Objective
We examined the influence of physician and hospital volume on the morbidity and mortality of radical hysterectomy for cervical cancer.
Study Design
Women who underwent radical hysterectomy for cervical cancer between 2003 and 2007 were examined. The effect of surgeon and hospital volume on morbidity and mortality was examined using multivariable generalized estimating equations.
Results
A total of 1536 women who underwent radical hysterectomy were identified. Patients treated by high-volume surgeons had fewer medical complications (odds ratio, 0.55; 95% confidence interval, 0.34–0.88) and shorter lengths of stay (odds ratio, 0.49; 95% confidence interval, 0.25–0.98). After adjustment for case mix and surgeon volume, hospital volume had no independent effect on any of the variables of interest.
Conclusion
High-volume surgeons have fewer postoperative medical complications, shorter lengths of stay, and lower transfusion requirements. Hospital volume appears to have only a minor influence on outcomes after radical hysterectomy.
Radical hysterectomy is the surgical treatment of choice for stage IB-IIA cervical cancer. The procedure consists of the en bloc resection of the uterus and surrounding tissues, including the parametria, upper vagina, and uterosacral ligaments. Although radical hysterectomy is associated with excellent oncologic outcomes, the surgery is also accompanied by substantial morbidity. A number of patient and tumor factors influence the morbidity of radical hysterectomy.
The importance of surgical volume on outcome has now been demonstrated for a number of procedures; those patients operated on by high-volume surgeons and at high-volume hospitals have superior outcomes. In an analysis of 14 cardiovascular and cancer surgeries, mortality was reduced for all of the procedures when the operations were performed in high-volume hospitals. Surgeon volume appears to have a similar effect; in an analysis of over 450,000 patients surgeon volume mediated a large proportion of the variation in outcome. Although the volume-outcome relationship has been documented for a number of procedures, the effects of volume appear to be most pronounced for high-risk cardiovascular and oncologic procedures.
The effect of surgical volume on the morbidity and mortality of radical hysterectomy for cervical cancer has received little attention. Radical hysterectomy is a technically demanding operation, typically performed by gynecologic oncologists who have completed subspecialty training. Given the declining incidence of cervical cancer in developed countries, radical hysterectomy is now a procedure that is performed relatively infrequently, even among some subspecialists. In concert, this suggests that surgical volume may influence the outcomes of radical hysterectomy. We examined the influence of physician and hospital volume on the morbidity and mortality of radical hysterectomy for cervical cancer.
Materials and Methods
Data from the Perspective (Premier, Charlotte, NC) database was used. Perspective is a nationwide, voluntary, fee-supported database developed to measure quality and resource use. The database collects inpatient data from more than 500 acute-care hospitals throughout the United States. The data are updated quarterly and audited regularly to ensure quality and integrity. Perspective collects data on patient demographics, clinical characteristics, procedures, and all billed services. Perspective is validated and has been used in a number of outcomes studies.
Women who underwent abdominal radical hysterectomy ( International Classification of Disease, 9th revision code 68.6) for invasive cervical cancer between 2003 and 2007 were analyzed. Patients who underwent a minimally invasive procedure were excluded. Predictor variables in the analysis included age (<60 or >60 years of age), race, marital status, date of diagnosis (2003-2005 vs 2006-2007), and region of residence (midwest, northeast, south, west). Comorbidity was estimated using the Charlson Index. Hospital characteristics, including location (urban or rural), type (teaching or nonteaching), and size (<400, 400-600, <600 beds), were recorded.
The volume of each patient’s surgeon and hospital was calculated. As not all surgeons contributed patients during the entire study period, we calculated annual surgeon volume by dividing the total number of procedures by the number of years in which an individual surgeon performed at least 1 radical hysterectomy. A similar classification schema for annualized hospital volume was developed. Surgeon and hospital volume cutpoints were then selected to divide patients into approximately equal volume-based tertiles: low, intermediate, and high.
Procedure-associated morbidity was assessed using International Classification of Disease, 9th revision codes for known radical hysterectomy-related complications. Complications were classified into the following groups: (1) operative complications (bladder injury, ureteral injury, intestinal injury, vascular injury, other operative injury), (2) perioperative surgical complications (reoperation, postoperative hemorrhage, wound complication, venous thromboembolism), and (3) medical complications (cardiovascular, pulmonary, gastrointestinal, renal, infectious, neurologic). Length of hospital stay was defined as the number of days from the procedure until discharge. Rates of transfusion and intensive care unit (ICU) use were calculated. Readmissions within 60 days of the procedure for any of the complications defined above were recorded. Perioperative death was defined as death during the hospitalization during which the patient underwent the procedure.
