If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Corresponding author at: Department Health Sciences, Faculty of Science & Talma Institute, Vrije Universiteit Amsterdam, De Boelelaan 1085, 1081 HV Amsterdam, the Netherlands.
Department of Health Economics, School of Business and Economics & Talma Institute, Vrije Universiteit, Amsterdam, the NetherlandsDepartment of Obstetrics and Gynaecology, Amsterdam University Medical Centers, Amsterdam, the Netherlands
Grow School of Oncology and developmental biology, Maastricht University Medical Center, Maastricht, the NetherlandsDepartment of Obstetrics and Gynaecology, Maxima Medical Center, Veldhoven, the Netherlands
Department of Health Economics, School of Business and Economics & Talma Institute, Vrije Universiteit, Amsterdam, the NetherlandsZilveren Kruis Health Insurance, Leiden, the Netherlands
Idiopathic heavy menstrual bleeding negatively affects women’s quality of life.
•
The stepped care approach shows practice variation between Dutch hospitals.
•
Women underwent 0.63 treatments (range, 0.36–1.00) before endometrial ablation.
•
Women underwent 0.96 treatments (range, 0.56–1.45) before hysterectomy.
•
Scope exists to reduce hysterectomy rates by implementing less invasive therapies.
Abstract
Introduction
Heavy menstrual bleeding (HMB) affects a quarter of all women, with half having no structural cause. Dutch guidelines recommend a stepped care approach to the management of such idiopathic HMB, starting with medication or a levonorgestrel-releasing intrauterine device (LNG-IUD), before progressing to endometrial ablation, and ultimately, hysterectomy. However, practice variation between hospitals could lead to suboptimal health outcomes and increased healthcare costs for some women.
Objectives
To evaluate adherence to stepped care for women with idiopathic HMB and to identify practice variation among Dutch hospitals.
Study design
This population-based cross-sectional study used Dutch insurance claims data from primary and secondary care for all women with idiopathic HMB referred to a gynecologist between January 2019 and December 2020. We calculated the average number of treatments in the 3 years before each treatment step at each hospital, making adjustments for age, socioeconomic status, and ethnicity. Variation in medical practice was measured by the coefficient of variation (CV).
Results
We studied 20,715 women treated with LNG-IUDs (56%), endometrial ablation (36%), laparoscopic hysterectomy (13%), or vaginal hysterectomy (4%) in 93 hospitals. Before endometrial ablation, on average 47% used medication (hospital range 27%–71%; CV 0.17) and 16% used an LNG-IUD (hospital range 8%–29%, CV 0.32). Before hysterectomy, 52% (hospital range 28%–65%, CV 0.16) used medication, 21% (hospital range 6%–38%, CV 0.35) used an LNG-IUD, and 23% underwent endometrial ablation (hospital range 0%–59%, CV 0.55). On average, women underwent 0.63 (hospital range 0.36–1.00, adjusted rate 0.40–0.98, CV 0.17) and 0.96 (hospital range 0.56–1.45, adjusted rate 0.56–1.44, CV 0.18) treatments before endometrial ablation and hysterectomy, respectively.
Conclusions
Considerable practice variation exists among Dutch hospitals in the stepped care approach to idiopathic HMB. Improving adherence to this approach could improve quality of care and reduce costs.
Heavy menstrual bleeding (HMB) is defined as excessive menstrual blood loss that interferes with a woman’s physical, social, or emotional and material quality of life.[
] The International Federation of Gynecology and Obstetrics (FIGO) has developed the PALM-COEIN classification system for the causes of abnormal uterine bleeding (AUB), where PALM refers to structural disorders (i.e., polyp; adenomyosis; leiomyoma; malignancy and hyperplasia) and COEIN refers to non-structural disorders (i.e., coagulopathy; ovulatory dysfunction; endometrial; iatrogenic and not yet classified).[
] However, Dutch guidelines recommend a stepped care approach for the treatment of idiopathic (non-structural) HMB, taking care to exclude patients with iatrogenic AUB who might use anticoagulants (i.e., patients in an AUB-COEN cohort). Consistent with other national guidelines,[
] this approach prioritizes low invasive therapies, including oral medication and levonorgestrel-releasing intrauterine device (LNG-IUD) placement, and recommends surgery if these fail. Surgical interventions then include endometrial ablation, which reduces HMB by destroying the functional endometrium, [
] Both surgical interventions incur greater risks for women and higher financial costs for society, and as such, should be reserved until earlier steps have been saturated in a stepwise manner. Several studies have assessed variations between regions or hospitals in surgical treatment for idiopathic HMB, [
] but we are aware of none focusing on practice variation in the stepped care approach. Research in this area could provide insights into how we might improve guideline adherence and reduce overtreatment for women with idiopathic HMB.
