Objective
We sought to analyze use of alternative treatments and pathology among women who underwent hysterectomy in the Michigan Surgical Quality Collaborative.
Study Design
Perioperative hysterectomy data including demographics, preoperative alternative treatments, and pathology results were analyzed from 52 hospitals participating in the Michigan Surgical Quality Collaborative from Jan. 1 through Nov. 8, 2013. Women who underwent hysterectomy for benign indications including uterine fibroids, abnormal uterine bleeding (AUB), endometriosis, or pelvic pain were eligible. Pathology was classified as “supportive” when fibroids, endometriosis, endometrial hyperplasia, adenomyosis, adnexal pathology, or unexpected cancer were reported and “unsupportive” if these conditions were not reported. Multivariable analysis was done to determine independent associations with use of alternative treatment and unsupportive pathology.
Results
Inclusion criteria were met by 56.2% (n = 3397) of those women who underwent hysterectomy (n = 6042). There was no documentation of alternative treatment prior to hysterectomy in 37.7% (n = 1281). Alternative treatment was more likely to be considered among women aged <40 years vs those aged 40-50 and >50 years (68% vs 62% vs 56%, P < .001) and among women with larger uteri. Unsupportive pathology was identified in 18.3% (n = 621). The rate of unsupportive pathology was higher among women age <40 years vs those aged 40-50 and >50 years (37.8% vs 12.0% vs 7.5%, P < .001), among women with an indication of endometriosis/pain vs uterine fibroids and/or AUB, and among women with smaller uteri.
Conclusion
This study provides evidence that alternatives to hysterectomy are underutilized in women undergoing hysterectomy for AUB, uterine fibroids, endometriosis, or pelvic pain. The rate of unsupportive pathology when hysterectomies were done for these indications was 18%.
See related editorial, page 257
Hysterectomy is the most commonly performed major gynecologic surgery in the United States. Over 400,000 hysterectomies are performed annually in the United States and it is estimated that 1 in 3 women will have had a hysterectomy by age 60 years. Of the benign hysterectomies performed in the United States, 68% are done for the primary indication of abnormal uterine bleeding (AUB), uterine leiomyomata, and endometriosis. The American Congress of Obstetricians and Gynecologists (ACOG) supports the use of alternatives to hysterectomy including hormonal management, operative hysteroscopy, endometrial ablation, and use of the levonorgestrel intrauterine device (IUD) as primary management of these conditions in many cases. Although use of these alternative treatments has recently led to a decrease in the utilization of hysterectomy, assessing the appropriateness of hysterectomy continues to be a target for quality improvement. Applying appropriateness criteria to hysterectomy, overutilization has been estimated to range from 16-70%.
Using data from the Michigan Surgical Quality Collaborative (MSQC), we sought to examine the medical records for documentation of alternative treatment prior to hysterectomy and whether the specimen pathology was supportive or unsupportive of the need for surgery.
Materials and Methods
This study was performed using a dataset from the MSQC. Funded by the Blue Cross and Blue Shield of Michigan/Blue Care Network, MSQC is a coalition of 52 academic and community hospitals across the state of Michigan that voluntarily participate in this statewide surgical quality improvement collaborative using an audit and feedback system as well as regular meetings and site visits. At each participating hospital, specially trained, dedicated nurse abstractors collect patient characteristics, intraoperative processes of care, and 30-day postoperative outcomes from general and vascular surgery and hysterectomy cases in accordance with established policies and procedures. To reduce sampling error, a standardized data collection methodology is employed that uses only the first 25 cases of an 8-day cycle (alternating on different days of the week for each cycle). The clearly defined and standardized data collection methodology is routinely validated through scheduled site visits, conference calls, and internal audits. The collection of hysterectomy-specific data (eg, parity, indication for surgery, pathology) in MSQC began on Jan 1, 2013. This study met the criteria for “exempt” status by the University of Michigan Institutional Review Board-Medical (HUM00073978).
Inclusion criteria for this study were women aged >18 years; a preoperative indication of AUB, uterine fibroids, chronic pelvic pain, and/or endometriosis; and hysterectomy performed from Jan. 1, through Nov. 8, 2013. Exclusion criteria included a preoperative indication of high-grade cervical dysplasia, endometrial hyperplasia with atypia, any cancer, pelvic mass, family history of cancer, or pelvic organ prolapse. Abstracted data included demographics, preoperative health history, preoperative nonsurgical treatments, and pathology from the hysterectomy specimens. Access to outpatient data and/or office notes was not routinely available.
