Abstract
Objective
The incidence and mortality rates of endometrial cancer are increasing globally, including in Japan. Quality of cancer care is promoted through guideline adherence. This study aimed to establish quality indicators (QIs) for endometrial cancer and explore the factors contributing to treatment nonadherence.
Methods
QIs and pattern-of-care indicators (PCIs) were developed using the Research and Development/University of California Los Angeles modified Delphi method. QIs reflect desirable healthcare patterns, whereas PCIs address treatment areas with lacking evidence. Data from the Hospital-Based Cancer Registry and Diagnosis Procedure Combination Survey were used. Patients diagnosed or treated between January 1 and December 31, 2020 were included. The reasons for nonadherence were collected. Logistic regression was used to analyze the factors influencing adherence, including age, body mass index, comorbidities, facilities, and recurrence risk.
Results
Of the 35 proposed QI candidates, 8 QIs and 9 PCIs were selected, predominantly focusing on surgical aspects. Adherence rates varied, with peritoneal lavage cytology being the highest (93.1%), and postoperative hormone replacement therapy (HRT) for patients aged <45 years being the lowest (30.9%), when focusing on process indicators. Reasons for nonadherence included patient preference and medical comorbidities as significant factors. Multivariate analysis highlighted age, clinical stage, and Barthel index as significant contributors to nonadherence.
Conclusion
We developed QIs to comprehensively assess endometrial cancer treatment. Adherence rates are favorable; however, HRT has a low adherence rate. Factors leading to nonadherence include advanced age and incomplete activities of daily living, particularly in advanced stages.
Keywords: Guideline Adherence, Endometrial Neoplasms, Process Assessment, Health Care
Synopsis
We developed quality indicators for endometrial cancer. Postoperative hormone replacement therapy has the lowest adherence rate among indicators. Aging, advanced stage, and low Barthel index contributed significantly to nonadherence across all stages and patient preference and comorbidities are the main reasons for nonadherence in advanced stages.
INTRODUCTION
The incidence of endometrial cancer is continuously increasing in conjunction with a declining birth rate [1]. Globally, the rate of new uterine cancer cases increased by 0.6% annually from 2010 to 2019 [2]. This upward trend has also been observed in Japan, where the incidence rates increased from 15.5 per 100,000 population in 2009 to 27.6 per 100,000 population in 2019 [3,4]. Concurrently, the mortality rate of endometrial cancer continues to increase, with a global average increase of 1.7% per year from 2010 to 2019 [2]. Similarly, an increasing trend was observed in Japan, rising from 2.9 per 100,000 population in 2010 to 4.3 per 100,000 population in 2019 [5].
Treatment options for endometrial cancer, such as radiation therapy and chemotherapy, have shown lower efficacy than those for cervical and ovarian cancers. Surgery has long been the mainstay for endometrial cancer, offering high cure rates in the early stages but with higher mortality rates in advanced stages [6].
Focusing on cancer types characterized by high incidence rates, the standardization and enhancement of overall cancer care quality are being promoted by assessing adherence rates to various guidelines. In the field of gynecology, the guidelines established by European Society of Gynaecological Oncology (ESGO)/European Society for Radiotherapy and Oncology/European Society of Pathology address cervical [7,8], endometrial [9], and ovarian cancers [10], specifying treatment protocols for each cancer. In Japan, quality indicators (QIs) for cervical cancer were formulated in 2013 [11] and revised in 2022, whereas QIs for ovarian cancer were established in 2018 [12]. In the process of developing QIs, we identified pattern-of-care indicators (PCIs). QI encompasses healthcare patterns considered desirable for adherence. For cases where these practices are not implemented, facilities identify the reasons for nonadherence in each case and calculate predicted adherence rates to elucidate the current status of healthcare as comprehensively as possible. PCI pertains to the treatment content of medical policies that lack established evidence or present challenges in selecting a single treatment approach. PCIs are based on values for the most widely adopted healthcare patterns, to explore other treatment approaches.
Previous research has identified the characteristics of patient populations with low adherence to clinical practice guidelines in the management of endometrial cancer. In Germany, age was found to significantly contribute to decreased adherence [13]. Conversely, in the United States, a multi-ethnic nation with a non-universal healthcare system, previous studies have revealed lower adherence among racial groups such as Black, Hawaiian or the Pacific Islanders, Latinx, and Alaskan native populations, as well as individuals with low socioeconomic status [14]. Additionally, a correlation has been established between adherence and prognosis, whether measured by progression-free survival or overall survival (OS) [14].
In Japan, efforts to standardize the quality of care for endometrial cancer were undertaken in 2021. Members of the Guideline Development Committee established QIs for aspects deemed crucial within the guidelines and measurable using our dataset. This study aimed to introduce the defined QIs and explore the reasons for nonadherence. By aligning the reasons for nonadherence from prior studies with those gathered in our project, we highlight the risk factors for patients prone to nonadherence.
MATERIALS AND METHODS
1. Development of QIs
To establish QIs, the Research and Development/University of California Los Angeles modified Delphi method was used [15]. In this method, the committee members initially proposed candidate indicators for assessing the quality of care for endometrial cancer based on guideline recommendations. The members then assessed these candidate indicators for their validity and measurement requirements, rating them on a scale of 1 to 9. After the initial ratings, the members discussed and modified the indicators. The second round of ratings followed this discussion. Indicators were considered valid if their median ratings for both “validity” and “measurement needs” were ≥7. Using the rating of QI candidates, if an indicator had a median rating of ≥7 for “measurement needs” but <6 for “validity,” it was categorized as a PCI.
2. Data source for the adherence of QI/PCI
We implemented the developed indicators in a database that combined the national database of the Hospital-Based Cancer Registry (HBCR) [16] and health service utilization data from the Diagnosis Procedure Combination (DPC) survey. The HBCR is a database that collects basic information on new cancer cases at hospitals, as cancer care hospitals designated by the Ministry of Health are required to maintain a registry. Moreover, HBCR operates in non-designated hospitals. The HBCR data are submitted to the National Cancer Center once a year and compiled into a national database. The HBCR data contain patient demographics and clinical cancer information, such as clinical and pathological stages, Union for International Cancer Control (UICC) tumor-node-metastasis classifications, tumor location, and histopathological findings based on the International Classification of Diseases Oncology third edition (ICD-O-3). The DPC survey data contain information on health services provided that were equivalent to fee-for-service health insurance claims. Hospitals that participated in the HBCR were invited to participate as part of a national quality monitoring program. Cases obtained from facilities registered in the HBCR account for approximately 70% of all cases in Japan and DPC data were contributed by approximately 80% of the facilities registered in the HBCR. Therefore, this dataset includes approximately 50% of all cases in Japan [17].
