The Impact of Annual Surgical Volume on Patient Outcomes in Laparoscopic Hysterectomy

Introduction

More than 450,000 hysterectomies are performed in the United States each year, making it the most common non-pregnancy–related surgical procedure for women.^1–3 Hysterectomy can be performed by laparotomy or with minimally invasive techniques—via the vagina or with conventional or robotic laparoscopy. Minimally invasive hysterectomies carry significantly fewer risks than those done by laparotomy but require more surgical experience and technical training to accomplish safely. ⁴ That need for increased skill unfortunately means most hysterectomies are still performed via laparotomy, despite recommendations that a minimally invasive approach be applied when feasible. In the United States, 30.5% of hysterectomies are performed laparoscopically.^2,5–7

Major complications due to hysterectomy are rare; however, because hysterectomy is such a common procedure, efforts to decrease associated risk can have significant impacts on public health.^8,9 Associations between surgical volume and outcomes have been demonstrated in many other surgical fields.^10–13 In gynecology, that relationship has been investigated with abdominal myomectomy, hysteroscopic myomectomy, cancer surgery, pubovaginal sling placement, and pelvic reconstruction.^11,14–16 However, data regarding laparoscopic hysterectomy (LH) and the effects of surgeon volume on outcome are limited and inconsistent.^10,11 Some studies compared one type of LH procedure to another (total LH [TLH] versus laparoscopic assisted vaginal hysterectomy [LAVH] or laparoscopic assisted supracervical hysterectomy [LASH]).^17,18 Even fewer data exist regarding the impact of surgeon volume on resource utilization for LH.^17,19–23

This study examined the association of annual surgeon volume on surgical complications and resource utilization during LH in a large, tertiary-care teaching hospital.

Materials and Methods

Under a Cedars–Sinai institutional review board–approved protocol, a retrospective review was conducted on the charts of all patients undergoing LH for any indication between July 2011 and March 2014 (686 charts). Cedars–Sinai is a large, tertiary-care, academic not-for-profit hospital with providers including a mix of faculty and private physicians with marked variability in the mix of obstetric, gynecologic, surgical, and outpatient care. Approximately 60% of the hysterectomies during the study period were performed laparoscopically and 12% were performed vaginally; the route was determined at the discretion of the primary surgeon, who was either a gynecologist or a gynecologic oncologist. Both benign and malignant indications were included, but radical and robotic-assisted hysterectomies were excluded because of their increased complexity and because they are only performed by gynecologic oncologists at this institution. After further excluding the charts of patients without complete follow-up data, 539 patients' charts remained available for analysis.

Data regarding surgeon and patient characteristics were abstracted directly from electronic medical and hospital records (Table 1). Based on previous studies, a low-volume surgeon was defined as a surgeon who performed fewer than 5 LHs per year, an intermediate-volume surgeon did between 5 and 10 LHs per year, and a high-volume surgeon did >10 LHs per year. ²⁰ Volume was recorded according to the primary surgeon, although resident involvement occurred in all surgeries.

Intraoperative complications included visceral or vascular injury. Postoperative complications within 30 days of surgery were categorized using the validated Clavien–Dindo classification system, with grades I and II designated as minor complications and grades III and IV designated as major complications.^24,25 Resource utilization included: estimated blood loss (EBL); length of surgery; conversion to laparotomy; length of stay (LOS) in the hospital; and 30-day readmission. Length of surgery was defined in minutes from the initial incision to the completion of skin closure. LOS was defined as number of days in 24-hour increments, from the procedure until discharge. High resource utilization included conversion to laparotomy, readmission, EBL >100 mL, length of surgery >164 minutes, and LOS beyond 1 day postoperation. These specific cutoff points were chosen because they exceeded the 75th percentile for each variable in the study population.

