480.x), acute deep venous thrombosis (453.x), urinary tract infection (559.0), fever (780.6), pulmonary edema (276.6, 428.1, 518.4), acute renal failure (593.9), ureteral injury (568.x and 593.3), bladder injury (578.x and 596.6), ileus (560.1), small bowel obstruction (560.9), bowel injury (998.2), and blood transfusion (990.x). Acute mortality is coded independently within the database.
In addition, a modified Charlson index was used to determine preoperative morbidity as previously described.17 In brief, the comorbid conditions were used to create a score from 0 to 5, with 5 indicating the most severe conditions.
The SPARCS database provides license numbers for the surgeons associated with each surgical procedure. The database was used to identify the total number of hysterectomies performed by each surgeon over the course of the study period. The surgeons' license numbers were used as unique identifiers only; there was no attempt to link this number to obtain surgeons' names, institutions, or subspecialty training. The number of hysterectomies performed by each surgeon was divided by the number of years in which that surgeon performed at least one hysterectomy to generate an annual average. Hysterectomies performed for all indications were included to calculate each surgeon's case average. This average was used to stratify surgeons into the following categories: fewer than 10 or 10 or more hysterectomies per year; fewer than 15 or 15 or more hysterectomies per year; and fewer than 20 or 20 or more hysterectomies per year.
Statistical analyses were performed using Pearson χ2 and Student t test. P for trend was calculated using Cuzick nonparametric test for trend across ordered groups.
Number needed to treat, with 95% confidence intervals, was calculated for acute morbidities and mortality as the reciprocal of the difference in outcomes between the high-volume and low-volume surgeons.18 Linear and logistic regression models assessing the effect of surgeons' volume of cases on length of stay and acute morbidity after controlling for known confounders were created. To account for each surgeon contributing multiple patients to the cohort, a model to account for clustering was created using each surgeon's license number. All statistical analyses were performed using STATA 9 software.
Hysterectomies performed for benign indications during the study period are enumerated in Table 1. From 2001 through 2006, a total of 146,494 hysterectomies were performed in New York State. Of these, 103,181 (70.4%) were abdominal, 26,660 (18.2%) were vaginal, and 16,653 (11.4%) were laparoscopically assisted. The total number of hysterectomies performed annually declined over the study period from 25,249 in 2001 to 21,409 in 2006. There was a decrease in the percentage of abdominal hysterectomies (from 71.6% to 68.1%) and vaginal hysterectomies (from 19.1% to 17.1%) performed, with an increase in the number of laparoscopically assisted hysterectomies (from 9.3% to 14.8%; P for trend <.001).
Using the license numbers provided by the database, the number of hysterectomies performed by each surgeon in New York State was calculated. Descriptive statistics of each surgeon's average are given in Table 2. The average and median number of hysterectomies performed per surgeon per year in New York State are 6.8 and 3.1 cases, respectively. The range is broad, with one surgeon performing an average of 106 hysterectomies per year. However, the range between the 25th and 75th percentile is quite narrow at 1 and 8.9 cases per year, respectively. In Figure 1, providers are categorized in three ways: as performing fewer than 10 compared with 10 or more hysterectomies per year; as performing fewer than 15 compared with 15 or more hysterectomies per year; and as performing fewer than 20 compared with 20 or more hysterectomies per year. Overall, 26% of hysterectomies are performed by surgeons who perform fewer than 10 hysterectomies per year. Thirty-seven percent of hysterectomies in New York City are performed by surgeons who perform fewer than 10 hysterectomies per year, whereas this percentage decreases to 19% for providers in regions outside of New York City (P<.001). Surgeons outside of New York City were more likely to perform 15 or more (64%) and 20 or more (47%) hysterectomies per year when compared with surgeons in New York City (48% and 38%, respectively; P<.001 for both).
Figure 2 shows the type of hysterectomy performed stratified by the average number of hysterectomies performed per year for each surgeon listed in the database. The percentage of abdominal hysterectomy is 81% for surgeons performing fewer than 10 hysterectomies per year compared with 67% for surgeons performing more than 10 hysterectomies per year (P<.001). At each break point (10, 15, or 20 cases), vaginal and laparoscopically assisted hysterectomies are more likely as the surgeon's volume increases. For surgeons performing fewer than 10 hysterectomies per year, vaginal (12%) and laparoscopically assisted hysterectomies (7%) are less common compared with surgeons performing at least 10 hysterectomies (20% and 13%, respectively; P<.001). For surgeons performing fewer than 20 hysterectomies per year, vaginal (15%) and laparoscopically assisted hysterectomies (9%) are less common compared with surgeons performing at least 20 hysterectomies (22% and 15%, respectively; P<.001).
