|
|
||||||||
ORIGINAL RESEARCH |
From the 1Department of Obstetrics and Gynecology and the 2Division of Research, Kaiser Permanente Northern California.
| ABSTRACT |
|---|
|
|
|---|
METHODS: All women, 20 years or age or older, who were undergoing hysterectomy for benign indications in Kaiser Permanente Northern California from 1994 to 2003 were identified. We analyzed hysterectomy rates by type, indication, and age group. Changes over time were analyzed with the Cochran-Armitage test for linear trend.
RESULTS: From 1994 to 2003, there were 32,321 hysterectomies performed for benign indications. Hysterectomy rates showed a significant decline, from 4.01 per 1,000 women in 1994 to 3.41 per 1,000 women in 2003 (P for trend < .001). The relative proportions of all hysterectomies performed as laparoscopically assisted vaginal hysterectomy (LAVH) peaked at 13.0% in 1995 and then steadily declined to 3.9% in 2003 (P for trend < .001), whereas the relative proportion of subtotal abdominal hysterectomy increased from 6.9% in 1994 to 20.8% in 2003 (P for trend < .001). Hysterectomy rates declined 11.2% for uterine leiomyoma (relative risk [RR] 0.89, 95% confidence interval [CI] 0.830.95), 33.1% for endometriosis (RR 0.67, 95% CI 0.590.76), and 18.6% for uterine prolapse (RR 0.81, 95% CI 0.720.92). The relative proportion performed for uterine leiomyoma was consistently greater than for all other indications combined.
CONCLUSION: The rates of hysterectomy for benign indications are decreasing. The type of hysterectomy changed significantly, with LAVH performed less frequently and subtotal abdominal hysterectomy increasing in popularity. Uterine leiomyoma remains the most common indication for benign hysterectomy.
LEVEL OF EVIDENCE: II-2
| MATERIALS AND METHODS |
|---|
|
|
|---|
Kaiser Permanente Northern California members are cared for exclusively by physicians and allied health professionals in The Permanente Medical Group, a multispecialty medical partnership of over 5,500 physicians, including over 500 ob-gyn physicians and 3 ob-gyn residency programs. The Kaiser Health Plan members have direct access to ob-gyn services. Surgical decisions are made autonomously between physicians and health plan members; there is no preauthorization or other administrative permission required. During the years of this study, there were no surgical or admission co-pays, and surgical care was a completely covered benefit of the health plan. The Permanente Medical Group physicians are salaried, and their reimbursement is not dependent upon the number of procedures performed or avoided.
Kaiser Permanente Northern California maintains comprehensive computerized clinical databases, including a hospitalization database that contains primary discharge diagnosis, 11 secondary diagnoses, primary procedure, and 7 secondary procedures using International Classification of Diseases, 9th Revision (ICD-9) codes. These databases capture information about hysterectomies performed in Kaiser Permanente Northern California ambulatory surgical centers and short stay admissions.
For each year from 1994 to 2003, study eligibility was limited to women who were 20 years of age or older on or before midnight of December 31 of the year prior. Additionally, continuous Kaiser Permanente Northern California membership for 12 months of a given study year was required to facilitate calculation of annual hysterectomy rates. All eligible women who had a hysterectomy for benign disease were identified. Total abdominal hysterectomy (TAH) was identified by procedure code 68.4 and subtotal abdominal hysterectomy by code 68.3. Vaginal hysterectomy was identified by codes 68.5 and 68.59 without associated laparoscopic codes. Laparoscopically assisted vaginal hysterectomy (LAVH) was identified by codes 68.51, 68.31, or vaginal hysterectomy with associated laparoscopic codes (54.21, 65.01, 65.31, 65.41, 65.53, 65.54, 65.63, or 65.64). Laparoscopic supracervical hysterectomy and total laparoscopic hysterectomy was defined as either 1) ICD-9 code 68.31 (introduced for laparoscopic supracervical hysterectomy, October 1, 2003) or 2) concomitant ICD-9 code 68.3 (subtotal hysterectomy) or 68.4 (total abdominal hysterectomy) AND associated laparoscopic code. Only 10 laparoscopic supracervical hysterectomy procedures were identified, and they were included in the LAVH group for analysis. Data on radical hysterectomy (68.7) were not collected. Women were excluded if any of the discharge diagnoses included cancer of the reproductive tract or precancerous condition of the reproductive tract using the following codes: cervical cancer (180), uterine cancer (179, 182, 630), ovarian cancer (183), other genital cancers (184), carcinoma in situ of cervix, uterus, or other female genital (233.1, 233.2, 233.3), and endometrial hyperplasia (621.3).