TLH included patients in whom the entire uterine cervix and body were detached from the surrounding supporting structures laparoscopically and the vaginal cuff sutured either laparoscopically or vaginally. LAVH was defined as detaching the uterine body from its supporting structures laparoscopically and detaching the cervix from its attachments vaginally. LASH was defined as laparoscopically amputating the uterine corpus at the level of the uterine arteries and leaving the cervix in situ. ⁷

Univariate analyses, using distribution-appropriate tests, were performed to examine differences in surgeons' ages and the number of years the surgeons were in practice, patients' clinical factors and surgical procedures, surgical outcomes, and resource utilization among the different surgical-volume cohorts. χ², or Fisher's exact tests, were used for categorical variables and Student's t-tests or Wilcoxon signed-rank tests were used for continuous variables, as appropriate. Adjusted logistic regression models were used to assess the significance of surgeon volume on intraoperative and postoperative surgical complications, as well as resource utilization, adjusted for patient age, body mass index (BMI), uterine weight, and previous abdominal surgery. All comparisons were considered to have reached statistical significance at a p-value of ≤0.05. SAS, version 9.3 (Cary, NC) was used for all statistical testing.

Results

Forty-six surgeons performed 539 laparoscopic hysterectomies during the study period. Four surgeons were categorized as high-volume, 7 as intermediate-volume, and 35 (76%) as low-volume. Three of the 4 highest-volume surgeons were gynecologic oncologists; the remaining 1 had fellowship training in minimally invasive gynecologic surgery. Two intermediate-volume surgeons had fellowship training in gynecologic oncology, and 1 low-volume surgeon had fellowship training in minimally invasive surgery. There was no difference in the median age among low-, intermediate- and high-volume surgeons (46 versus 55 versus 49 years old; p = 0.4); and no difference in the years in practice among the surgeon groups (17 versus 20 versus 18; p = 0.5). One hundred and thirty-seven patients (26%) had surgery performed by a low-volume surgeon, 115 (21%) by an intermediate-volume surgeon, and 287 (53%) by a high-volume surgeon.

Women whose surgeries were performed by high-volume surgeons (> 10 LHs per year) tended to be older and to have a lower BMI (Table 1). There were no differences in the frequency of comorbidities or histories of prior abdominal surgery among women whose surgeries were performed by high-, intermediate-, or low-volume surgeons. Fibroids were the indication for hysterectomy in one-third of the patients. Low- and intermediate-volume surgeons performed hysterectomy for fibroids more commonly, with corresponding higher uterine weights, than high-volume surgeons. High-volume surgeons were more likely to operate on patients with malignant diagnoses and were more likely to perform concomitant bilateral salpingo-oophorectomy or staging at the time of hysterectomy. There were no other concomitant surgeries performed other than bilateral salpingo-oophorectomy (BSO) or staging procedures in this cohort.

TLH was the most common type of hysterectomy, performed in 60% of the cases, followed by LASH in 31% of the cases, and LAVH in 9% of the cases. High- and intermediate-volume surgeons performed TLH predominantly, while low-volume surgeons performed more LASH procedures. LASH procedures were more common in younger women with fibroids, while TLH procedures were more common in older women with malignancies.

On univariate analysis (Table 2), women who were operated on by high-volume surgeons had statistically significantly lower average EBLs than women operated on by intermediate- and low-volume surgeons. The median EBL for low-volume surgeons was double that of high-volume surgeons, and approximately one-third of women operated on by low-volume surgeons had an EBL >100 mL, compared to only 15% of women operated on by high volume surgeons (p = <0.001).

The median length of surgery was higher for low-volume surgeons than for intermediate- or high-volume surgeons, although this difference was not statistically different. One-third of women whose surgeries were performed by low-volume surgeons exceeded 164 minutes, compared to 21% of women operated on by intermediate- and 23% of women operated on by high-volume surgeons, respectively (p = 0.04). High-volume surgeons had fewer patients with LOS >1 day, although this was not statistically significant. Conversion to laparotomy and readmission rates were slightly higher in women operated on by high-volume surgeons. Surgeon volume was inversely related to resource utilization in that there was a statistically significant downward trend in high resource utilization as surgeon volume increased (trend p-value = 0.006). Low-volume surgeons had the highest rate of high resource utilization.

Intraoperative and postoperative complications were uncommon, occurring in 2% and 6% of the entire study cohort, respectively. Any surgical complications occurred in 8%, 5%, and 8% of patients whose surgery was performed by low-, intermediate- and high-volume surgeons, respectively. The rate of postoperative complications was higher than the rate of intraoperative complications, irrespective of the surgeon volume. There were no vascular injuries or hematomas. No patient had more than one intraoperative complication (urinary or bowel injuries); however, patients might have had more than one postoperative surgical complication. There were no significant differences in intraoperative, postoperative, or any complications among the patients operated on by low, intermediate, or high-volume surgeons.

