Jacoby, Vanessa L. MD, MAS1; Autry, Amy MD1; Jacobson, Gavin MD2; Domush, Robert MD1; Nakagawa, Sanae MA1; Jacoby, Alison MD1
The introduction of laparoscopy into benign gynecology has changed hysterectomy practice patterns dramatically. The proportion of hysterectomies performed laparoscopically has increased significantly during the past 20 years, from 0.3% in 1990 to 11.8% in 2003.1,2 This rise in laparoscopy has been coupled with a marked decrease in the rate of abdominal hysterectomies from 74% to 60% and a slight decrease in the proportion of vaginal hysterectomies from 24% to 22%.1,2 Gynecologists may favor a laparoscopic approach compared with an abdominal approach because of shorter hospital stays, faster recovery times, less blood loss, and fewer infections. However, these benefits are weighed against the reported increase in urologic injury with laparoscopic hysterectomy.3 In contrast, compared with vaginal hysterectomy, the laparoscopic route has not shown any significant advantages in randomized trials.3
Although rates of laparoscopic hysterectomy have increased overall, the rise in laparoscopy may not be distributed uniformly. Several factors have been associated with differential use of laparoscopic hysterectomy, including geographic location, race/ethnicity, income, and health insurance status.1,4,5 However, these studies have not examined the full range of factors that may influence the decision to perform laparoscopy, in particular the indication for hysterectomy. Our aim is to describe nationwide laparoscopy practice patterns and to identify independent demographic, clinical, and health-system factors associated with the use of laparoscopy in a diverse population of women undergoing hysterectomy for benign conditions.
MATERIALS AND METHODS
This is a cross-sectional analysis of the 2005 Nationwide Inpatient Sample, a national database sponsored by the Agency for Healthcare Research and Quality. The Nationwide Inpatient Sample is a 20% stratified random sample of discharges from all community hospitals in the United States. Hospitals considered for sampling include nonfederal, general, and specialty short-term hospitals, including public and academic facilities. The sampling scheme of the Nationwide Inpatient Sample represents approximately 90% of all hospitals. It is the largest all-payer database of hospital discharges, with 8 million hospital stays in the 2005 Nationwide Inpatient Sample from 37 states. The 2005 Nationwide Inpatient Sample was the most recent version of this database available at the time of our analysis. The study was deemed exempt by the institutional review board at the University of California, San Francisco.
Each record in the Nationwide Inpatient Sample contains a maximum of 15 procedure codes and 15 diagnostic codes classified using both the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM), and the Clinical Classification software developed by the Agency for Healthcare Research and Quality. The Clinical Classification software is a categorization scheme that collapses ICD-9 codes into clinically meaningful categories that are useful for descriptive analysis.
Procedure codes from the ICD-9-CM were used to categorize all women age 18 years or older who underwent hysterectomy into one of three groups: 1) abdominal (683.9 for subtotal abdominal hysterectomy, 684.9 for total abdominal hysterectomy), 2) vaginal (685.9), or 3) laparoscopic (683.1 for subtotal laparoscopic hysterectomy, 685.1 for laparoscopic-assisted vaginal hysterectomy). In 2005, there was not an ICD-9-CM for total laparoscopic hysterectomy. Women with ICD-9-CM procedure codes for cesarean delivery with concomitant hysterectomy or with any ICD-9-CM code or Clinical Classification Software code for cancer of the gynecologic, gastrointestinal, or genitourinary tract, lymphoma, or malignant neoplasm without specification of site were excluded.