This study evaluates overall adherence by, and variation between, Dutch hospitals in the stepped care approach to managing women with idiopathic HMB according to Dutch practice guidelines.
Material and methods
Study design
We conducted a population-based cross-sectional study of Dutch insurance claims data from primary and secondary care between January 2019 and December 2020. The average number of treatments in the 3 years before each treatment step at each hospital are calculated, with adjustments for age, socioeconomic status (SES), and ethnicity.
Dutch setting and data sets
Most healthcare services in the Netherlands are covered by mandatory health insurance, including those for HMB. Women with gynecological complaints first consult their general practitioner, who only refers to gynecologists in hospitals or independent treatment centers if they cannot resolve the issue in primary care. The healthcare budget in secondary care then follows the Diagnosis Treatment Combination (DTC) codes for referral, which include information about the medical activities performed for each health complaint. Claims data submitted to Dutch health insurers can be obtained from their executive agency, Vektis, which provided national data for all care received by women in our study from January 1, 2016, to December 31, 2020. We enriched this dataset with demographic and socioeconomic data by postal code for 2018 from Statistics Netherlands. [
We enrolled all women with DTC code G11 (cycle abnormalities, including postmenopausal bleeding) recorded between January 1, 2016, and December 31, 2020. Given that this population can be heterogeneous, the following criteria were used to include only patients with idiopathic HMB according to the modified AUB-COEN classification (i.e., excluding iatrogenic AUB as a non-structural cause).
We only included women treated by LNG-IUD, endometrial ablation, or hysterectomy for HMB. Other diagnoses requiring these treatments should take another DTC code, except postmenopausal bleeding. We mitigated this risk by only enrolling women younger than 55 years (menopause normally occurs between 45 and 55 years old), because although rare, older women are more likely to undergo hysterectomy for postmenopausal bleeding. [
] Additionally, we used a lower age limit of 30 years because pregnancy is contraindicated after endometrial ablation.
Women who met the following criteria were excluded: those with DTC codes G15 (uterine fibroids) or G18 (contraception) in addition to G11, to limit those with HMB due to anatomic causes (e.g., intracavitary polyps or myomas) or using LNG-IUD for contraception; those with a history of breast cancer or venous thrombosis who may be unable to use hormonal medication or an LNG-IUD; and those who underwent therapeutic hysteroscopy, to exclude women with HMB due to intracavitary polyps or myomas. Given that existing guidelines recommend vaginal or laparoscopic hysterectomy for idiopathic HMB, we further excluded patients who underwent abdominal hysterectomy to minimize the risk of incorrectly including women with fibroids (though DTC G15 should be registered in these cases). [
The American College of Obstetricians and Gynecologists. Choosing the Route of Hysterectomy for Benign Disease 2017 [updated 2021]. Available from: https://www.acog.org/clinical/clinical-guidance/committee-opinion/articles/2017/06/choosing-the-route-of-hysterectomy-for-benign-disease.
] To ensure complete retrospective treatment information, we also excluded women with only LNG-IUD, endometrial ablation, or hysterectomy recorded before January 1st, 2019. By using anticoagulant medication to identify woman with a history of venous thrombosis, most cases of iatrogenic HMB were excluded. Appendix A provides a complete list of medications, DTC codes, and healthcare activities used to identify participants.