Table 3 shows each type of postoperative surgical complication by surgeon volume using the Clavien–Dindo classification system. ²⁴ One patient, whose surgery was performed by a high-volume surgeon, had major postoperative surgical complications, including admission to the intensive care unit and death. Despite the older age and more-frequent diagnoses of malignancy, with concomitant staging procedures, among the patients operated on by high volume surgeons, there were no significant differences in surgical complications by surgeon volume.

On multivariate analysis, after adjusting for age, BMI, uterine weight, and prior abdominal surgery, there was a nearly twofold increased odds of higher resource utilization if the surgery was performed by a low- versus a high-volume surgeon (odds ratio: 1.75; confidence interval: 1.12–2.76; p = 0.05). There were no statistically significant differences in the intraoperative or postoperative surgical complications among the patients operated on by low-, intermediate-, and high-volume surgeons.

Discussion

In this cohort of 539 laparoscopic hysterectomies, surgeon volume affected outcomes as reflected primarily in resource utilization. Hysterectomies by low-volume surgeons were associated with higher blood loss, longer operating times, and a trend towards longer LOS than hysterectomies performed by intermediate- or high-volume surgeons. Low-volume surgeons had almost twofold increased odds of a patient needing higher resource utilization, compared to intermediate- or high-volume surgeons, after adjusting for appropriate confounders. Rates of intraoperative and postoperative surgical complications were low, and did not differ significantly among the groups, although high-volume surgeons operated on older women with more frequent diagnoses of malignancy, which required concomitant staging procedures.

One of the largest series to examine the role of surgeon volume and outcome for LHs found that those performed by high-volume surgeons (>14 LHs annually) produced modest reductions in complications (including intraoperative injuries, surgical-site complications, and medical complications) and costs, compared to low-volume surgeons (< 6 LHs annually). ²⁰ However, other studies of LH have not replicated this effect, showing only minimal reductions in operating time and conversion rates, or having low complication rates regardless of surgeon volumes.^15,19,26 Complication rates in the current cohort were consistent with those reported in other large studies of LHs.^{17,19–23,27–29} Seventy-five percent of cases in the current series had an EBL <100 mL, length of surgery <164 minutes, and LOS <1 day, which was the rationale to use these cutoff points to identify outliers in resource utilization in the study group. These cutoff points were comparable to those used in other large studies on surgical outcome for LH.^{17–22,27–29}

The median EBL for hysterectomy performed by a low-volume surgeon was twice that of the hysterectomies performed by high-volume surgeons: 100 mL versus 50 mL. While this difference may be of uncertain clinical significance, the frequency of an EBL >100 mL in LH performed by a low-volume surgeon was double that of LH performed by a high-volume surgeon: 32% versus 15%. Similar trends were noted for longer length of surgery and LOS, which contributed to the higher resource utilization for low-volume surgeons.

One of the challenges in interpreting current literature on the relationship between surgeon volume and outcomes is the variation in the definition of a high-volume laparoscopic surgeon used in the individual studies. Some studies included a large number of surgeons with a wide range of surgical volumes, while other studies were limited to a small number of very high-volume subspecialty-trained providers. The definition of high-volume surgeons in the literature ranges from more than a cumulative number of 30 laparoscopic hysterectomies, to >51 laparoscopic cases per year.^{17,19–22,29} Very few studies have reported outcomes among surgeons with <5 hysterectomies per year.^19,20,26 A recent meta-analysis of surgical outcomes in gynecologic surgery that included abdominal and vaginal hysterectomies as well as LHs concluded that surgeons performing <1 hysterectomy per month had higher rates of adverse outcomes. ³⁰ The lack of consensus regarding the minimum number of complex minimally invasive surgeries necessary to achieve—and maintain—surgical proficiency underscores the need to develop practical guidelines to determine and enhance minimally invasive surgical competence. ⁵

The current study included a group of 46 gynecologic surgeons with variations in their years of practice, their minimally invasive surgical and fellowship training, and their complex laparoscopic surgical volumes. The definition of low-, intermediate-, and high-volume surgeons in the current study fell within the range of most studies of LH, and included some of the lowest-volume surgeons, with <5 LHs per year.^{19,20,22,26,28} The rationale to include three tiers of surgeon volumes was determined specifically to include a large number of generalist gynecology surgeons who comprised the majority of the low- and intermediate-surgeon volume groups. Low-volume surgeons comprised 76% of the surgeons in this cohort and performed 26% of all the LHs. This finding was consistent with Wallenstein et al., who reported in 2012 that low-volume surgeons (<6 cases annually) performed the majority of surgeries in a commercial database of more than 124,000 LHs. ²⁰ These findings, therefore, are more generalizable to the gynecologic practices found in medical centers and community hospitals, compared to academic centers. ²⁸

Regarding the specific type of LH, the inherent variations in the indications for surgery and surgeon preferences can influence the surgical outcomes. Some studies have been limited to patients with fibroids, while other studies, like the present one, have included patents with benign and malignant diagnoses. High-volume surgeons used TLH in almost 80% of their cases. High-volume surgeons not only performed more LHs, but also this use of a single surgical approach could have facilitated increased skill and enhanced efficiency. In contrast, low-volume surgeons varied their approaches, potentially missing an opportunity to develop skill, and contributing to the higher EBLs and longer lengths of surgery observed in this cohort. Comparative analysis of laparoscopic approaches has also been inconclusive, and few data exist comparing all specific types of LH procedures to one another.^{17,19,23,27,29} LASH procedures in the current study were more common among younger women with larger uteri and lower BMIs, which echoes the findings reported by one of the only studies to compare different types of LH on surgical outcomes. ¹⁷

Strengths of the current study included its large sample size and diverse group of providers, the comprehensive medical review of all hospital-based records, and the use of multivariate analyses, which have only been included in a few studies investigating outcomes of LH and surgical volume.^17,20,23,26 The current authors did not report data on any learning-curve effect among the surgeons in this cohort over the study period, an effect that has been observed by other researchers and that might have had an impact on outcomes.^19,25 The study was not designed to evaluate the effect of fellowship training on patient outcome. Although the high-volume surgeons were subspecialists, only 1 had a minimally invasive surgery fellowship. Few studies have evaluated the impact of fellowship training on patient outcomes in laparoscopic gynecologic procedures.^17,26

Despite meticulous chart abstraction, a retrospective analysis can miss complications that would be found when using a prospective protocol. The current study did not investigate the healthcare costs associated with each procedure and their associated complications—an area in need of further investigation. Outcomes for vaginal hysterectomy, known to be the route associated with lowest cost, ⁶ were not examined in this study, in part, due to there being fewer cases of this kind. The low frequency of vaginal hysterectomy performed during the study period (12%) represents another avenue for improved resource utilization.

Within the current study's categorization of resource utilization, metrics, such as readmission and length of hospitalization, were included, as these two factors increase the cost of a hospital stay. Conversion to laparotomy has been shown to be less cost-effective; thus, that metric toward higher resource utilization was included.^31,32 Moreover, due to the rarity of blood transfusions, the current authors opted to examine EBL. EBL has been shown to contribute to reoperations and adverse surgical outcomes; thus, the current authors reason that higher EBL increases resource utilization. ³³ Finally, length of surgery has been associated with increased adverse events and readmissions.^34–35 The current authors understand that their categorization of resource utilization was just a proxy for possible estimates in costs and could not reflect the impact of volume on costs accurately.

Conclusions

These data support a recommendation that surgeons should invest in maintaining their laparoscopic skills. Though there is no consensus for the amount of LHs that constitute high- or low-surgical volume. The data from the current study indicated that that gynecologic surgeons should perform at least 5 LH procedures annually. Similar to studies in other surgical specialties, the current findings demonstrate that patients operated on by low-volume surgeons have higher resource utilization, which can impact patient outcomes and hospital efficiency.^19,20,22,26 As reimbursement is increasingly tied to performance metrics, measured competency in LH must include both surgical complications and resource utilization. Efforts to decrease the variations in healthcare utilization, while maintaining quality, will be increasingly important to hospitals and healthcare delivery systems that monitor surgeon volumes. ³⁶ Additional training resources, such as simulation centers and access to advanced laparoscopic and vaginal surgery skills courses, will be pivotal to helping gynecologic surgeons keep their skills honed and to optimize patient outcomes.