3. Patient selection
Patients registered from January 1, 2020 to December 31, 2020 in the HBCR and diagnosed or treated within the registered facilities were selected through the DPC system, which covers the period from October 1, 2019 to March 31, 2022. Patients with recurrence and those without treatment administered were excluded. The data specifically targeted cases with localization coded as C54 and a histological type corresponding to endometrial cancer according to ICD-O-3.2, focusing solely on malignant (/3) cases. The following histological types of endometrial cancers were included: endometrioid carcinoma (8263, 8380, 8382, 8383, and 8570), clear cell carcinoma (8310), serous carcinoma (8441 and 8480), carcinosarcoma (8980), and other (8000, 8001, 8005, 8010, 8020, 8054, 8140, 8265, 8323, 8260, and 8981). Stages were determined based on either the pathological stage for cases with surgical intervention or the clinical stage defined by International Federation of Gynecology and Obstetrics 2008 for cases without surgical intervention, referred to as the “Integrated Stage.” Since vascular invasion status cannot be captured in this dataset, the postoperative recurrence risk classification was defined as follows: low recurrence risk group included T1aN0M0 for Grade 1/Grade 2 endometrioid adenocarcinoma (G1/G2); intermediate recurrence risk group included T1bN0M0 for G1/G2 and T1aN0M0 for Grade 3 endometrioid adenocarcinoma (G3); and high recurrence risk group included T1bN0M0 for serous and clear cell carcinoma, T2-3N1M1 for all histological types, and carcinosarcoma in any stage.
4. Reasons for nonadherence
The analysis results were communicated to participating hospitals, who were requested to provide feedback on the reasons for nonadherence to the QIs (PCI was not collected because of its nature). For privacy protection, categories with fewer than ten patients were not reported. Clinically valid reasons were categorized as “sufficient reasons,” whereas those that were invalid were labeled as “insufficient reasons.” Cases judged as “sufficient reasons” implied that standard treatment was performed if those reasons were absent or that treatment was performed but not recorded due to data limitations. Therefore, the adjusted adherence rate was calculated using the following formula: Observed Adherence Rate + (100 − Observed Adherence Rate) × (Percentage of “Sufficient Reasons”).
5. Analysis of characteristics of nonadherence patients
To analyze the factors contributing to non-adherence, logistic regression analysis was conducted adjusting for age, integrated stage, body mass index (BMI), Barthel index (BI), comorbidities, and facility types (designated cancer care hospital/others). Although general conditions of cancer patients are usually measured by performance status (PS), they were not included in this dataset, for which we used BI, an activity of daily living (ADL) measure, as a proxy. Patient selection are shown in the flowchart in Fig. 1. Because information of patients such as comorbidity, BMI, and BI was recorded as part of discharge summary, those who was not hospitalized during the study period did not have the information, and thus, excluded from the analysis (n=350, 2.7%). Additionally, 533 patients whose stage and/or BI were missing were excluded. Cases without comorbidity input, excluding outpatient cases, amounted to 2,182. Generally, the number of comorbidities tends to correlate with the age of patients with cancer [18]. These patients without recorded comorbidities were treated as having no comorbidities. Owing to suspected data errors in BMI, cases with values beyond the 99th percentile were subsequently excluded from the analysis. Comorbidities were converted to age-adjusted Charlson Comorbidity Index (ACCI) scores for each patient by summing the individual comorbidities. Facility attributes were classified based on whether they corresponded to cancer care hospitals designated by the Ministry of Health, Labor, and Welfare (Designated Cancer Care Hospitals). Age and BMI were treated as continuous variables, whereas stage, recurrence risk and facility attributes were treated as categorical variables. Data entry in discharge summary is manually performed by individual healthcare workers, which increases the risk of missing values and reduces reliability [19]. Because of data reliability, we assumed that the BI was not suitable for evaluation as a continuous value and was converted into a binary variable (0 for 100 points, 1 for <100 points); ACCI was also converted into a set of binary variable (0 for <3 points for stages I–III, 0 for <7 points for stage IV, and 1 for higher points), because existing literature has demonstrated that cases with an ACCI of ≥3 are associated with poorer prognosis compared to cases with fewer points mainly in early stages [20].
Fig. 1. Patient selection.
ADL, activity of daily living; BMI, body mass index.
6. Ethical considerations
The study protocol was approved by the Institutional Review Board of the National Cancer Center of Japan (approval No. 2013-081). Following the Research Ethics Guidelines, we provided the patients with the opportunity to opt-out from the study by posting the information of this study on the institution’s website.
RESULTS
1. Patient backgrounds
Table 1 presents the patient demographics of the cases in 2020 that were considered for the evaluation of the indicators. The mean age of the patients was 60.4 years (standard deviation [SD]=12.7). The histological types were as follows: endometrioid carcinoma, 81.4% (10,672/13,119); clear cell carcinoma, 2.2% (294/13,119); and serous carcinoma, 5.7% (749/13,119). The mean BMI was 24.9 (SD=5.8). The mean ACCI for all cases was 2.8 (SD=1.3). The mean BI was 97.4 (SD=11.8), 89.3% (11,709/13,119) having 100 points.