The primary outcome was the use of medical management or minor procedures as treatment alternatives prior to hysterectomy. Counseling regarding the use of hormonal therapy, a levonorgestrel IUD, pain management, endometrial ablation, and/or hysteroscopy prior to hysterectomy was abstracted from the medical record. Information regarding uterine artery embolization was not documented in the database, and thus was not available for analysis. Each alternative treatment was counted as “considered” if documentation was available indicating that the patient declined, was unable to tolerate, or had an unsuccessful course of treatment.
Pathology reports were examined to determine if there were findings supporting the need for hysterectomy. Pathology findings were considered “supportive” if cervical or uterine fibroids, cervical dysplasia (cervical intraepithelial neoplasia [CIN] 2 or CIN 3), simple endometrial hyperplasia with atypia, complex endometrial hyperplasia with or without atypia, adenomyosis, endometriosis, benign pelvic masses, or cancer of any type were identified. In contrast, pathology was defined as “unsupportive” if the cervix, endometrium, and myometrium, fallopian tubes, and ovaries were all described as “normal” or “unremarkable” or were not described. Pathology was also considered “unsupportive” if only the following were used to describe the cervix or endometrium: low-grade cervical dysplasia (CIN 1), cervicitis, cervical hyperplasia, nabothian cysts, cervical or endometrial polyps, inactive/atrophic endometrium, endometritis, metaplasia, or proliferative or secretory endometrium. Specimen weights (uterine or uterine + adnexa) were abstracted from the pathology report.
All statistical analyses were performed using software (STATA, version 13; College Station, TX). The demographics, medical comorbidities, surgical details, and pathology were compared among those women who did and did not have documentation of alternative treatment using χ 2 analysis and Student t tests. Women were divided into age groups of <40, 40-50, and >50 years and χ 2 analysis and analysis of variance were used to make comparisons. A multivariable logistic regression model to evaluate what predicted use of alternative treatment was developed. Factors with a plausible relationship to the utilization of alternative treatment to hysterectomy or with a significant ( P < .05) or marginally significant ( P < .15) relationship in bivariate analysis were initially included in the model. Factors with no significant relationship were removed from the model unless there was a marginally significant relationship and there was a plausible relationship to the outcome of interest. In a similar fashion, multivariable logistic regression was developed to analyze associations between clinical factors and the finding of unsupportive pathology. A significance level of alpha = .05 was used for all bivariate analyses.
Results
From Jan. 1 through Nov. 8, 2013, data were available for 6042 hysterectomies performed at the 52 participating MSQC hospitals. Inclusion criteria were met by 56.2% (n = 3397) ( Figure 1 ). Demographics and medical comorbidities are presented in Table 1 . The primary indications for the hysterectomy were endometriosis and/or chronic pelvic pain in 9.2% (n = 311), AUB and/or uterine fibroids in 49.1% (n = 1667), and a combination of endometriosis, chronic pain, AUB, and/or fibroids in 48.1% (n = 1419). Documentation of at least 1 alternative treatment was identified in 62.3% (n = 2116). Women with a preoperative indication of uterine fibroids/AUB were less likely to receive at least 1 alternative treatment than those women with endometriosis or a combination of endometriosis/pain and uterine fibroids/AUB (58.2% vs 68.8% vs 65.7%, P < .001). Women who had documentation of any alternative treatment prior to hysterectomy were younger than those who did not (43.1 ± 7.8 vs 44.9 ± 8.5, P < .001). Hypertension was more common in those who did not use alternative treatment and white women were more likely than black women and women classified as neither white nor black to receive alternative treatment. Parity, body mass index, insurance, and other common medical comorbidities (pulmonary disease, coronary artery disease, diabetes, history of deep venous thrombosis) did not differ among the groups. Due to the difference in age observed among those with and without documentation of alternative treatment, we then analyzed cases by the following age groups: <40, 40-50, and >50 years ( Table 1 ). Women aged <40 years were more likely to have documentation of alternative treatment, more likely to be Hispanic, more likely to smoke, more likely to have a history of deep venous thromboembolism, and less likely to have hypertension, diabetes, or coronary artery disease compared to women aged 40-50 or >50 years. The age groups did not differ in body mass index or in the rates of chronic obstructive pulmonary disease or bleeding disorders.