Predictor variables were classified as demographic (age, race/ethnicity, income, region of the country, rural/urban hospital setting), clinical (surgical diagnosis, concomitant adnexal surgery), or health system (primary payer, hospital bed count, hospital teaching status). Data on age, race/ethnicity, income, region, and hospital setting were derived from predefined Nationwide Inpatient Sample categories. Primary expected payer and hospital size and teaching status also were extrapolated from available Nationwide Inpatient Sample categorical variables. The seven categories for surgical diagnosis were constructed using the following codes: 1) fibroids: ICD-9-CM codes 218.0, 218.1, 218.2, or 218.9; 2) endometriosis: Clinical Classification software code 169; 3) pelvic infection: ICD-9-CM codes 614.0–614.9, 615.0, 615.1, 615.9, 616.10, 616.11, or 616.2–616.5; 4) prolapse: ICD-9-CM codes 618.0–618.05, 618.09, 618.1–618.4, 618.6–618.8, 618.81–618.83, 618.89, or 618.9; 5) abnormal bleeding: ICD-9-CM codes 626.0–626.6, 626.8–627.1, or 626.7; and 6) pelvic pain: ICD-9-CM diagnosis codes 625.0 or 625.2–625.5. Each patient had a maximum of 15 diagnosis codes listed at the time of surgery; these codes were not treated as mutually exclusive in our analysis because many patients have multiple indications for undergoing hysterectomy. Each indication was represented in the model by a separate indicator variable. The resulting odds ratio (OR) for each indication can be interpreted as the odds of undergoing laparoscopy for women with that diagnosis compared with women without that diagnosis, holding all other indications and other covariates constant.
Women who underwent concomitant adnexal surgery with hysterectomy were placed in the predictor group unilateral salpingo-oophorectomy or bilateral salpingo-oophorectomy. This category included women who underwent either unilateral or bilateral oophorectomy or salpingo-oophorectomy, which we identified using the following ICD-9-CM codes: 656.1, 656.3, 655.1, 655.3, 656.2, 656.4, 655.2, 655.4, 654.9, 654.1, 653.1, 653.9. Total charges are reported in the Nationwide Inpatient Sample based on the overall dollar amount charged for the entire hospital stay excluding professional fees and noncovered charges. Length of stay is calculated in the Nationwide Inpatient Sample by subtracting the admission date from the discharge date with same-day stays coded as 0 days.
To account for the sampling design of the Nationwide Inpatient Sample, special survey procedures were used in SAS 9.12 (SAS Institute, Inc., Cary, NC). Thus, all analyses use the inverse probability of selection weights provided in the data set and account for stratification of the sample by geographic region, type of control (public, not-for-profit, proprietary), location (urban or rural), teaching status, and bed count (small, medium, large). The analyses also account for clustering of patient outcomes within hospitals, the primary sampling units. The subgroup and overall totals we present reflect the inverse probability weights and thus can be interpreted as estimates of totals in the target population.
Logistic regression was used to assess the independent associations of demographic, clinical, or health-system factors associated with undergoing laparoscopic compared with abdominal hysterectomy and laparoscopic compared with vaginal hysterectomy. Predictors were selected a priori on substantive grounds. All were included in the multivariable models to avoid inflation of the type I error rate potentially induced by model selection. The very large Nationwide Inpatient Sample size accommodates this large number of predictors. Total charges and length of stay between hysterectomy approaches were assessed using t-tests.
To account for the influence of the large number of observations with missing data for race/ethnicity (28% in the laparoscopy compared with abdominal hysterectomy analysis and 27% in the laparoscopy compared with vaginal hysterectomy analysis), all analyses are presented with the missing values for race/ethnicity treated as a separate category (as indicated in table footnotes). In this approach, all women with missing data for race/ethnicity are included in the analyses.
In the 2005 Nationwide Inpatient Sample, there were 518,828 women who underwent hysterectomy for benign gynecologic conditions who were included in our analysis; 14% of hysterectomies were laparoscopic, 64% abdominal, and 22% vaginal. Table 1 demonstrates the general characteristics of the hysterectomy population by surgical route. The majority of women were white and had private insurance, irrespective of hysterectomy approach. The most common surgical diagnosis for vaginal hysterectomy was prolapse (62%), whereas fibroids (62%) was the most common diagnosis for abdominal hysterectomy and abnormal bleeding (53%) the most common diagnosis for laparoscopic hysterectomy. The majority of women who underwent either laparoscopic (60%) or abdominal (68%) hysterectomy had concomitant unilateral or bilateral salpingo-oophorectomy compared with only 26% of women who had vaginal hysterectomy. Approximately 40% of all hysterectomies, irrespective of route, occurred in the South.