Table 1. Demographic data of patients with endometrial cancer in the Hospital-Based Cancer Registry in 2020.
| Variables | Values | |
|---|---|---|
| Age (yr) | 60.4±12.7 | |
| Integrated stage | ||
| I | 9,379 (71.5) | |
| II | 782 (5.9) | |
| III | 1,600 (12.2) | |
| IV | 1,018 (7.8) | |
| Unknown | 340 (2.6) | |
| Postoperative recurrence risk | ||
| Low risk | 5,556 (42.4) | |
| Intermediate risk | 1,666 (12.7) | |
| High risk | 3,128 (23.8) | |
| Unknown | 2,769 (21.1) | |
| Histological types | ||
| Endometrioid carcinoma | 10,672 (81.4) | |
| Clear cell carcinoma | 294 (2.2) | |
| Serous carcinoma | 749 (5.7) | |
| Carcinosarcoma | 639 (4.9) | |
| Others | 765 (5.8) | |
| Barthel index | 97.4±11.8 | |
| 100 points | 11,709 (89.3) | |
| Lower than 100 points | 867 (6.6) | |
| Unknown | 543 (4.1) | |
| ACCI | 2.8±1.3 | |
| Lower than 4 in stage I–III or 8 in stage IV | 10,259 (78.2) | |
| Higher than 5 in stage I–III or 9 in stage IV | 337 (2.6) | |
| Unknown | 2,532 (19.2) | |
| Comorbidities | ||
| MI, CHF | 26 (0.2) | |
| Cerebrovascular disease | 2 (<0.1) | |
| Dementia | 89 (0.7) | |
| DM without complications | 1,474 (11.2) | |
| DM with complications | 840 (6.4) | |
| HD, CKD | 68 (0.5) | |
| BMI (kg/m2) | 24.9±5.8 | |
| Unknown | 355 (2.7) | |
| First treatment | ||
| Surgery | 4,062 (31.0) | |
| Radiotherapy | 209 (1.6) | |
| Chemotherapy | 989 (7.5) | |
| Surgery + Chemotherapy | 4,042 (30.8) | |
| Radiotherapy + Chemotherapy | 46 (0.4) | |
| Others | 3,773 (28.8) | |
| Facility attributes | ||
| Designated hospitals | 11,265 (85.9) | |
| Non-designated hospitals | 1,854 (14.1) | |
Values are presented as number (%) or mean ± standard deviation.
ACCI, age-adjusted Charlson Comorbidity Index; BMI, body mass index; CHF, congestive heart failure; CKD, chronic kidney disease; DM, diabetes mellitus; HD, hemodialysis; MI, myocardial infarction.
2. QIs developed
From the 35 QI candidates suggested by the specialist panel, 8 QIs and 9 PCIs were selected (Table 2). There are 10 indicators related to surgery out of a total of 17 indicators. Items related to the evaluation of the clinical stage before surgery consisted of 2 indicators (ut6, ut9x); postoperative complications consisted of 2 indicators (ut7, ut8); surgery for advanced stages consisted of 3 indicators (ut2, ut2x, ut3x); and perioperative treatment consisted of 5 items (ut3, ut4, ut4x, ut5x, ut6x). In 2020, the top 2 indicators with high adherence rates were undesirable procedures, with complications after laparoscopic surgery at 0.8% (26/3,309, 95% confidence interval [CI]=0.5–1.1) and complications after laparotomy at 1.8% (230/7,245; 95% CI=1.5–2.1). The lowest indicator was postoperative hormone replacement therapy (HRT) for patients under 45 years old, with a rate of 30.9% (244/789; 95% CI=27.7–34.3).
Table 2. Adherence rates for QIs and PCIs.
| Indicator categories & numbers | Indicators | No. of target patients | Adherence rate (%) | |||
|---|---|---|---|---|---|---|
| Target patients | Patients who had specified care | |||||
| QI | ||||||
| Process | ||||||
| ut1 | Peritoneal lavage cytology for endometrial patients with cancer | 10,552 | 93.1 | |||
| Patients with endometrial cancer who underwent surgery for malignant uterine tumors | Intraoperative peritoneal lavage or ascitic fluid cytology during surgery | |||||
| ut2 | Surgery for cStage III endometrial cancer | 1,100 | 87.5 | |||
| Patients with cStage III endometrial cancer | Patients who underwent total hysterectomy or surgery for malignant uterine tumors | |||||
| ut3 | Adjuvant therapy for intermediate recurrent risk in endometrial cancer | 1,511 | 63.6 | |||
| Patients with endometrial cancer patients with intermediate risk of recurrence and underwent surgery for malignant uterine tumors | Patients who received adjuvant chemotherapy or radiotherapy | |||||
| ut4 | Adjuvant chemotherapy for high recurrent risk in endometrial cancer | 2,389 | 78.5 | |||
| Patients with endometrial cancer with high risk of recurrence and underwent surgery for malignant uterine tumors (excluding M1 cases) | Patients who underwent adjuvant chemotherapy | |||||
| ut5 | Postoperative HRT for patients with endometrial cancer (aged 45 and under) | 789 | 30.9 | |||
| Patients with endometrial cancer aged 45 and under who underwent surgery for malignant uterine tumors in pStage I or II | Postoperative hormone replacement therapy within 1 yr | |||||
| Outcome | ||||||
| ut6 | Postoperative stage of patients with endometrial cancer who underwent laparoscopic surgery* | 3,162 | 7.2 | |||
| cStage I endometrial cancer patients with laparoscopic/robotic surgery | Patients with pStage II or higher | |||||
| ut7 | Complications following laparoscopic surgery for endometrial cancer* | 3,309 | 0.8 | |||
| Patients with endometrial cancer who underwent laparoscopic/robotic surgery | Patients who underwent surgery for complications within 30 days postoperatively | |||||
| ut8 | Complications following open surgery for endometrial cancer* | 7,245 | 1.8 | |||
| Patients with endometrial cancer who underwent surgery for malignant uterine tumors (open only) | Patients who underwent surgery for complications within 30 days postoperatively | |||||
| PCI | ||||||
| Process | ||||||
| ut1x | Laparoscopic surgery for cStage IA endometrial cancer | 6,142 | 48.8 | |||
| Patients with cStage IA endometrial cancer who underwent surgery for malignant uterine tumors | Patients with endometrial cancer who underwent laparoscopic/robotic surgery | |||||
| ut2x | Surgery in patients with cStage IV endometrial cancer | 1,005 | 50.4 | |||
| Patients with cStage IV endometrial cancer | Patients who underwent total hysterectomy or surgery for malignant uterine tumors | |||||
| ut3x | Surgery in patients with cStage IVB endometrial cancer | 945 | 34.6 | |||
| Patients with cStage IVB endometrial cancer | Patients who underwent total hysterectomy or surgery for malignant uterine tumors without chemo/radiotherapy prior to surgery | |||||
| ut4x | Neoadjuvant chemotherapy in patients with cStage IV endometrial cancer | 1,005 | 11.2 | |||
| Patients with cStage IV endometrial cancer | Patients who underwent neoadjuvant chemotherapy prior to total hysterectomy or surgery for malignant uterine tumors | |||||
| ut5x | Selection of adjuvant chemotherapy regimen for endometrial cancer | 3,848 | 10.6 | |||
| Patient who received adjuvant chemotherapy after surgery for malignant uterine tumors (excluding M1) | Adriamycin and cisplatin as adjuvant chemotherapy | |||||
| Outcome | ||||||
| ut6x | Adjuvant radiotherapy for high-risk recurrent endometrial cancer | 2,389 | 1.4 | |||
| Patient with high risk of recurrence endometrial cancer who underwent surgery for malignant uterine tumors (excluding M1) | Adjuvant radiotherapy | |||||
| ut7x | Surgery for malignant uterine tumors after luteal hormone therapy | 154 | 7.1 | |||
| Patients with stage IA endometrial cancer who received luteal hormone therapy without undergoing hysterectomy | Patient who underwent total hysterectomy within 5 months to 1 yr after starting luteal hormone therapy | |||||
| ut8x | Interval for cytological examinations after surgery for malignant uterine tumors | 10,552 | 47.9 | |||
| Patients with endometrial cancer who underwent surgery for malignant uterine tumors | Patients who underwent cytological examinations scheduled at intervals of 120 days or less within the first year following the initial outpatient visit after surgery | |||||
| ut9x | Patients with cStage IA with pathological stage higher than stage IB after surgery for malignant uterine tumors* | 6,142 | 18.0 | |||
| Patients with cStage IA who underwent surgery for malignant uterine tumors | Patients with increased pathological stage (higher than stage IB) | |||||
HRT, hormone replacement therapy; PCI, pattern-of-care indicator; QI, quality indicator.