Primary indication for benign hysterectomy was assigned by the first listed diagnostic code for uterine leiomyoma (218), endometriosis (617), or uterine prolapse/incontinence (618, 625.6, and 788.3), and "other" was assigned for the remaining diagnoses. A similar methodology for assigning the leading cause of hysterectomy has been used previously.2 Some studies have separated out abnormal uterine bleeding from an "other" category.3,4 However our preparatory work found that, overall, 12.5% of hysterectomies fit into the "other" group, of which only 5.1% were for abnormal uterine bleeding, and therefore we did not separate out this indication.
Data on race was not available for analysis because Kaiser Permanente Northern California does not routinely collect racial data on its members. However, periodic random health surveys of members are performed every 3 years on less than 1% of the membership, with approximately a 55% response rate. These surveys include racial data. For survey years 1996, 1999, and 2002, among women who were more than 20 years of age, the following racial data were reported: white/Hispanic ranged from 64.7% to 68.9%; African Americans ranged from 6.5% to 7.0%, Hispanics ranged from 9.3% to 11.4%, and Asians ranged from 12.9% to 15.6%.
Annual benign hysterectomy rates per 1,000 eligible women were calculated. Corrections for previous hysterectomy were not made to the denominator, an approach similar to that in previous large national hysterectomy studies. Results of periodic random health surveys by Kaiser Permanente Northern California suggested no significant change in the percentage of women with previous hysterectomy. In 1996, 18.5% of women in the sample surveyed self-reported a previous hysterectomy, compared with 17.7% in 2002 (internal Kaiser data). In addition to indication for hysterectomy and type of hysterectomy, cases were analyzed by 5-year age groups (2024, 2529, 3034, 3539, 4044, 4549, 5054, and
55 years). Statistical software (SAS 9.1; SAS Institute Inc, Cary, NC) was used for analyses. Changes over time were tested by the Cochran-Armitage test for linear trend.5 All tests were 2-sided; P values of .05 or less were considered to indicate significance. Relative risks (RRs) and 95% confidence intervals (CIs) are presented when comparing rates between 2 specific years. This study was approved by the Kaiser Permanente Northern California Institutional Review Board on April 9, 2004.
| RESULTS |
|---|
|
|
|---|
|
Statistical trend tests of the annual rates for total abdominal hysterectomy, vaginal hysterectomy, and LAVH all showed significant declines, whereas the rate for subtotal abdominal hysterectomy increased significantly (Table 2). Compared with hysterectomy rates in 1994, rates in 2003 decreased 26.1% (RR 0.74, 95% CI 0.690.79) for total abdominal hysterectomy, decreased 7.4% (RR 0.93, 95% CI 0.841.02) for vaginal hysterectomy, decreased 73.2% (RR 0.27, 95% CI 0.220.33) for LAVH, and increased 156% (RR 2.56, 95% CI 2.192.99) for subtotal abdominal hysterectomy. Overall, the relative proportion of vaginal hysterectomy trended down (P = .02), fluctuating from a peak of 27.1% in 1997 to a nadir of 21.9% in 2001, whereas the relative proportion of LAVH declined steadily (P < .001) from 13.0% in 1995 to 3.9% in 2003 (Fig. 1). The relative proportion of total abdominal hysterectomy also trended down (P < .001), ranging from 49% to 57%, whereas the relative proportion of subtotal abdominal hysterectomy increased 3-fold from 6.9% in 1994 to 20.8% in 2003 (P for trend < .001).