Outcomes
We developed three primary and seven secondary indicators, as defined in Table 1, using the healthcare activity codes detailed in Appendix B. Analyses were based on the index treatment, which we defined as the most recent LNG-IUD insertion, endometrial ablation, or hysterectomy for idiopathic HMB between January 1, 2019, and December 31, 2020. The three primary outcomes show the average number of treatments for the period (nationwide and by hospital) before the three index treatments, while the seven secondary outcomes provide more detailed insights into the number of treatments performed before the index treatments. Finally, we report the hospital range to demonstrate the spread of practice variation among hospitals. This describes the range between the hospitals with the lowest and highest number of treatments per woman for a given index treatment.
Table 1Primary and secondary indicators.
MAIN INDICATORS
Index treatment
Nominator
Denominator
Outcome/Indicator
LNG-IUD
Number of treatments (medication) per woman prior to LNG-IUD insertion
Number of women with an LNG-IUD insertion
Indicator 1: Average number of treatments per woman prior to LNG-IUD insertion
Endometrial ablation
Number of treatment (medication or LNG-IUD) per woman prior to endometrial ablation
Number of women who underwent endometrial ablation
Indicator 2: Average number of treatments per woman prior to endometrial ablation
Hysterectomy
Number of treatments (medication/LNG-IUD/endometrial ablation) per woman prior to hysterectomy
Number of women who underwent a hysterectomy
Indicator 3: Average number of treatments per woman prior to hysterectomy
SECONDARY INDICATORS
Index treatment
Nominator
Denominator
Outcome/Indicator
Endometrial ablation
Number of women without treatment prior to endometrial ablation
Number of women who underwent endometrial ablation
Indicator 4: % of women without treatment prior to endometrial ablation
Endometrial ablation
Number of women who underwent only one treatment (medication or LNG-IUD) prior to endometrial ablation
Number of women who underwent endometrial ablation
Indicator 5: % of women with one treatment prior to endometrial ablation
Endometrial ablation
Number of women that used medication and had an LNG-IUD inserted prior to endometrial ablation
Number of women who underwent endometrial ablation
Indicator 6: % of women with two different treatments prior to endometrial ablation
Hysterectomy
Number of women without any treatment prior to hysterectomy
Number of women who underwent a hysterectomy
Indicator 7: % of women without treatment prior to hysterectomy
Hysterectomy
Number of women who underwent only one treatment (medication/LNG-IUD/endometrial ablation) prior to hysterectomy
Number of women who underwent a hysterectomy
Indicator 8: % of women with one treatment prior to hysterectomy
Hysterectomy
Number of women who underwent two different treatments (medication/LNG-IUD/endometrial ablation) prior to hysterectomy
Number of women who underwent a hysterectomy
Indicator 9: % of women with two treatments prior to hysterectomy
Hysterectomy
Number of women that underwent three different treatments (medication/LNG-IUD/endometrial ablation) prior to hysterectomy
Number of women who underwent a hysterectomy
Indicator 10: % of women with three treatments prior to hysterectomy
The characteristics of the included women and treating hospitals are reported for the whole study sample and separately for each index treatment. Corrections were also performed for age, SES, and ethnicity.
It is widely established that the risk of thrombosis increases with age and body mass index among women using oral contraceptives. Consequently, these women may be less likely to opt for an oral contraceptive. [
] Although the Dutch guideline does not advocate deviation in these cases, women may have preferences that differ from those of the average population after considering the pros and cons. We corrected for age with five consecutive 5-year periods from 30 to 54 years, but we could not directly control for obesity because our dataset contained no information on body mass index. SES, based on three average household income percentiles (low, 0–40; medium, 41–79; high, ≥80), was used as a proxy of obesity due to the correlation between lower SES and obesity. We also corrected for ethnicity because of its association with uterine myomas, using the percentage of non-Western immigrants per postal code as a proxy. Although we excluded women with a record of DTC G15 (uterine fibroids), registration bias could have resulted in DTC code G11 instead of G15 for those who underwent hysterectomy for myomas. [
] Finally, to adjust for possible bias due to the COVID-19 pandemic, we included a variable for whether the index treatment took place in 2019 or 2020.
Ten (ordered) logistic models were used to estimate the number of treatments based on the case-mix characteristics, allowing calculation of the weighted numbers of expected treatments before each index treatment and assessment of the proportional odds assumption. To estimate the extent of practice variation between hospitals for all treatments, and to compare practice variation for index treatments, we calculated the coefficient of variation (CV) for each indicator as a descriptive measure of heterogeneity. For this, we divided the standard deviation (SD) by the mean.