An assessment of postoperative morbidity was performed, including known short-term complications associated with hysterectomy, and this was stratified by surgeon's case volume. The results are shown in Table 3. The percentage of hysterectomies with postoperative morbidity or mortality for hysterectomies performed by surgeons with an average of fewer than 10 cases as compared with 10 or more cases was 16.5% compared with 11.7%, respectively (P<.001). All aspects of acute morbidity, with the exception of bladder injury and pulmonary edema, were significantly reduced when comparing hysterectomies performed by surgeons with fewer than 10 cases per year to those with 10 or more cases per year. When comparing hysterectomies performed by surgeons performing fewer than 10 hysterectomies per year compared with those performing at least 10 per year, the performance of 20.8 hysterectomies by surgeons performing at least 10 hysterectomies would be sufficient to avoid one case of acute morbidity. The comparable statistic for surgeons performing at least 15 and at least 20 hysterectomies per year was 33.4 and 46.7, respectively. Deaths were reported in 0.21% of women undergoing hysterectomy by surgeons with fewer than 10 cases per year compared with 0.06% for surgeons with at least 10 cases per year (P<.001).
A multivariable logistic regression was performed to estimate the effect of surgeon volume on hysterectomy-associated morbidity, controlling for mode of hysterectomy, comorbidity index, insurance status, and age. A cluster analysis model was included, and Table 4 shows the results of this analysis. The presence of preoperative comorbidities was the strongest predictor of increased morbidity (odds ratio 3.25). However, lower surgical volume (fewer than 10 hysterectomies per year) was associated with a 36% increase in the incidence of acute morbidity after controlling for other confounders. Morbidity was increased for both abdominal and laparoscopically assisted hysterectomies when compared with vaginal hysterectomies. Medicaid insurance status and age 50 or older were also associated with increased morbidity.
Figure 3 describes the length of hospital stay in days, stratified by type of hysterectomy performed and volume of the operating surgeon. Both the type of hysterectomy (abdominal compared with vaginal compared with laparoscopically assisted) and the surgical volume were correlated with length of stay. Abdominal hysterectomy was associated with a longer hospital stay when compared with vaginal or laparoscopically assisted hysterectomy in all categories. Length of stay for abdominal hysterectomy was 4.09 for surgeons averaging fewer than 10 hysterectomies per year as compared with 3.43 for surgeons averaging 10 or more hysterectomies per year (P<.001). Declines in the length of stay were also noted at the break points of 15 and 20 hysterectomies per year. A similar trend was noted for women in the vaginal and laparoscopically assisted hysterectomy categories. The longest length of stay was associated with abdominal hysterectomies performed by surgeons averaging fewer than 10 cases per year (4.09 days); the shortest length of stay was associated with laparoscopically assisted hysterectomies performed by surgeons averaging 20 or more cases per year (1.86 days).
A linear regression model was created to determine the effect of length of stay by surgeon's case volume while controlling for hysterectomy type, preoperative comorbid conditions, and postoperative morbidity (Table 5). A significant decrease in length of stay was noted when comparing surgeons who performed fewer than 10 compared with 10 or more hysterectomies per year, with the latter group posting a reduction of 0.44 days. Both preoperative and postoperative morbidity rates were associated with increased length of stay, adding 3.86 and 3.06 days, respectively. Abdominal hysterectomy was associated with a modest, but statistically significant, increase in length of stay when controlling for these other factors. Laparoscopically assisted hysterectomy was associated with a 0.36-day reduction in length of stay when compared with vaginal hysterectomy.
The availability of surgeons' license numbers allowed us to evaluate the effect of surgical volume on the type of hysterectomy performed. Surgeons who perform more than 10 hysterectomies per year are more likely to use vaginal and laparoscopic-assisted routes for the procedure. These less invasive routes were associated with a decrease in length of stay, and this decrease was more pronounced with increasing surgical volume.
The association of surgical volume with improved outcomes persisted in our multivariable logistic regression analysis. There have been many reports regarding the benefit of surgical volume for complex procedures.8,9,12,13 Although hysterectomy is a major procedure, it is not considered as being highly complex, as evidenced by the assigned relative value units. One previous report has linked surgical volume to outcomes for benign hysterectomy cases. Hanstede et al19 performed a single-institution retrospective review of hysterectomies and found that the performance of more than 10 hysterectomies per year led to an improvement in outcomes. However, their study included oncology-related hysterectomies and specifically excluded laparoscopically assisted hysterectomies, so the results are not wholly applicable to our data set.
We chose cutoffs of 10, 15, and 20 cases per year to evaluate the effects of volume on procedure type and outcome. Approximately 25% of hysterectomies in New York State are performed by physicians who perform fewer than 10 hysterectomies per year, making 10 cases per year the first quartile for our data. There is literature supporting the use of 10 cases per year as a benchmark in complex gynecologic procedures, including ovarian cancer staging,10,20 as well as in routine hysterectomy.19 Using this cutoff, we were able to see a significant decrease in acute morbidity for those surgeons with more than 10 cases per year. In comparison, there was diminished incremental benefit to performing more than 15 (or more than 20) hysterectomies per year. This implies that the appropriate definition of a high-volume surgeon for hysterectomy is 10 cases per year. We did not evaluate the hospitals in which the surgeries were performed; however, this has been found to be a significant variable.6,21–23
We also report a significant difference in mortality rates when cases are stratified by a surgical volume of fewer than 10 compared with 10 or more hysterectomies per year. The accepted mortality rate for benign hysterectomy is low, at 6.0 per 10,000 cases.24 This was the rate we found among surgeons performing 10 or more hysterectomies per year. We found a surprising increase in mortality for low-volume surgeons, at 0.21%, despite excluding both obstetric and oncology-related cases. In contrast, Hanstede et al19 found a stable, low mortality rate for both high- and low-volume surgeons in their cohort. Juillard et al23 did not find a difference in hysterectomy-associated mortality among high- and low-volume hospitals. Thus, our finding, although intriguing, bears repeating in other large data sets.
Population density in New York State ranges broadly, including populous urban areas and rural farmlands. In New York City, there is a greater likelihood of having a low-volume surgeon perform a hysterectomy when compared with other regions of the state. Given the large population of subspecialists in gynecologic surgery found in the city, this result came as a surprise. However, several authors argue that after meeting a minimum number of physicians per capita, additional physicians yield minimal benefits while potentially increasing costs.25,26 We presume that the relative decrease in the number of physicians outside of the main urban area increases the likelihood of being referred to a high-volume surgeon. This is an interesting finding that may also be relevant to physicians' practices in other large metropolitan areas.
The use of a mandatory state-wide database enabled us to evaluate a large pool of individuals in diverse hospital systems. However, there are several important limitations. Although compliance with the database is mandatory for all acute care hospitals in New York State, the accuracy of the reported data are not validated. Specifically, surgical procedures require accurate coding to be captured for this project. Also, because this is an inpatient database, individuals who were discharged the same day as their minimally invasive hysterectomy were not counted. This may skew the data and lead to under-reporting of minimally invasive procedures. However, it has been previously reported that 0.1% of hysterectomies performed in New York State were in an ambulatory setting.27 This percentage may be increasing over time and may be informative for future reporting. Despite these limitations, the current study is able to provide previously unknown information regarding patterns of care and short-term outcomes for women undergoing hysterectomy in New York State.
In summary, performance of an average of 10 hysterectomies per year is associated with improved outcomes and may be a reasonable benchmark for high-volume status. High-volume surgeons are more likely to perform minimally invasive hysterectomies, with attendant decreases in morbidity and length of stay. High-volume surgeons also have a decrease in short-term morbidity independent of the type of hysterectomy performed. Surgeons in less populous areas (with fewer competing physicians) were more likely to be high-volume surgeons.
1. Whiteman MK, Hillis SD, Jamieson DJ, Morrow B, Podgornik MN, Brett KM, et al. Inpatient hysterectomy surveillance in the United States, 2000–2004. Am J Obstet Gynecol 2008;198:34.e1–7.
2. Wu JM, Wechter ME, Geller EJ, Nguyen TV, Visco AG. Hysterectomy rates in the United States, 2003. Obstet Gynecol 2007;110:1091–5.
3. Farquhar CM, Naoom S, Steiner CA. The impact of endometrial ablation on hysterectomy rates in women with benign uterine conditions in the United States. Int J Technol Assess Health Care 2002;18:625–34.
4. Jacobson GF, Shaber RE, Armstrong MA, Hung YY. Changes in rates of hysterectomy and uterine conserving procedures for treatment of uterine leiomyoma. Am J Obstet Gynecol 2007;196:601.e1–6.
5. Merrill RM. Hysterectomy surveillance in the United States, 1997 through 2005. Med Sci Monit 2008;14:24–31.
6. Weeks WB, West AN, Wallace AE, Lee RE, Goodman DC, Dimick JB, et al. Reducing avoidable deaths among veterans: directing private-sector surgical care to high-performance hospitals. Am J Publ Health 2007;97:2186–92.
7. Birkmeyer JD, Finlayson SR, Tosteson AN, Sharp SM, Warshaw AL, Fisher ES. Effect of hospital volume on in-hospital mortality with pancreaticoduodenectomy. Surgery 1999;125:250–6.
8. Showstack JA, Rosenfeld KE, Garnick DW, Luft HS, Schaffarzick RW, Fowles J. Association of volume with outcome of coronary artery bypass graft surgery. Scheduled vs nonscheduled operations. JAMA 1987;257:785–9.
9. Birkmeyer JD, Stukel TA, Siewers AE, Goodney PP, Wennberg DE, Lucas FL. Surgeon volume and operative mortality in the United States. N Engl J Med 2003;349:2117–27.
10. Bristow RE, Zahurak ML, Diaz-Montes TP, Giuntoli RL, Armstrong DK. Impact of surgeon and hospital ovarian cancer surgical case volume on in-hospital mortality and related short-term outcomes. Gynecol Oncol 2009;115:334–8.
11. Hannan EL, O'Donnell JF, Kilburn H Jr, Bernard HR, Yazici A. Investigation of the relationship between volume and mortality for surgical procedures performed in New York State hospitals. JAMA 1989;262:503–10.
12. Hannan EL, Radzyner M, Rubin D, Dougherty J, Brennan MF. The influence of hospital and surgeon volume on in-hospital mortality for colectomy, gastrectomy, and lung lobectomy in patients with cancer. Surgery 2002;131:6–15.
13. Laffel GL, Barnett AI, Finkelstein S, Kaye MP. The relation between experience and outcome in heart transplantation. N Engl J Med 1992;327:1220–5.
14. Nieboer TE, Johnson N, Lethaby A, Tavender E, Curr E, Garry R, et al. Surgical approach to hysterectomy for benign gynaecological disease. The Cochrane Database of Systematic Reviews 2009, Issue 3. Art. No.: CD003677. DOI: 10.1002/14651858.CD003677.pub4.
15. Sculpher M, Manca A, Abbott J, Fountain J, Mason S, Garry R. Cost effectiveness analysis of laparoscopic hysterectomy compared with standard hysterectomy: results from a randomised trial. BMJ 2004;328:134–40.
16. Ellstrom M, Ferraz-Nunes J, Hahlin M, Olsson JH. A randomized trial with a cost-consequence analysis after laparoscopic and abdominal hysterectomy. Obstet Gynecol 1998;91:30–4.
17. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol 1992;45:613–9.
18. Laupacis A, Sackett DL, Roberts RS. An assessment of clinically useful measures of the consequences of treatment. N Engl J Med 1988;318:1728–33.
19. Hanstede MM, Wise LA, Stewart EA, Feldman S. The relation of annual surgeon case volume to clinical outcomes and resource utilization in abdominal hysterectomy. J Reprod Med 2009;54:193–202.
20. Elit L, Bondy SJ, Paszat L, Przybysz R, Levine M. Outcomes in surgery for ovarian cancer. Gynecol Oncol 2002;87:260–7.
21. Ioka A, Tsukuma H, Ajiki W, Oshima A. Influence of hospital procedure volume on ovarian cancer survival in Japan, a country with low incidence of ovarian cancer. Cancer Sci 2004;95:233–7.
22. Kumpulainen S, Grenman S, Kyyronen P, Pukkala E, Sankila R. Evidence of benefit from centralised treatment of ovarian cancer: a nationwide population-based survival analysis in Finland. Int J Cancer 2002;102:541–4.
23. Juillard C, Lashoher A, Sewell CA, Uddin S, Griffith JG, Chang DC. A national analysis of the relationship between hospital volume, academic center status, and surgical outcomes for abdominal hysterectomy done for leiomyoma. J Am Coll Surg 2009;208:599–606.
24. Wingo PA, Huezo CM, Rubin GL, Ory HW, Peterson HB. The mortality risk associated with hysterectomy. Am J Obstet Gynecol 1985;152:803–8.
25. Goodman DC, Grumbach K. Does having more physicians lead to better health system performance? JAMA 2008;299:335–7.
26. Fisher ES, Wennberg DE, Stukel TA, Gottlieb DJ, Lucas FL, Pinder EL. The implications of regional variations in Medicare spending. Part 2: health outcomes and satisfaction with care. Ann Intern Med 2003;138:288–98.
27. Farquhar CM, Steiner CA. Hysterectomy rates in the United States 1990–1997. Obstet Gynecol 2002;99:229–34.
© 2010 by The American College of Obstetricians and Gynecologists.