There were several independent predictors of undergoing laparoscopic hysterectomy compared with abdominal hysterectomy in the multivariable model (Table 2). Women age 35–49 years (P<.001) and those 55 years or older (P<.02) were less likely to undergo laparoscopic hysterectomy compared with women 18–34 years. Race/ethnicity was a significant predictor; African-American, Latina, and Asian women had 40–50% lower odds of laparoscopic hysterectomy (P<.001). Women with household incomes in all three categories less than $61,000 were less likely to undergo laparoscopy (OR 0.75, 95% confidence interval [CI] 0.61–0.91 for the lowest income level, less than $37,000). Compared with the Northeast, the West had a significantly higher laparoscopy rate (OR 1.77, 95% CI 1.20–2.62). A diagnosis of fibroids or pelvic infection was associated with approximately 30% lower odds of laparoscopy (P<.001), and the odds of laparoscopy were 29% lower in women who underwent concomitant unilateral or bilateral salpingo-oophorectomy (P<.001). Women with Medicare, Medicaid, or no health insurance (self-pay or no charge/charity) were 29–59% less likely to undergo laparoscopic hysterectomy, irrespective of race/ethnicity. Hospital setting (rural compared with urban), teaching status, and bed count were not associated with surgical route.
In comparison with vaginal hysterectomy, several factors were found to be associated with the odds of undergoing laparoscopic hysterectomy (Table 3). Older age was associated with decreased odds of laparoscopy, particularly for women age 55 years or older, who were 62% less likely to undergo laparoscopy (P<.001), even accounting for prolapse as the surgical diagnosis. Low household income also was associated with lower rates of laparoscopy; women in all three income categories less than $61,000 were less likely to have laparoscopic hysterectomy (OR 0.76, 95% CI 0.61–0.95 for less than $37,000). Race/ethnicity and region of the country were not statistically significant predictors of laparoscopy. Women with fibroids (P<.001), endometriosis (P<.001), pelvic infection (P<.001), or pelvic pain (P=.003) were more likely to undergo laparoscopy, but those with uterine prolapse had 80% lower odds of laparoscopy (P<.001). Women who had concomitant unilateral or bilateral salpingo-oophorectomy were nearly six times as likely to undergoing laparoscopic hysterectomy (P<.001). Women with Medicare, Medicaid, or no health insurance (self-pay or no charge/charity) were 27–71% less likely to undergo laparoscopy. Hospital setting and teaching status were not associated with laparoscopy rates, but patients in medium–bed count and large–bed count hospitals were about 30% less likely to have laparoscopic hysterectomy.
Length of hospital stay and total charges for the surgery and hospital admission were significantly different between surgical routes. The laparoscopic approach had the shortest average length of stay at 1.65 days, compared with 1.86 for vaginal hysterectomy and 3.07 for abdominal hysterectomy (P<.001 for pair-wise comparisons, laparoscopy as reference group). However, laparoscopic hysterectomy had the highest total charges at $18,821, compared with $17,839 for abdominal hysterectomy and $14,121 for vaginal (P<.001 for pair-wise comparisons, laparoscopy as reference group).
In this large, nationwide sample of more than 500,000 women who underwent hysterectomy for benign conditions, we identified several independent factors that are associated with undergoing a laparoscopic procedure compared with a vaginal or abdominal hysterectomy. As expected, clinical factors such as surgical diagnosis and concomitant adnexal surgery were associated independently with the use of laparoscopy. For instance, women with possible extrauterine disease (endometriosis or pelvic infection) were more likely to undergo laparoscopic compared with vaginal hysterectomy. However, race/ethnicity, income, and insurance also were associated highly with the use of laparoscopy. These findings indicate that nonclinical patient characteristics influence the use of laparoscopy and that medically underserved women may be excluded from the widespread use of this technology.