As sensitivity analyses, we repeated the descriptive and statistical analyses for idiopathic HMB treatments performed in 2- and 4-year periods before the index treatment and calculated the outcomes per year (2019 and 2020) to assess the impact of COVID-19. The analyses were also repeated separately for women aged 30–45 and 45–54 years. Given that pregnancy is contraindicated after endometrial ablation or hysterectomy, we expected the data to be skewed to the end of reproductive age. [
We used SAS (version 9.4; SAS Institute Inc., Cary, NC, USA) for data storage and analysis.
Ethical approval
The compliance officers of Vektis approved the study, but to reduce the risk of traceability, they withheld permission to publish data for indicators associated with less than 20 women per hospital. The study was exempt from institutional review board approval because data were gathered retrospectively and analyzed anonymously.
Results
Participants
Of the 418,513 women with DTC code G11 registered between 2016 and 2020, we finally enrolled 20,715 (Fig. 1). Table 2 shows the baseline descriptive data for the study population by index treatment, and Table 3 specifies the type of medication used before that treatment.
Fig. 1Study flow diagram The Figure shows the number of women included and excluded per eligibility criterion.
Non-western inhabitants in postal code of residence (proxy for ethnicity), %
12
11
11
*The index treatment refers to the most recent treatment for HMB between January 1st, 2019, and December 31st, 2020, including LNG-IUD placement, endometrial ablation, and hysterectomy. ** Laparoscopic hysterectomy = 2041; Vaginal hysterectomy = 749. Abbreviations: CV, coefficient of variation; LNG-IUD, levonorgestrel-releasing intrauterine device; SD, standard deviation.
Table 3Specification of medication before index treatments.
Medication before index treatment (generic name)*
Trade names
ATC code
n
Ethinylestradiol
L02AA03
0
Ethinylestradiol & Desogestrel
Marvelon, Mercilon
G03AA09
217
Ethinylestradiol & Drospirenone
Yasmin
G03AA12
327
Ethinylestradiol & Gestodene
Minulet
G03AA10
56
Ethinylestradiol & Levonorgestrel
Lovette, Stediril 30, Microgynon 20, 30 & 50
G03AA07
3214
Ethinylestradiol & Norgestimaat
Cilest
G03AA11
20
Levonorgestrel
G03AC03
0
Levonorgestrel
G03AD01
10
Medroxyprogesterone
Provera
G03DA02
1343
Norethisteron
Primolut
G03DC02
1837
Tranexamic acid
Cyklokapron
B02AA02
4195
*The index treatment refers to the most recent treatment for HMB between January 1st, 2019, and December 31st, 2020, including LNG-IUD placement, endometrial ablation, and hysterectomy.
Abbreviations: ATC, Anatomical Therapeutic Chemical (code).
Case-mix correction was possible for 11,567 of the 11,635 women who underwent LNG-IUD insertion in 78 hospitals (Fig. 2). The average number of medications (including tranexamic acid) used before LNG-IUD insertion was 0.33 per woman (hospital range 0.14–0.64; adjusted, 0.15–0.67; CV, 0.24).
Fig. 2Case-mix adjusted proportion of women who used medication in the 3 years prior to LNG-IUD placement. Hospital number is the arbitrary number assigned to each hospital. Case-mix adjusted proportion is ordered from lowest to highest. Purple horizontal bar shows the average among hospitals. Women could have received no treatment and/or medication (hormonal or tranexamic acid) before LNG-IUD placement. Case-mix adjustment was performed for age, social-economic status and ethnicity. Abbreviations: LNG-IUD, levonorgestrel-releasing intrauterine device. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Case-mix correction was possible for 7,467 of the 7,514 women who underwent endometrial ablation in 60 hospitals (Fig. 3). The average number of treatments (i.e., medication or IUD insertion) before endometrial ablation was 0.63 per woman (hospital range, 0.36–1.00; adjusted, 0.40–0.98; CV, 0.17). Concerning the secondary indicators, 46% of women received no treatment before ablation (hospital range, 21%–64%; adjusted, 22%–61%; CV, 0.17), 45% received one treatment (hospital range, 32%–64%; adjusted, 32%–63%; CV, 0.14), and 9% received two treatments (hospital range, 0%–21%; adjusted, 1%–20%; CV, 0.41).