| Variable | Alternative treatment (n = 2116) | No alternative treatment (n = 1281) | P value | Aged <40 y n = 942 | Aged 40-50 y n = 1800 | Aged >50 y n = 655 | P value |
|---|---|---|---|---|---|---|---|
| Age, y a | 43.1 ± 7.8 | 44.9 ± 8.5 | < .001 | ||||
| Parity b | 2 (1–3) | 2 (1–3) | .87 | 2 (1–3) | 2 (1–3) | 2 (1–3) | .51 |
| Race, % (n) | .01 | < .001 | |||||
| White | 72.7 (1539) | 68.1 (873) | 78.0 (735) | 67.4 (1214) | 70.7 (463) | ||
| Black | 20.7 (438) | 23.5 (301) | 16.6 (156) | 24.6 (442) | 21.5 (141) | ||
| All other | 6.6 (139) | 8.4 (107) | 5.4 (51) | 8.0 (144) | 7.8 (51) | ||
| Ethnicity, % (n) | .50 | .01 | |||||
| Non-Hispanic | 90.6 (1917) | 90.5 (1159) | 88.0 (829) | 91.3 (1643) | 92.2 (604) | ||
| Hispanic | 1.7 (35) | 2.2 (28) | 2.9 (27) | 1.5 (27) | 1.4 (9) | ||
| Unknown | 7.8 (164) | 7.3 (94) | 9.1 (86) | 7.2 (130) | 6.4 (42) | ||
| BMI a | 30.9 ± 7.4 | 30.8 ± 8.1 | .44 | 30.6 ± 7.9 | 30.8 ± 7.5 | 31.2 ± 8.5 | .36 |
| Smoker, % (n) | 26.7 (566) | 26.3 (337) | .78 | 37.5 (353) | 23.8 (428) | 18.6 (122) | < .001 |
| COPD, % (n) | 2.7 (58) | 2.3 (30) | .48 | 2.4 (23) | 2.3 (42) | 3.5 (23) | .25 |
| CAD, % (n) | 1.2 (25) | 1.8 (23) | .14 | .74 (7) | 1.3 (23) | 2.8 (18) | .003 |
| HTN, % (n) | 22.7 (480) | 26.5 (339) | .013 | 12.3 (116) | 23.3 (419) | 43.4 (284) | < .001 |
| DVT, % (n) | 2.2 (47) | 3.1 (40) | .11 | 3.6 (34) | 2.3 (41) | 1.8 (12) | .05 |
| Diabetes, % (n) | 5.3 (111) | 6.6 (85) | .09 | 3.7 (35) | 5.6 (101) | 9.2 (60) | < .001 |
| Bleeding disorder, % (n) | 1.1 (23) | 0.7 (9) | .26 | 1.3 (12) | 0.9 (16) | 0.6 (4) | .38 |
| Uterine weight, g a | 230 ± 318.8 (n = 1159) | 317.3 ± 502.9 (n = 705) | < .001 | 173.6 ± 280.3 (n = 556) | 293 ± 447.5 (n = 1012) | 328.1 ± 396.5 (n = 296) | < .001 |
| Uterine + adnexa weight, g a | 347.5 ± 497.2 (n = 547) | 243 ± 347.2 (n = 914) | < .001 | 172.4 ± 261.3 (n = 359) | 315.5 ± 433.1 (n = 153) | 324.6 ± 471.8 (n = 40) | < .001 |
| Alternative treatment, % (n) | — | — | —- | 67.6 (637) | 61.9 (1115) | 55.6 (364) | < .001 |
Alternative treatments documented in the database were also analyzed. These treatments included hormonal therapy, pain management, levonorgestrel IUD, hysteroscopy, and endometrial ablation. How frequently each alternative was considered among hysterectomy cases analyzed by age group is presented in Table 2 . The frequency with which each alternative treatment was considered and then undertaken (eg, failed or not tolerable) is illustrated in Figure 2 . The rate at which the levonorgestrel IUD was “considered” and attempted was much lower than that for hormonal therapy, hysteroscopy, or endometrial ablation. The levonorgestrel IUD was “considered” by 12% and only one third of those either failed or found it not tolerable. The total number of treatments considered among women of different age groups is provided in Figure 3 . For purposes of analyses, alternative treatments were further categorized as “medical only” if documentation of any combination of the following was found: hormonal therapy, pain management, and levonorgestrel IUD, and “minor surgery only” if only hysteroscopy and/or endometrial ablation were attempted ( Figure 4 ).
| Type of alternative treatment, % (n) | Aged <40 y n = 637 | Aged 40-50 y n = 1115 | Aged >50 y n = 364 | P value |
|---|---|---|---|---|
| Hormonal therapy | 51.5 (328) | 46.8 (522) | 40.1 (146) | .002 |
| Levonorgestrel IUD | 12.7 (81) | 12.4 (138) | 9.3 (34) | .234 |
| Endometrial ablation | 44.1 (281) | 48.3 (538) | 31.6 (115) | < .001 |
| Hysteroscopic surgery | 32.7 (208) | 33.3 (371) | 44.8 (163) | < .001 |
| Pain management | 25.4 (162) | 20.6 (230) | 16.5 (60) | .003 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