*Indicators marked are not recommended care; thus, lower adherence rates are interpreted as indicating better practices.
3. Reasons for nonadherence
While we collected the reasons for nonadherence, the number of retrieval facilities varied across indicators, with ut1 having the highest number at 362, but its median remained at 113 (interquartile range=257) (Table 3). A detailed analysis of ut7 was not performed because of the collection of reasons for fewer than ten cases. For ut1, as the analysis did not limit the stage, many responses indicated a reluctance for data collection because the clinical findings before and during surgery indicated an advanced stage. In indicators related to the treatment of advanced stages (ut2–4), reasons for nonadherence such as “elderly” and “poor PS” emerged, along with an observed increase in cases citing “patient preference.” Moreover, there was a growing trend of citing internal medical comorbidities as reasons for the inability to perform each indicator. The introduction of HRT after treatment had the lowest adherence rate at 30.9% (244/789). A total of 168 reasons for nonadherence was obtained, including cases with insufficient reasons accounting for 68.4% (115/168), and cases based on patient preference accounting for 12.5% (21/168). The values uniformly improved when the reasons for nonadherence were considered. The adjusted adherence rate showed a significant increase, with the most notable increase being a 31.1% rise in postoperative chemotherapy for intermediate-risk cases (94.7% from 63.4%), whereas changes in other areas were minimal, such as a 0.7% decrease in open surgery complications (1.8%–1.1%).
Table 3. Reasons for nonadherence as reported by facilities.
| Indicator categories & numbers | Indicators | Original adherence rate (%) | No. of reported nonadherence reasons | Converted as nonadherence | Converted as adherence | Adjusted adherence rate (%) | |||
|---|---|---|---|---|---|---|---|---|---|
| Insufficient reasons (%) | Patient preference (%) | Clinical reasons (%) | Details of clinical reasons (%) | Data malfunction (%) | |||||
| ut1 | Peritoneal lavage cytology for endometrial cancer | 9,826/10,552 (93.1) | 362 | 25.1 | N/A | 4.4 | Multiple cancers: 18.8 | 70.2 | 98.3 |
| Other comorbidities: 6.3 | |||||||||
| Already pathologically diagnosed with cancerous ascites: 12.5 | |||||||||
| Due to stage 4B: 6.3 | |||||||||
| Macroscopic peritoneal dissemination: 6.3 | |||||||||
| Recurrence: 6.3 | |||||||||
| Conducted peritoneal washings instead of ascitic fluid cytology: 6.3 | |||||||||
| Unattainable due to intra-abdominal adhesions: 6.3 | |||||||||
| Abandoned due to severe obesity: 6.3 | |||||||||
| Cancer diagnosis made postoperatively: 25.0 | |||||||||
| ut2 | Surgery for cStage III endometrial cancer | 963/1,100 (87.5) | 33 | N/A | 21.2 | 69.7 | Poor PS: 7.4 | N/A | 99.6 |
| Liver impairment/kidney impairment: 3.7 | |||||||||
| Aging: 14.8 | |||||||||
| Multiple cancers: 14.8 | |||||||||
| Cerebro/cardiovascular diseases: 7.4 | |||||||||
| Other comorbidities: 11.1 | |||||||||
| Received radiation therapy initially for hemostasis due to extensive genital bleeding, followed by chemotherapy: 3.7 | |||||||||
| Judged difficult to achieve complete resection in preoperative assessment: 14.8 | |||||||||
| Deemed unresectable during laparotomy: 7.4 | |||||||||
| ut3 | Adjuvant therapy for intermediate recurrent risk in endometrial cancer | 961/1,511 (63.6) | 152 | 14.5 | 46.7 | 38.7 | Conducted at other facilities: 3.8 | N/A | 94.7 |
| Poor PS: 2.3 | |||||||||
| Liver impairment/kidney impairment: 1.5 | |||||||||
| Respiratory diseases: 1.5 | |||||||||
| Aging: 15.4 | |||||||||
| Postoperative complications: 1.5 | |||||||||
| Mental illness/cognitive impairment: 4.6 | |||||||||
| Multiple cancers: 2.3 | |||||||||
| Cerebro/cardiovascular diseases: 2.3 | |||||||||
| Other comorbidities: 6.2 | |||||||||
| No vascular invasion: 2.3 | |||||||||
| Refusal by relatives: 0.8 | |||||||||
| ut4 | Adjuvant chemotherapy for high recurrent risk in endometrial cancer | 1,876/2,389 (78.5) | 154 | 9.1 | 34.4 | 53.9 | Conducted at other facilities: 16.9 | N/A | 98.0 |
| Clinical trials: 1.2 | |||||||||
| Poor PS: 7.2 | |||||||||
| Liver impairment/kidney impairment: 9.6 | |||||||||
| Respiratory diseases: 1.2 | |||||||||
| Aging: 25.3 | |||||||||
| Postoperative complications: 4.8 | |||||||||
| Mental illness/cognitive impairment: 14.5 | |||||||||
| Multiple cancers: 4.8 | |||||||||
| Cerebro/cardiovascular diseases: 4.8 | |||||||||
| Other comorbidities: 7.2 | |||||||||
| Performed after additional surgery: 1.2 | |||||||||
| M1 identified before determining treatment plan: 1.2 | |||||||||
| ut5 | Postoperative HRT for patients with endometrial cancer (aged 45 and under) | 244/789 (30.9) | 168 | 68.4 | 12.5 | 19.0 | Conducted at other facilities: 15.6 | 0 | 52.7 |
| Liver impairment/kidney impairment: 15.6 | |||||||||
| Postoperative complications: 3.1 | |||||||||
| Mental illness/cognitive impairment: 12.5 | |||||||||
| Multiple cancers: 25.0 | |||||||||
| Cerebro/cardiovascular diseases: 3.1 | |||||||||
| Other comorbidities: 15.6 | |||||||||
| Ovarian preservation: 9.4 | |||||||||
| ut6 | Postoperative stage of patients with endometrial cancer undergoing laparoscopic surgery* | 227/3,162 (7.2) | 74 | 71.6 | N/A | 16.2 | Conducted at other facilities: 16.7 | N/A | 4.9 |
| Multiple cancers: 8.3 | |||||||||
| Mental illness/cognitive impairment: 8.3 | |||||||||
| Other comorbidities: 8.3 | |||||||||
| Did not convert to laparotomy as T1a was confirmed intraoperatively: 58.3 | |||||||||
| ut7 | Complications following laparoscopic surgery for endometrial cancer* | 26/3,309 (0.8) | 9 | N/A | N/A | N/A | N/A | N/A | N/A |
| ut8 | Complications following open surgery for endometrial cancer* | 130/7,245 (1.8) | 32 | 59.3 | 0 | 40.6 | Clinical trials: 7.7 | 0 | 1.1 |
| Liver impairment/kidney impairment: 30.8 | |||||||||
| Other comorbidities: 61.5 | |||||||||
HRT, hormone replacement therapy; N/A, not applicable; PS, performance status.
*Indicators marked are not recommended care; thus, lower adherence rates are interpreted as indicating better practices.
4. Multivariate analysis for patient characteristics of nonadherence cases
In the multivariate analysis, age was significantly associated with achievement rates in 4 out of 5 indicators (ut2, ut3, ut4, and ut5). The clinical stage (Integrated Stage and Recurrence Risk Classification) also had a significant relationship with 2 out of 5 indicators (ut4, ut5) (Table 4). Poor ADL, indicating that the BI does not meet the maximum score of 100 points, significantly contributed to 4 out of the 5 items (ut1, ut2, ut3, ut4). Two factors were also identified to be related to nonadherence to specific indicators. BMI had a significant relationship with the performance of ascitic fluid cytology and the initiation of postoperative HRT (ut1, ut5). ACCI is significantly associated with the non-introduction of postoperative adjuvant therapy (ut3, ut4). Facility attributes contributed minimally to the achievement rates.
Table 4. Multivariate analysis for factors associated with adherence.
| Indicator categories & numbers | Indicators | Age | Integrated stage 2 | Integrated stage 3 | Integrated stage 4 | Integrated stage unknown | BMI | Barthel index (binary) | ACCI (binary) | Facility attributes | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| OR (95% CI) | p-value | OR (95% CI) | p-value | OR (95% CI) | p-value | OR (95% CI) | p-value | OR (95% CI) | p-value | OR (95% CI) | p-value | OR (95% CI) | p-value | OR (95% CI) | p-value | OR (95% CI) | p-value | ||
| ut1 | Peritoneal lavage cytology for endometrial cancer | 1.00 (0.99–1.01) | 0.831 | 0.89 (0.66–1.20) | 0.432 | 1.08 (0.85–1.37) | 0.541 | 0.65 (0.48–0.89) | 0.008 | 1.25 (0.72–2.16) | 0.425 | 0.98 (0.97–0.99) | 0.004 | 0.58 (0.43–0.77) | <0.001 | 1.22 (1.00–1.49) | 0.055 | 0.95 (0.76–1.19) | 0.640 |
| ut2 | Surgery for cStage III endometrial cancer | 0.96 (0.95–0.98) | <0.001 | - | - | - | - | - | - | - | - | 0.96 (0.95–1.04) | 0.967 | 0.27 (0.16–0.44) | <0.001 | 0.99 (0.52–1.87) | 0.971 | 1.00 (0.54–1.84) | 0.991 |
| ut3 | Adjuvant therapy for intermediate recurrent risk in endometrial cancer | 0.94 (0.93–0.95) | <0.001 | - | - | - | - | - | - | - | - | 0.99 (0.96–1.01) | 0.182 | 0.27 (0.16–0.47) | <0.001 | 0.47 (0.35–0.62) | <0.001 | 1.03 (0.75–1.43) | 0.836 |
| ut4 | Adjuvant chemotherapy for high recurrent risk in endometrial cancer | 0.94 (0.93–0.95) | <0.001 | 0.69 (0.50–0.95) | 0.024 | 1.26 (0.92–1.71) | 0.144 | NA | NA | 1.96 (1.16–3.31) | 0.012 | 0.99 (0.97–1.01) | 0.357 | 0.26 (0.18–0.37) | <0.001 | 0.37 (0.28–0.49) | <0.001 | 1.08 (0.78–1.48) | 0.655 |
| ut5 | Postoperative HRT for patients with endometrial cancer (aged 45 and under) | 0.93 (0.90–0.96) | <0.001 | 0.54 (0.30–0.98) | 0.041 | NA | NA | NA | NA | NA | NA | 0.96 (0.90–0.96) | <0.001 | 0.23 (0.05–1.01) | 0.052 | 1.43 (0.99–2.06) | 0.055 | 0.73 (0.44–1.20) | 0.212 |
ACCI, age-adjusted Charlson Comorbidity Index; BMI, body mass index; CI, confidence interval; HRT, hormone replacement therapy; NA, not applicable; OR, odds ratio.
DISCUSSION
We developed the QIs and PCIs to comprehensively assess the care of patients with endometrial cancer. A total of 8 QIs and 9 PCIs were developed in this study. Among these, there were 5 process indicators for QIs and 8 for PCIs. Additionally, there were 3 outcome indicators for QIs and one for PCIs. As a result, these QIs and PCIs aligned with the various indicators of the ESGO, which are considered world standards [9]. The adherence rates were generally high, except for items related to the HRT.
The ESGO QIs are limited to surgical indicators, including 2 indicators for preoperative evaluation, 4 for surgical techniques in the uterine body, 7 for surgical techniques in the omentum and lymph nodes, 2 for interventions in advanced stages, 2 for complications and reoperations, and 1 for postoperative adjuvant therapy. Each item had its own specified minimum target. The actual achievement rates of the established QIs are expected to be reported in the future.
The QIs established in the current study were not limited to surgery but the dominant indicators targeted surgical care, accounting for 10 out of 17 items. In terms of the surgical indicators, adherence rates for all QIs exceeded 80%. In cases where curative treatment is possible in advanced stages and surgery is desirable, the ESGO QIs suggest that surgery should be performed in over 75% of stage III–IV cases that meet these conditions. The ESGO and National Comprehensive Cancer Network (NCCN) guidelines [9,21], as well as the Japanese guidelines, recommend surgery for cases in which complete resection is achievable, even in advanced stages. In stage III disease, particularly IIIA and IIIC, where lesions involve the adnexa and lymph nodes, complete resection is feasible. While cases of stage IIIB disease with invasion of the vaginal or parametrial tissues are less common, poor treatment outcomes are observed with chemotherapy and radiation therapy, making surgery preferable, if feasible. In Japan, the adherence rate for uterine cancer surgery (ut2) is considered very high at 87.5%. For complications requiring invasive interventions and the rate of reoperation, ESGO’s QIs recommend aiming for values <2%. In Japan, as indicated in ut7 and ut8, these targets were achieved. However, our dataset lacks information on lymph node intervention and therefore cannot be assessed.
The low adherence rate to adjuvant therapy for intermediate risk endometrial cancer (63.6%), is often attributed to patient preferences. Among the 550 cases where treatment was not administered, reasons for non-adherence were collected for 152 cases (27.6%). Of these, 46.7% preferred not to undergo treatment based on patient preference. For treatment of intermediate risk cases, Japanese guidelines provide a 2C recommendation level based on several studies comparing pelvic radiotherapy with cisplatin-based chemotherapy or sequential adjuvant chemotherapy and radiotherapy [22,23]. These studies suggest less benefit for intermediate risk endometrial cancer patients than for those in the high risk group. Consequently, clinicians may describe this treatment as a controversial option, and patients often opt to forgo it. Regarding the low rate of adherence to HRT after treatment (30.9%), Japan tends to have lower rates than the global standards for gynecological cancers (e.g., ovarian cancer: 22.6% [12]). One reason for this is the widespread use of traditional herbal medicines to alleviate subjective symptoms, as observed in the ovarian cancer QI [12]. Notably, such treatments only provide relief for subjective symptoms, and the effects of hormones, including reducing the risk of developing conditions such as osteoporosis, dyslipidemia, and cardiovascular problems, cannot be replaced [24]. At present, the conventional belief that HRT based on estrogen alone increases the recurrence of endometrial cancer persists [25]. This belief, along with the fear of recurrence among patients and healthcare providers, has hindered its widespread adoption, despite evidence showing the safety of HRT with an estrogen and progesterone combination. Because the safety of HRT after surgery for early-stage endometrial cancer is established, further promotion of HRT is desirable [26].
Through the collection of reasons for nonadherence, items related to the treatment of advanced stages (ut2–ut4), such as “age,” “poor PS,” and “patient preference” were frequently mentioned. Although the number of cases was small, the increasing proportion of cases that cited “patient preference” as a reason for nonadherence for advanced-stage cancer is noteworthy. Multivariate analysis revealed that increasing age and incomplete ADL (BI not reaching 100 points) were related to a decline in adherence in ut2–ut4. Additionally, a trend was observed for citing internal medical comorbidities as reasons for nonadherence to indicators for advanced stages. In the multivariate analysis, the ACCI was a significant factor in the introduction of adjuvant therapy postoperatively. The adherence rate for HRT after treatment was notably low at 30.9% (244/789), with 68.4% (115/168) of cases having insufficient reasons, and 12.5% (21/168) of cases citing patient preferences. In multivariate analysis, HRT was less likely to be initiated in patients with higher age, stage, and BMI. Individuals with a high BMI are presumed to have a higher thrombotic tendency with HRT than individuals with a normal BMI, resulting in HRT deferral [27]. Although there were concerns about inter-facility variations, no significant differences related to facility attributes were observed for all items, indicating that standardization of care was likely achieved, albeit the possibility of certain variations in each group.
In a previous study in Germany, adherence decreased with increasing age for all items (evaluated as a continuous variable) and significantly declined for individuals aged 70 years and older (evaluated categorically), which aligned with our study [13]. This study did not yield statistically significant results concerning comorbidity. One potential reason may be the threshold used to define the presence of comorbidities. Our study defined high comorbidity at a threshold of 4 points—2 points allocated to solid tumors and an additional 2 points for other comorbidities—whereas the German study used a threshold of 1 point. As a result, our high-comorbidity group more selectively included patients with severe conditions, enhancing sensitivity and enabling us to detect statistically significant differences. Additionally, a study from Austria, analyzing stage I cases from 1995–2008, found that a CCI score of ≥2, reflecting non-solid tumor comorbidities, correlated with a decreased rate of treatment implementation [28]. In Germany, with stage I as the reference for adherence, guideline adherence increased with advancing stage, showing odds ratio (OR) 2.0 and OR 2.6 for stages II and III, and an OR of 1.2, higher than stage I but lower than stage II or III, for stage 4 [13]. Similarly, in a U.S. study, guideline adherence was higher for stage II (75.8%) and III (77.9%), with the lowest adherence observed in stage IV (35.7%) [14].
In our study, multivariable analysis identified stage progression as a significant indicator affecting guideline adherence, specifically in the evaluation of adjuvant chemotherapy for high risk endometrial cancer (ut4) and postoperative HRT (ut5). As shown in Table 3, considering factors such as patient preference and clinical reasoning led to adherence rates increasing by approximately 20%, from 78.5% to 98.0% for ut4 and from 30.9% to 52.7% for ut5. These findings imply that clinical factors, including patient preferences and comorbidities, indirectly reduce guideline adherence in advanced-stage cases. Although previous studies did not collect data on reasons for non-implementation, it is likely that these factors also contributed to reduced adherence in stage 4.
Regarding prognosis, a previous study from the United States emphasized adherence to the NCCN guidelines as an independent factor for prognosis improvement [14]. Similarly, a German study found that patients who achieved guideline adherence for surgery, cancer drug therapy, and lymph node dissection experienced a significant extension in OS [13]. Although the prognosis of the cases in 2020 could not be determined at this time, future investigations may involve an assessment of adherence and prognosis.
Our study had several limitations. Regarding the evaluation of adherence rates, facilities registered in the HBCR and participating in the QI project may be more motivated to improve cancer care, potentially leading to an overestimation of the absolute values compared with the overall medical standards in Japan. Additionally, the DPC data submitted by each medical institution may not capture treatments performed at other facilities. Through the collection of reasons for nonadherence, 3 out of 7 indicators reported that multiple cases had received care at other facilities. Item-specific limitations include the lack of vascular invasion status in the recurrence risk classification, leading to underestimation. For example, in ut3, “no vascular invasion” was reported as the reason for nonadherence in several cases. The inability to distinguish the presence of ovarian preservation as a procedural step leads to an underestimation of the ut5. Although we collected the reasons for nonadherence, the median number of facilities that reported reasons for nonadherence remained low at 113/499 (22.6%); therefore, the obtained reasons may not necessarily represent the entirety of the reasons for nonadherence. Among the reasons entered, those considered sufficient were further narrowed. In the multivariate regression analysis of factors associated with adherence, lower adherence was related to the absence of PS data collection, resulting to the use of the BI and ACCI as substitutes, which may not accurately depict the overall patient condition. Furthermore, the ACCI considers 2,182 cases with missing comorbidity data as “no comorbidities,” leading to an underestimation.
In conclusion, we developed QIs to comprehensively assess endometrial cancer treatment. Adherence rates were generally favorable; however, HRT has a low adherence rate. Factors leading to nonadherence include advanced age and incomplete ADL, particularly in advanced stages. To maintain the validity of established QIs, it is essential to incorporate new evidence and conduct appropriate reviews, such as modifying or discontinuing measurements, as needed.
ACKNOWLEDGEMENTS
We express our heartfelt gratitude to the participating facilities and practitioners who contributed to this project. We are also grateful to Mrs. Yuriko Nishikawa for supporting our project through clerical work throughout the establishment of quality indicators and completion of this manuscript.
Footnotes
Conflict of Interest: Hidemichi Watari received honoraria from MSD K.K and Eisai Co., Ltd. All remaining authors declare no conflicts of interest.
- Conceptualization: K.T., R.R., N.S., K.Y., K.M., T.H., S.M., T.M., T.T., T.Y.1, T.Y.2, M.T., W.H., H.T.
- Data curation: K.T., I.Y., Y.Y., H.T.
- Formal analysis: K.T.
- Funding acquisition: H.T.
- Investigation: K.T., I.Y., R.R., K.A.
- Methodology: K.T., I.Y., R.R.
- Project administration: K.T., H.T.
- Resources: K.T.
- Software: H.T.
- Supervision: N.S., H.T.
- Validation: K.T., K.A.
- Visualization: K.T.
- Writing - original draft: K.T.
- Writing - review & editing: I.Y., R.R., N.S., K.Y., K.M., T.H., S.M., T.M., T.T., T.Y.1, T.Y.2, M.T., W.H., Y.Y., K.A., H.T.
T.Y.1, Yoshito Terai; T.Y.2, Yasuhisa Terao.
References
- 1.United Nations Population Fund. The problem with “too few” [Internet] New York, NY: United Nations Population Fund; 2023. [cited 2024 Mar 5]. Available from: https://www.unfpa.org/swp2023/too-few#:~:text=Worldwide%2C%20fertility%20has%20fallen%20from,births%20per%20woman%20by%202050. [Google Scholar]
- 2.National Cancer Institute. Cancer stat facts: uterine cancer [Internet] Bethesda, MD: National Cancer Institute; 2024. [cited 2023 Nov 15]. Available from: https://seer.cancer.gov/statfacts/html/corp.html. [Google Scholar]
- 3.Cancer Information Service, National Cancer Center (JP) Cancer statistics. National estimates: cancer incidence data (1975-2015). (Monitoring of Cancer Incidence in Japan [MCIJ]) [Internet] Tokyo: National Cancer Center; 2021. [cited 2023 Nov 15]. Available from: https://ganjoho.jp/reg_stat/statistics/stat/cancer/18_corpus_uteri.html. [Google Scholar]
- 4.Cancer Information Service, National Cancer Center (JP) Cancer statistics. National cancer incidence data (2016-2020). [Internet] Tokyo: National Cancer Center; 2024. [cited 2025 Mar 10]. Available from: https://ganjoho.jp/reg_stat/statistics/data/dl/excel/cancer_incidenceNCR (2016-2019).xls. [Google Scholar]
- 5.Cancer Information Service, National Cancer Center (JP) National cancer mortality data (1958-2022) [Internet] Tokyo: National Cancer Center; 2023. [cited 2024 Apr 28]. Available from: https://ganjoho.jp/reg_stat/statistics/stat/cancer/18_corpus_uteri.html. [Google Scholar]
- 6.Crosbie EJ, Kitson SJ, McAlpine JN, Mukhopadhyay A, Powell ME, Singh N. Endometrial cancer. Lancet. 2022;399:1412–1428. doi: 10.1016/S0140-6736(22)00323-3. [DOI] [PubMed] [Google Scholar]
- 7.Cibula D, Planchamp F, Fischerova D, Fotopoulou C, Kohler C, Landoni F, et al. European Society of Gynaecological Oncology quality indicators for surgical treatment of cervical cancer. Int J Gynecol Cancer. 2020;30:3–14. doi: 10.1136/ijgc-2019-000878. [DOI] [PubMed] [Google Scholar]