|
|
The hysterectomy rate for uterine leiomyoma was greater than all other indications combined in each year studied (Table 3). Annual hysterectomy rates declined significantly for the 3 indications identified (P for trend < .001 for each) (Table 3). Compared with hysterectomy rates in 1994, rates in 2003 decreased 11.2% (RR 0.89, 95% CI 0.830.95) for uterine leiomyoma, decreased 33.1% (RR 0.67, 95% CI 0.590.76) for endometriosis, decreased 18.6% (RR 0.81, 95% CI 0.720.92) for prolapse, and decreased 5.7% (RR 0.94, 95% CI 0.831.08) for "other" indications. Figure 2 shows the relative proportions of hysterectomy by indication. Although tests for trend showed a statistically significant increase in the relative proportion of hysterectomy for uterine leiomyoma (P = .006) and "other" indications (P = .03) and a decrease for endometriosis (P = .006) and uterine prolapse (P = .02), the overall trends were relatively flat.
|
|
Hysterectomy rates were consistently highest for women aged 4044 years and 4549 years, although the increase in rates for these 2 age groups was not statistically significant (P for trend = .28 and .06, respectively; Fig. 3). Statistical tests for trend showed a significant decrease in hysterectomy rates for all other age groups (P < .01 for each). Figure 4 shows the relative proportions of hysterectomy by indication in each age group. Uterine leiomyoma was the most common indication for hysterectomy among the age groups from 3554 years, and uterine prolapse was the most common indication for women age 55 years or older.
|
|
| DISCUSSION |
|---|
|
|
|---|
We do not believe there is anything unique to the Kaiser Permanente Northern California health care model that would explain the decreasing hysterectomy trends found in our study. Patients have easy access to physicians and also have the opportunity to freely seek second opinions from other surgeons within the medical group. There have never been any organizational policies or concerted efforts to decrease the number of hysterectomies. We believe the individual practice patterns of The Permanente Medical Group gynecologists are similar to those of the geographic communities in which they practice. They have the same opportunities and motivations to learn new surgical techniques. We do believe that there is a culture within the organization that encourages the delivery of cost-effective medicine, but this does not necessarily translate to efforts to decrease hysterectomy rates.
It is possible that increased use of laparoscopic and hysteroscopic procedures, development and acceptance of endometrial ablation devices, introduction of a progestin-based intrauterine device, and the emergence of uterine artery embolization have begun to substitute for hysterectomy in our population. Kaiser Permanente Northern California members may have greater access to alternative treatments because these procedures are often covered by the health plan, as are drug benefits encompassing intrauterine devices and depot leuprolide. Utilization trend data on these alternative treatments during our study period could help measure their impact on hysterectomy rates and is currently an active area of research at Kaiser Permanente Northern California. The level of information available to patients on different treatment alternatives may also be greater among our population. Additionally, changes in patient preferences for hysterectomy or alternative treatments, as well as changes in provider preferences and counseling, could also contribute to decreasing hysterectomy rates. Although there have been no centrally directed efforts from Kaiser Permanente Northern California to decrease hysterectomy rates or promote alternative treatments, there may be unappreciated incentives for the health plan and local facility efforts by The Permanente Medical Group physicians to encourage alternative treatments. Given that the decrease in hysterectomy rates varied by indication, most likely a number of different factors are contributing.
We report significant changes in the type of hysterectomy being performed. Although the relative proportion of vaginal hysterectomies remained relatively stable, the relative proportion of LAVH decreased significantly with a corresponding increase in hysterectomies performed abdominally. Additionally, there was a dramatic change among type preference for abdominal hysterectomy, with the relative proportion of subtotal hysterectomies increasing 3-fold, from 7% to 21%. Although the abdominal hysterectomy rate by laparotomy (total plus subtotal abdominal hysterectomy) decreased from 2.58 per 1,000 women in 1994 to 2.41 per 1,000 in 2003, the increase in relative proportion of subtotal hysterectomy caused the relative proportion of abdominal hysterectomy by laparotomy to increase from 64.1% in 1994 to 70.6% in 2003. Farquhar and Steiner3 also noted a 3-fold increase in subtotal hysterectomy in U.S. data, but it peaked at only 2%, whereas a 10-year report on Danish women found a 5-fold increase reaching 22% in 1998.6 This change in preference is hard to explain, given the lack of evidence to support this practice. Most likely physician and patient perception of the merit of subtotal hysterectomy had a significant influence: women may believe subtotal hysterectomy may have less impact on sexuality, while physicians might consider the procedure less likely to effect bowel and bladder function and to have lower rates of infectious morbidity, be less likely to lead to prolapse, and have lower operative time and less blood loss than total abdominal hysterectomy. Given the lack of convincing evidence favoring this approach, the motivating reasons behind this change in practice pattern remain unclear.