Women without private health insurance, including those with Medicaid, Medicare, self-pay, or no charge/charity, were less likely to undergo a laparoscopic approach to hysterectomy compared with both the abdominal and vaginal route. Overall, hysterectomy reimbursement from Medicaid and Medicare is significantly lower than that from most private insurers, and these programs reimburse $100–$200 less for a laparoscopic approach compared with abdominal or vaginal. Therefore, physicians have a financial disincentive to perform laparoscopy in women enrolled in these insurance programs. Similarly, when the hospital absorbs the cost of care for uninsured women (no charge/charity) or patients pay out-of-pocket (self-pay), the desire to contain costs may drive the decision to avoid laparoscopy.
Race/ethnicity was a predictor of surgical route, with African-American, Latina, and Asian women less likely to undergo laparoscopic hysterectomy compared with abdominal hysterectomy, even accounting for confounding by surgical diagnosis and all other factors in our multivariable model. Two previous studies also have found lower rates of laparoscopy among nonwhite women undergoing hysterectomy, although these analyses did not account for the effect of diagnosis on the choice of surgical route.1,4 We controlled for surgical diagnosis, including the presence of uterine fibroids, and still found racial/ethnic differences in the use of laparoscopy compared with abdominal hysterectomy. Differences in surgical approach by race/ethnicity are not unique to hysterectomy. African-American and Latina adults who undergo appendectomy also are less likely to receive laparoscopic surgery,6 and African-American patients have lower rates of laparoscopy compared with open cholecystectomy,7 suggesting a larger trend in laparoscopy practice. We were unable to determine whether patient preference, provider bias, or some other unmeasured factor was responsible for the racial/ethnic differences we observed in hysterectomy route. Further investigation is warranted to better understand these differences in practice and why they did not persist for the comparison of laparoscopy with vaginal hysterectomy.
Irrespective of race or ethnicity, lower income was associated with lower odds of undergoing laparoscopy in our analysis. This is consistent with studies of breast cancer and lung cancer patients that have found differences in surgical treatment based on socioeconomic status.8,9 There likely are several factors that contribute to lower laparoscopy rates among low-income hysterectomy patients. First, low-income women may have lower health literacy and therefore inquire less frequently about newer technologies for hysterectomy. Low health literacy has been shown to affect surgical decision making and to decrease postoperative satisfaction.10 In addition, low-income women may not have access to gynecologists who have the training and skill to perform laparoscopic procedures or hospitals that have the required laparoscopic equipment.
Older age was associated with lower rates of laparoscopy in our analyses. In women younger than 35 years, providers and patients may be highly focused on cosmetic outcomes and opt for smaller incisions with laparoscopy compared with laparotomy. Age differences in surgical route were most pronounced comparing laparoscopy with vaginal hysterectomy, irrespective of uterine prolapse as the surgical diagnosis. There was a steady decline in the likelihood of laparoscopy for each increasing age category above 34 years, with the oldest group of women 55 years or older 62% less likely to undergo laparoscopy. Increasing parity or pelvic relaxation or both with increasing age may explain these lower rates of laparoscopy compared with vaginal hysterectomy. Unfortunately, we were unable to assess these variables in the Nationwide Inpatient Sample database. In addition, younger patients might seek out younger gynecologists with less training and comfort in performing vaginal hysterectomy compared with a laparoscopic approach.11
There are several limitations to our analysis. The Nationwide Inpatient Sample is a large database that receives input from 32 state databases. Therefore, errors in coding and classification of all predictor variables are possible. In addition, the data set does not provide information on outpatient hysterectomies or physician characteristics such as specialty training, sex, surgical volume, or practice setting that may effect surgical route.12–14 Finally, there was a large amount of missing data (approximately 30%) for the race/ethnicity predictor. However, all of our models included women with missing data for this predictor in the complete case analysis, and we do not believe that missing data for race/ethnicity would be associated systematically with the outcome of interest—hysterectomy surgical route. In addition, models that excluded women with missing data for race/ethnicity were not significantly different from the complete case analyses, and the magnitude of associations we found for race/ethnicity are unlikely to be affected by possible residual confounding owing to missing data.