Fig. 3Case-mix adjusted average number of treatments in the 3 years before endometrial ablation among the treating hospitals. Hospital number is the arbitrary number assigned to each hospital. Case-mix adjusted average number of treatments is ordered from lowest to highest. Purple horizontal bar shows the average number of treatments. Women could have received no treatment, medication (hormonal or tranexamic acid), and/or LNG-IUD placement before endometrial ablation. Case-mix adjustment was performed for age, social-economic status and ethnicity. Abbreviations: LNG-IUD, levonorgestrel-releasing intrauterine device. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Of the 2,790 women who underwent hysterectomy (laparoscopic, 2041; vaginal, 749), case-mix correction was possible for 2,777 women treated in 55 hospitals (Fig. 4). The average number of treatments (i.e., medication, LNG-IUD, or ablation) was 0.96 per woman (hospital range 0.56–1.45, adjusted rate 0.56–1.44, CV 0.18) in the three years before hysterectomy. Concerning the secondary indicators, 33% of women received no treatment before hysterectomy (hospital range, 10%–56%; adjusted, 11%–56%; CV, 0.24), 42% received one treatment (hospital range, 27%–68%; adjusted, 27%–68%; CV, 0.20), 22% received two treatments (hospital range, 4%–38%; adjusted, 4%–38%; CV 0.36), and 4% received three treatments (hospital range 0%–13%; adjusted, 0%–13%; CV, 0.93).
Fig. 4Case-mix adjusted average number of treatments in the 3 years before laparoscopic or vaginal hysterectomy among the treating hospitals. Hospital number is the arbitrary number assigned to each hospital. Case-mix adjusted average number of treatments is ordered from lowest to highest. Purple horizontal bar shows the average number of treatments. Women could have received no treatment, medication (hormonal or tranexamic acid), LNG-IUD placement, and/or endometrial ablation before hysterectomy. Case-mix adjustment was performed for age, social-economic status and ethnicity. Abbreviations: LNG-IUD, levonorgestrel-releasing intrauterine device. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
As shown in Appendix C, the sensitivity analyses produced results consistent with the main analyses.
Discussion
Our analyses reveal that guideline adherence can be improved for women who undergo surgery for idiopathic HMB. Significant percentages of women aged 30–55 years who underwent endometrial ablation (46%) or hysterectomy (33%) did not receive another treatment in the 3 years before surgery. The largest variation was observed for women undergoing endometrial ablation before hysterectomy.
Existing literature on practice variation in the management of idiopathic HMB has partially addressed our research question, showing mixed results on guideline adherence and regional variation in the use of endometrial ablation and hysterectomy. [
] In a smaller historical cohort of 16 women, Ali et al. found that only 25% had received an LNG-IUD and that none underwent endometrial ablation before hysterectomy. [
] Our results are consistent with these, suggesting that women may receive suboptimal care for idiopathic HMB due to consistent failures to adopt the least invasive treatment first.
To improve healthcare quality and reduce costs, we must first identify overtreatment. Studies of practice variation in the management of HMB aim to tackle this by identifying where unwarranted practice variation exists. According to Wennberg, medical practice variation is unwarranted if it cannot be explained by the characteristics or preferences of women. [
] At the level of the individual patient, a woman might make an informed choice to deviate from the stepped-care model (e.g., opting not to use hormonal treatments). This cannot be considered either overtreatment or unwarranted practice variation because it is her right to choose. At the level of the hospital, it is also possible that we could find significantly more patients deviating from the stepped care model in a given center. In these cases, it should be possible to assess critically whether patient preference or suboptimal implementation of the stepped-care model could explain the observed differences. Finding and correcting the reasons for unwarranted practice variation in hospitals with suboptimal implementation could improve care for patients with HMB. For example, clinical decision aids could help to harmonize counseling for women with idiopathic HMB, making their preference integral to the decision-making process.
Studies into medical practice for idiopathic HMB often lack a clinical perspective because the non-invasive treatments before surgery are not known. [
] The performance indicators used in our study provide insights into adherence to the stepped care model and could be used as an audit and feedback instrument to reduce overtreatment. [
] An important limitation of such approaches is the requirement for extensive data acquisition, with the feasibility of implementing performance indicators negatively associated with its anticipated administrative burden. [
Keizer, de N.F., Maas, A.I.R., Menon, D., Ercole, A., Citerio, G., Lecky, F., Wilson, L., Cnossen, M.C., Polinder, S., Steyerberg, E.W., Jagt, van der M., Lingsma, H.F.
Development of a quality indicator set to measure and improve quality of ICU care for patients with traumatic brain injury.
] However, Vektis could provide claims data for all Dutch inhabitants treated in primary care, secondary care, and independent pharmacies, negating this issue.
Several limitations also warrant consideration. First, by capturing women with cycle irregularities, cases without idiopathic HMB may have been included. We reduced this heterogeneity by selecting women who underwent one of the three treatments for idiopathic HMB and by excluding women who underwent treatment for anatomic HMB (e.g., therapeutic hysteroscopy). Although some women could have undergone such treatment for other reasons, others may also have been excluded inadvertently. Furthermore, we cannot rule out the possibility of including women with endometrial hyperplasia, although we expect both the prevalence to be low in our population and for most of these patients to have DTC M14 (endometrial malignancy) registered.
Second, because the validity of our analyses depends on the specificity of the claims data, potential issues could arise from coding for endometrial ablation and LNG-IUD placement. For example, although the code for endometrial ablation (037177) can also be used for hysteroscopic removal of intracavitary adhesions, this procedure is rarely performed in the Netherlands and should not affect our findings. [
] Additionally, although the claim code for LNG-IUD placement (037180) can be used for copper IUDs, the latter are only used for contraception and require the use of another DTC code (G18).
Third, a trend for continued treatment has recently been observed after as many as 4 years of LNG-IUD placement (i.e., the longest period in our sensitivity analysis). This might indicate an underestimation of treatments given before the index treatment. [
Despite clear guidelines for the stepped care of idiopathic HMB, many women received no other treatments in the 3 years before endometrial ablation (46%) or hysterectomy (33%). Most variation between hospitals occurred in the implementation of endometrial ablation before hysterectomy (0%–59%). Improving adherence to the stepped-care approach, while taking care to consider patient preference, could improve care quality and reduce costs. Research should evaluate how implementing an audit cycle using claims-based quality indicators can help achieve this goal.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
We would like to thank Michiel ten Hove, Tijs van Gorp and Robin Stoof for their contribution to the data request to Vektis and Jordy Breet for the execution of the pilot study. Dr. Robert Sykes (Doctored Limited, www.doctored.org.uk) provided technical editing assistance for the final drafts of this manuscript.
Author contribution statement
MV, FP, BM, and MB designed the study. MV, EH were involved in the compliance procedure to retrieve the data from Vektis. TW analyzed the data and XK provided feedback to the statistical analyses performed. MV and FP drafted the manuscript. All authors reviewed and approved the final manuscript.
Funding information
None declared.
References
National Institute for Health and Care Excellence.2020;
The American College of Obstetricians and Gynecologists. Choosing the Route of Hysterectomy for Benign Disease 2017 [updated 2021]. Available from: https://www.acog.org/clinical/clinical-guidance/committee-opinion/articles/2017/06/choosing-the-route-of-hysterectomy-for-benign-disease.
Keizer, de N.F., Maas, A.I.R., Menon, D., Ercole, A., Citerio, G., Lecky, F., Wilson, L., Cnossen, M.C., Polinder, S., Steyerberg, E.W., Jagt, van der M., Lingsma, H.F.
Development of a quality indicator set to measure and improve quality of ICU care for patients with traumatic brain injury.
00043-X&id=gr1.jpg){kind=link}
00043-X&id=gr2.jpg){kind=link}
00043-X&id=gr3.jpg){kind=link}
00043-X&id=gr4.jpg){kind=link}