The finding of decreasing relative proportion of LAVH was unexpected. Weber and Lee4 reported the percentage of LAVH increased from less than 1% to 7.5%, which was similar to the increase estimated by Farquhar and Steiner3 of less than 1% to 9.9%. Keshavarz et al2 did not specifically report the percentage of LAVH but found that the proportion of all vaginal hysterectomies with concomitant laparoscopy increased from 13% to 28%. Although most studies suggest an increase in the percentage of LAVH in the 1990s, a recent survey found a decrease in LAVH from 1995 to 2000 among members of the American Association of Gynecologic Laparoscopists.7 There are several possible explanations for the decrease in LAVH. Purported advantages of LAVH exist only if it replaces abdominal hysterectomy without increasing morbidity or costs. Laparoscopically assisted vaginal hysterectomy is a technique often best suited for a smaller uterus. With hysterectomy rates for indications other than fibroids decreasing, there may be fewer hysterectomies performed on uteri in the appropriate size range for LAVH. With the decreasing hysterectomy rates, it could be more difficult to overcome the initial learning curve needed to feel confident with LAVH. Farquhar and Steiner3 reported that costs for LAVH doubled in the time period studied and by 1997 were greater than abdominal and vaginal hysterectomy. Weber and Lee4 also reported that median hospital charges in Ohio were greater for LAVH than for abdominal or vaginal hysterectomy. It is possible that an increase in hysterectomy by mini-laparotomy is responsible for the decline in relative proportion of LAVH, as well as the increase in subtotal hysterectomy. Informal surveys of ob-gyn chiefs at Kaiser Permanente Northern California facilities suggest this may be the case. The purported advantages of mini-laparotomy over LAVH include a less steep learning curve and less expense for equipment while maintaining similar short hospitalization time and rapid recovery. It is currently not possible to differentiate hysterectomy done by mini-laparotomy by ICD-9 coding.
Our findings that uterine leiomyoma was the most common indication for hysterectomy is consistent with national reporting. Although hysterectomy rates decreased for all indications studied as well as for most age groups, the increase in hysterectomy rate among women aged 4049 years was driven by hysterectomy for uterine leiomyoma. This is of particular importance because women aged 4049 years consistently had the highest hysterectomy rates. It is possible that either the actual or perceived success of alternative therapies for uterine leiomyoma is less than that for other indications. Another explanation is that alternative treatments for uterine leiomyoma may be used to treat more women and not necessarily replace hysterectomy. Two studies measuring the impact of endometrial ablation on hysterectomy rates concluded that these techniques were more additive than substitutive.8,9
Our study has several important limitations. Our findings are most applicable to the population served by Kaiser Permanente Northern California. We do not report on alternative treatments that could impact hysterectomy rates. We also do not report on race, although this limitation is not unique to our study. Keshavarz et al2 reported that race significantly influenced type of hysterectomy as well as hysterectomy rates for specific indications, but racial data were missing for 1720% of their study population. The study by Weber and Lee4 was missing racial demographics on 27% of their population, and Farquhar and Steiner3 reported that data on race was too incomplete to present. Our internal data did suggest that the overall percentage of African Americans in our population was low and stable at around 7%, but there were greater fluctuations for white/non-Hispanics, Hispanics, and Asians. The lack of specific racial data on women undergoing hysterectomy prevents analysis of the racial impact on overall hysterectomy rates as well as type- and indication-specific rates. Our data are also subject to coding inaccuracies. We did not measure changes in The Permanente Medical Group physician group demographics that could influence practice style, such as physician gender and surgical experience. Additionally, our hysterectomy rates are most likely underestimations because corrections for previous hysterectomy were not made in rate calculations.