The ideal surgical approach to hysterectomy should be guided by the patient’s clinical presentation, the experience and skill of the provider, evidence-based practice, and patient preferences. Although we believe that many hysterectomy decisions incorporate this complex set of factors, our analyses indicate that there may be a nationwide bias toward decreased use of laparoscopy among hysterectomy patients who are low income, do not have private health insurance, or are part of a racial or ethnic minority group. Although the benefit of laparoscopy for performing hysterectomy is debatable, particularly compared with vaginal hysterectomy, the option of this surgical approach should be readily available for appropriately selected patients. Further research is needed to better understand the etiology of these significant variations in nationwide laparoscopy practice.
1.Wu JM, Wechter ME, Geller EJ, Nguyen TV, Visco AG. Hysterectomy rates in the United States, 2003. Obstet Gynecol 2007;110:1091–5.
2.Farquhar CM, Steiner CA. Hysterectomy rates in the United States 1990-1997. Obstet Gynecol 2002;99:229–34.
3.Johnson N, Barlow D, Lethaby A, Tavender E, Curr E, Garry R. Surgical approach to hysterectomy for benign gynaecological disease. The Cochrane Database of Systematic Reviews 2006, Issue 2. Art No.: CD003677. DOI: 10.1002/14651858.CD003677.pub4.
4.Abenhaim HA, Azziz R, Hu J, Bartolucci A, Tulandi T. Socioeconomic and racial predictors of undergoing laparoscopic hysterectomy for selected benign diseases: analysis of 341487 hysterectomies. J Minim Invasive Gynecol 2008;15:11–5.
5.Jacobson GF, Shaber RE, Armstrong MA, Hung YY. Hysterectomy rates for benign indications. Obstet Gynecol 2006;107:1278–83.
6.Guller U, Jain N, Curtis LH, Oertli D, Heberer M, Pietrobon R. Insurance status and race represent independent predictors of undergoing laparoscopic surgery for appendicitis: secondary data analysis of 145,546 patients. J Am Coll Surg 2004;199:567–75.
7.Arozullah AM, Ferreira MR, Bennett RL, Gilman S, Henderson WG, Daley J, et al. Racial variation in the use of laparoscopic cholecystectomy in the Department of Veterans Affairs medical system. J Am Coll Surg 1999;188:604–22.
8.Greenwald HP, Polissar NL, Borgatta EF, McCorkle R, Goodman G. Social factors, treatment, and survival in early-stage non-small cell lung cancer. Am J Public Health 1998;88:1681–4.
9.Johantgen ME, Coffey RM, Harris DR, Levy H, Clinton JJ. Treating early-stage breast cancer: hospital characteristics associated with breast-conserving surgery. Am J Public Health 1995;85:1432–4.
10.Hawley ST, Janz NK, Hamilton A, Griggs JJ, Alderman AK, Mujahid M, et al. Latina patient perspectives about informed treatment decision making for breast cancer. Patient Educ Couns 2008;73:363–70.
11.Kenton K, Sultana C, Rogers RG, Lowenstein T, Fenner D, American Urogynecolgic Society Education Committee. How well are we training residents in female pelvic medicine and reconstructive surgery? Am J Obstet Gynecol 2008;198:567 e1–4.
12.Gretz H, Bradley WH, Zakashansky K, Nezhat F, Rahaman J, Chuang L, et al. Patient clinical factors influencing use of hysterectomy in New York, 2001-2005. Am J Obstet Gynecol 2008;199:349.e1–5.
13.Bickell NA, Earp JA, Garrett JM, Evans AT. Gynecologists’ sex, clinical beliefs, and hysterectomy rates. Am J Public Health 1994;84:1649–52.
14.Geller SE, Burns LR, Brailer DJ. The impact of nonclinical factors on practice variations: the case of hysterectomies. Health Serv Res 1996;30:729–50.