Frequency distributions between categoric variables were compared using the χ 2 test, whereas continuous variables were compared with 1-way analysis of variance (ANOVA). The unadjusted rates of individual complications and complication classes were first compared across surgeon and hospital volume strata. Generalized estimating equations were used to analyze the outcomes of interest, while controlling for other demographic and clinical variables. Models were constructed with adjustment for case mix as well as with adjustment for case mix and surgeon volume (when the primary variable of interest was hospital volume) and case mix and hospital volume (when the primary variable of interest was surgeon volume). To account for clustering the generalized estimating equations were fitted to correct for within-surgeon correlation (when hospital volume was the primary predictor) and within-hospital correlation (when surgeon volume was the primary predictor). Odds ratios (ORs) for the primary outcomes of interest are reported comparing the low-volume and high-volume tertiles with 95% confidence intervals (CIs). All P values were 2-sided. A P value of less than .05 was considered statistically significant. All analyses were performed with SAS version 9.2 (SAS Institute, Cary, NC).
Results
A total of 1536 women with cervical cancer who underwent radical hysterectomy were identified. The demographic characteristics of the cohort are displayed in Table 1 . Patients operated on by high-volume surgeons were more often married, resided in the southern United States, were located in urban areas, were treated at nonteaching hospitals, and were treated at large hospitals ( P < .05 for all). Women treated at high-volume hospitals were more frequently white, more often resided in the southern United States, lived in urban areas, were treated at nonteaching hospitals, and were treated at large hospitals ( P < .05 for all).
| Surgeon volume | Hospital volume | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Characteristic | Low | Intermediate | High | P value | Low | Intermediate | High | P value | ||||||
| Patients | 512 | 478 | 546 | 514 | 503 | 519 | ||||||||
| Hospitals | 109 | 28 | 15 | |||||||||||
| Surgeons | 234 | 60 | 39 | |||||||||||
| Mean annual procedures | ≤2.25 | 2.26–3.75 | <3.75 | ≤4.4 | 4.5–7.0 | <7.0 | ||||||||
| Age | .76 | .39 | ||||||||||||
| <60 | 425 | (83.0) | 405 | (84.7) | 459 | (84.1) | 422 | (82.1) | 426 | (84.7) | 441 | (85.0) | ||
| ≥60 | 87 | (17.0) | 73 | (15.3) | 87 | (15.9) | 92 | (17.9) | 77 | (15.3) | 78 | (15.0) | ||
| Race | .34 | < .0001 | ||||||||||||
| White | 327 | (63.9) | 290 | (60.7) | 345 | (63.2) | 312 | (60.7) | 301 | (59.8) | 349 | (67.2) | ||
| Black | 51 | (10.0) | 62 | (13.0) | 73 | (13.4) | 44 | (8.6) | 74 | (14.7) | 68 | (13.1) | ||
| Other | 134 | (26.2) | 126 | (26.4) | 128 | (23.4) | 158 | (30.7) | 182 | (25.5) | 102 | (19.7) | ||
| Insurance | .12 | .13 | ||||||||||||
| Medicare | 53 | (10.4) | 55 | (11.5) | 62 | (11.4) | 59 | (11.5) | 55 | (10.9) | 56 | (10.8) | ||
| Medicaid | 65 | (12.7) | 69 | (14.4) | 77 | (14.1) | 68 | (13.2) | 85 | (16.9) | 58 | (11.2) | ||
| Commercial | 335 | (65.4) | 308 | (64.4) | 368 | (67.4) | 333 | (64.8) | 318 | (63.2) | 360 | (69.4) | ||
| Uninsured | 42 | (8.2) | 23 | (4.8) | 27 | (5.0) | 30 | (5.8) | 29 | (5.8) | 33 | (6.4) | ||
| Unknown | 17 | (3.3) | 23 | (4.8) | 12 | (2.2) | 24 | (4.7) | 16 | (3.2) | 12 | (2.3) | ||
| Diagnosis | .70 | .23 | ||||||||||||
| 2003-2005 | 400 | (78.1) | 377 | (78.9) | 419 | (76.7) | 387 | (75.3) | 398 | (79.1) | 411 | (79.2) | ||
| 2006-2007 | 112 | (21.9) | 101 | (21.1) | 127 | (23.3) | 127 | (24.7) | 105 | (20.9) | 108 | (20.8) | ||
| Marital status | .002 | .17 | ||||||||||||
| Married | 251 | (49.0) | 224 | (46.9) | 278 | (50.9) | 251 | (48.8) | 240 | (47.7) | 262 | (50.5) | ||
| Single | 227 | (44.3) | 186 | (38.9) | 213 | (39.0) | 216 | (42.0) | 217 | (43.1) | 193 | (37.2) | ||
| Unknown | 34 | (6.6) | 68 | (14.2) | 55 | (10.1) | 47 | (9.1) | 46 | (9.2) | 64 | (12.3) | ||
| Region | < .0001 | < .0001 | ||||||||||||
| Midwest | 110 | (21.5) | 130 | (27.2) | 8 | (1.5) | 139 | (27.0) | 98 | (19.5) | 11 | (2.1) | ||
| Northeast | 71 | (13.9) | 54 | (11.3) | 49 | (9.0) | 68 | (13.2) | 85 | (16.9) | 21 | (4.1) | ||
| South | 248 | (48.4) | 260 | (54.4) | 468 | (85.7) | 225 | (43.8) | 295 | (58.7) | 456 | (87.9) | ||
| West | 83 | (16.2) | 34 | (7.1) | 21 | (3.9) | 82 | (16.0) | 25 | (5.0) | 31 | (6.0) | ||
| Location | .002 | < .0001 | ||||||||||||
| Urban | 499 | (97.5) | 469 | (98.1) | 546 | (100) | 492 | (95.7) | 503 | (100) | 519 | (100) | ||
| Rural | 13 | (2.5) | 9 | (1.9) | 0 | — | 22 | (4.2) | 0 | — | 0 | — | ||
| Teaching hospital | < .0001 | < .0001 | ||||||||||||
| No | 206 | (40.2) | 132 | (27.6) | 244 | (44.7) | 243 | (47.3) | 128 | (25.5) | 211 | (40.7) | ||
| Yes | 306 | (59.8) | 346 | (72.4) | 302 | (55.3) | 271 | (52.7) | 375 | (74.6) | 308 | (59.3) | ||
| Beds | < .0001 | < .0001 | ||||||||||||
| <400 | 179 | (35.0) | 132 | (27.6) | 83 | (15.2) | 229 | (44.6) | 103 | (20.5) | 62 | (12.0) | ||
| 400-600 | 163 | (31.8) | 175 | (36.6) | 149 | (27.3) | 180 | (35.0) | 236 | (46.9) | 71 | (13.7) | ||
| >600 | 170 | (33.2) | 171 | (35.8) | 314 | (57.5) | 105 | (20.4) | 164 | (32.6) | 386 | (74.4) | ||
| Charlson Index | .30 | .04 | ||||||||||||
| 1 | 273 | (53.3) | 284 | (59.4) | 311 | (57.0) | 291 | (56.6) | 271 | (53.9) | 306 | (59.0) | ||
| 2 | 147 | (28.7) | 128 | (26.8) | 152 | (27.8) | 133 | (25.9) | 142 | (28.2) | 152 | (29.3) | ||
| ≥3 | 92 | (18.0) | 66 | (13.8) | 83 | (15.2) | 90 | (17.5) | 90 | (17.9) | 61 | (11.8) | ||
| Oophorectomy | .02 | .59 | ||||||||||||
| Yes | 356 | (69.5) | 314 | (65.7) | 402 | (73.6) | 354 | (68.9) | 347 | (69.0) | 371 | (71.5) | ||
| No | 156 | (30.5) | 164 | (34.3) | 144 | (26.4) | 160 | (31.1) | 156 | (31.0) | 148 | (28.5) | ||
| Lymph node dissection | .06 | .04 | ||||||||||||
| Yes | 487 | (95.1) | 455 | (95.2) | 503 | (92.1) | 479 | (93.2) | 484 | (96.2) | 482 | (92.9) | ||
| No | 25 | (4.9) | 23 | (4.8) | 43 | (7.9) | 35 | (6.8) | 19 | (3.8) | 37 | (7.1) | ||
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