- 8.Chargari C, Tanderup K, Planchamp F, Chiva L, Humphrey P, Sturdza A, et al. ESGO/ESTRO quality indicators for radiation therapy of cervical cancer. Radiother Oncol. 2023;183:109589. doi: 10.1016/j.radonc.2023.109589. [DOI] [PubMed] [Google Scholar]
- 9.Concin N, Planchamp F, Abu-Rustum NR, Ataseven B, Cibula D, Fagotti A, et al. European Society of Gynaecological Oncology quality indicators for the surgical treatment of endometrial carcinoma. Int J Gynecol Cancer. 2021;31:1508–1529. doi: 10.1136/ijgc-2021-003178. [DOI] [PubMed] [Google Scholar]
- 10.Querleu D, Planchamp F, Chiva L, Fotopoulou C, Barton D, Cibula D, et al. European Society of Gynaecologic Oncology quality indicators for advanced ovarian cancer surgery. Int J Gynecol Cancer. 2016;26:1354–1363. doi: 10.1097/IGC.0000000000000767. [DOI] [PubMed] [Google Scholar]
- 11.Watanabe T, Mikami M, Katabuchi H, Kato S, Kaneuchi M, Takahashi M, et al. Quality indicators for cervical cancer care in Japan. J Gynecol Oncol. 2018;29:e83. doi: 10.3802/jgo.2018.29.e83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Kakuwa T, Watanabe T, Niino M, Kawata A, Satoh T, Matsumura N, et al. Quality of care measurement for patients with ovarian cancer in Japan. J Obstet Gynaecol Res. 2024;50:1182–1191. doi: 10.1111/jog.15961. [DOI] [PubMed] [Google Scholar]
- 13.Scharl S, Sprötge T, Gerken M, Scharl A, Ignatov A, Inwald EC, et al. Factors influencing treatment decision and guideline conformity in high-grade endometrial cancer patients: a population-based study. Arch Gynecol Obstet. 2022;305:203–213. doi: 10.1007/s00404-021-06140-5. [DOI] [PubMed] [Google Scholar]
- 14.Rodriguez VE, LeBrón AMW, Chang J, Bristow RE. Guideline-adherent treatment, sociodemographic disparities, and cause-specific survival for endometrial carcinomas. Cancer. 2021;127:2423–2431. doi: 10.1002/cncr.33502. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Fitch K, Bernstein SJ, Aguilar MD, Bernard B, Juan RL, Pablo L, et al. The RAND/UCLA appropriateness method user’s manual [Internet] Santa Monica, CA: RAND Corporation; 2001. [cited 2024 Mar 22]. Available from: https://www.rand.org/pubs/monograph_reports/MR1269.html. [Google Scholar]
- 16.Higashi T, Nakamura F, Shibata A, Emori Y, Nishimoto H. The national database of hospital-based cancer registries: a nationwide infrastructure to support evidence-based cancer care and cancer control policy in Japan. Jpn J Clin Oncol. 2014;44:2–8. doi: 10.1093/jjco/hyt013. [DOI] [PubMed] [Google Scholar]
- 17.Okuyama A, Tsukada Y, Higashi T. Coverage of the hospital-based cancer registries and the designated cancer care hospitals in Japan. Jpn J Clin Oncol. 2021;51:992–998. doi: 10.1093/jjco/hyab036. [DOI] [PubMed] [Google Scholar]
- 18.George M, Smith A, Sabesan S, Ranmuthugala G. Physical comorbidities and their relationship with cancer treatment and its outcomes in older adult populations: systematic review. JMIR Cancer. 2021;7:e26425. doi: 10.2196/26425. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Yamana H, Moriwaki M, Horiguchi H, Kodan M, Fushimi K, Yasunaga H. Validity of diagnoses, procedures, and laboratory data in Japanese administrative data. J Epidemiol. 2017;27:476–482. doi: 10.1016/j.je.2016.09.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Di Donato V, D’Oria O, Giannini A, Bogani G, Fischetti M, Santangelo G, et al. Age-adjusted Charlson Comorbidity Index predicts survival in endometrial cancer patients. Gynecol Obstet Invest. 2022;87:191–199. doi: 10.1159/000525405. [DOI] [PubMed] [Google Scholar]
- 21.Abu-Rustum N, Yashar C, Arend R, Barber E, Bradley K, Brooks R, et al. Uterine neoplasms, version 1.2023, NCCN clinical practice guidelines in oncology. J Natl Compr Canc Netw. 2023;21:181–209. doi: 10.6004/jnccn.2023.0006. [DOI] [PubMed] [Google Scholar]
- 22.Susumu N, Sagae S, Udagawa Y, Niwa K, Kuramoto H, Satoh S, et al. Randomized phase III trial of pelvic radiotherapy versus cisplatin-based combined chemotherapy in patients with intermediate- and high-risk endometrial cancer: a Japanese Gynecologic Oncology Group study. Gynecol Oncol. 2008;108:226–233. doi: 10.1016/j.ygyno.2007.09.029. [DOI] [PubMed] [Google Scholar]
- 23.Hogberg T, Signorelli M, de Oliveira CF, Fossati R, Lissoni AA, Sorbe B, et al. Sequential adjuvant chemotherapy and radiotherapy in endometrial cancer--results from two randomised studies. Eur J Cancer. 2010;46:2422–2431. doi: 10.1016/j.ejca.2010.06.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Pitkin J. Alternative and complementary therapies for the menopause. Menopause Int. 2012;18:20–27. doi: 10.1258/mi.2012.012001. [DOI] [PubMed] [Google Scholar]
- 25.Burger CW, van Leeuwen FE, Scheele F, Kenemans P. Hormone replacement therapy in women treated for gynaecological malignancy. Maturitas. 1999;32:69–76. doi: 10.1016/s0378-5122(98)00113-3. [DOI] [PubMed] [Google Scholar]
- 26.Shim SH, Lee SJ, Kim SN. Effects of hormone replacement therapy on the rate of recurrence in endometrial cancer survivors: a meta-analysis. Eur J Cancer. 2014;50:1628–1637. doi: 10.1016/j.ejca.2014.03.006. [DOI] [PubMed] [Google Scholar]
- 27.Canonico M, Plu-Bureau G, Lowe GDO, Scarabin PY. Hormone replacement therapy and risk of venous thromboembolism in postmenopausal women: systematic review and meta-analysis. BMJ. 2008;336:1227–1231. doi: 10.1136/bmj.39555.441944.BE. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Boll D, Verhoeven RH, van der Aa MA, Lybeert ML, Coebergh JW, Janssen-Heijnen ML. Adherence to national guidelines for treatment and outcome of endometrial cancer stage I in relation to co-morbidity in southern Netherlands 1995-2008. Eur J Cancer. 2011;47:1504–1510. doi: 10.1016/j.ejca.2011.03.023. [DOI] [PubMed] [Google Scholar]