Our study has several unique strengths. The Kaiser Permanente Northern California membership is a relatively large and stable population, with 1.2 million female members in June 1993, of which over 60% were still members 9 years later (internal Kaiser data). The databases are robust and complete, with rates calculated from the entire population and not extrapolated from sample estimates. Additionally, our databases capture hysterectomy done in ambulatory centers and those admitted for only short stay hospitalization.
Although we found a significant decrease in hysterectomy rates for benign disease in a large population, we could not elucidate the reasons for this. Future studies should focus on the potential impact of alternative surgical and medical treatments on hysterectomy rates. Research on disease-specific hysterectomy rates would be most valuable, as well as studies on motivating factors for patients and physicians regarding preferences for alternative treatments and hysterectomy types. Studies are also needed to investigate the unique risks and benefits associated with hysterectomy by mini-laparotomy. The impact of race on disease-specific hysterectomy rates needs investigation. Possible differences in access, practice style, or other undetermined factors between a large managed care organization such as Kaiser Permanente Northern California and commercial health care products would also be of interest.
| Footnotes |
|---|
The authors thank Bruce Folck for his invaluable assistance in database abstraction and statistical analysis.
The abstract for this study was accepted for oral presentation to the Annual Clinical Meeting of the American College of Obstetricians and Gynecologists in Washington, DC, May 9, 2006.
Corresponding author: Gavin Jacobson, MD, 395 Hickey Boulevard, Daly City, CA, 94015; e-mail: gavin.Jacobson{at}kp.org.
doi:10.1097/01.AOG.0000210640.86628.ff
| REFERENCES |
|---|
|
|
|---|
2. Keshavarz H, Hillis SD, Kieke BA, Marchbanks PA. Hysterectomy surveillance: United States, 19941999. MMWR CDC Surveill Summ. 2002;51(SS-5):18.
3. Farquhar CM, Steiner CA. Hysterectomy rates in the United States 19901997. Obstet Gynecol 2002;99:22934.
4. Weber AM, Lee JC. Use of alternative techniques of hysterectomy in Ohio, 19881994. N Engl J Med 1996;335:4839.
5. Armitage P. Tests for linear trends in proportions and frequencies. Biometrics 1955;11:37586.
6. Gimbel H, Settnes A, Tabor A. Hysterectomy on benign indication in Denmark 19881998: a register based trend analysis. Acta Obstet Gynecol Scand 2001;80:26772.[Medline]
7. Kives SL, Levy BS, Levine RL. Laparoscopic-assisted vaginal hysterectomy: American Association of Gynecologic Laparoscopists 2000 membership survey. J Am Assoc Gynecol Laparosc 2003;10:1358.[Medline]
8. Bridgman SA, Dunn KM. Has endometrial ablation replaced hysterectomy for the treatment of dysfunctional uterine bleeding? National figures. BJOG 2000;107:5314.[Medline]
9. 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;18:62534.[Medline]
This article has been cited by other articles:
![]() |
T. H.I. Brummer, T. T. Seppala, and P. S.M. Harkki National learning curve for laparoscopic hysterectomy and trends in hysterectomy in Finland 2000-2005 Hum. Reprod., April 1, 2008; 23(4): 840 - 845. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. M. Wu, M. E. Wechter, E. J. Geller, T. V. Nguyen, and A. G. Visco Hysterectomy Rates in the United States, 2003 Obstet. Gynecol., November 1, 2007; 110(5): 1091 - 1095. [Abstract] [Full Text] [PDF] |
||||
![]() |
V. Teplin, E. Vittinghoff, F. Lin, L. A. Learman, H. E. Richter, and M. Kuppermann Oophorectomy in Premenopausal Women: Health-Related Quality of Life and Sexual Functioning Obstet. Gynecol., February 1, 2007; 109(2): 347 - 354. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |