DOI: 10.1200/OP.21.00351 JCO Oncology Practice - published online before print September 28, 2021
Physician Influence on Variation in Receipt of Aggressive End-of-Life Care Among Women Dying of Ovarian Cancer
2Department of Gynecologic Oncology, University of Michigan, Ann Arbor, MI
3Wayne State University School of Medicine and Karmanos Cancer Institute, Detroit, MI
4Department of Epidemiology and Institute for Social Research, University of Michigan, Ann Arbor, MI
5Department of Internal Medicine, University of Michigan, Ann Arbor, MI
P.C. and L.P.W. contributed equally to senior authorship.
End-of-life care for women with ovarian cancer is persistently aggressive, but factors associated with overuse are not well understood. We evaluated physician-level variation in receipt of aggressive end-of-life care and examined physician-level factors contributing to this variation in the SEER-Medicare data set.
Medicare beneficiaries with ovarian cancer who died between 2000 and 2016 were included if they were diagnosed after age 66 years, had complete Medicare coverage between diagnosis and death, and had outpatient physician evaluation and management for their ovarian cancer. Using multilevel logistic regression, we examined physician variation in no hospice enrollment, late hospice enrollment (≤ 3 days), > 1 emergency department visit, an intensive care unit stay, terminal hospitalization, > 1 hospitalization, receiving a life-extending or invasive procedure, and chemotherapy (in the last 2 weeks).
In this sample of 6,288 women, 51% of women received at least one form of aggressive end-of-life care. Most common were no hospice enrollment (28.9%), an intensive care unit stay (18.6%), and receipt of an invasive procedure (20.7%). For not enrolling in hospice, 9.9% of variation was accounted for by physician clustering (P < .01). Chemotherapy had the highest physician variation (12.4%), with no meaningful portion of the variation explained by physician specialty, volume, region, or patient characteristics.
In this study, a meaningful amount of variation in aggressive end-of-life care among women dying of ovarian cancer was at the physician level, suggesting that efforts to improve the quality of this care should include interventions aimed at physician practices and decision making in end-of-life care.
Aggressive cancer care at the end of life does not meaningfully lengthen survival, and it is associated with worse patient quality of life and less satisfaction among surviving family members.1-3 Both ASCO and the National Quality Forum have end-of-life quality standards advising against care that is overly intensive or invasive.4,5 Despite these clinical guidelines, end-of-life care among women dying of ovarian cancer is persistently aggressive.6 Trends in multiple end-of-life emergency department (ED) visits and intensive care unit (ICU) stays are increasing, whereas probability of aggressive treatments has not improved.7
Why aggressive care for ovarian cancer at the end of life persists despite guidelines promoting more palliative care and less overuse of invasive treatments is not well understood?4,5,8 Existing studies exploring end-of-life care for women with ovarian cancer have focused on the influence of patient characteristics.7,9,10 A single-institution study found that women with ovarian cancer who received a provider recommendation for hospice were much more likely to enroll.11 Prior studies have found that for other cancer deaths, the physician's beliefs influenced variation in end-of-life care intensity and the patient's beliefs did not.12 However, to date, the influence of physicians on receipt of aggressive end-of-life care for ovarian cancer has not been evaluated. Understanding physicians' influence on the receipt of aggressive end-of-life care in this context is necessary to identify potential targets for future interventions for improving the quality of end-of-life care for women with ovarian cancer.
Therefore, the goal of this study was to evaluate receipt of aggressive end-of-life care among women dying of ovarian cancer using the SEER-Medicare data linkage and assess the amount of variation attributable to physicians. We also evaluated the association between physician-level characteristics, including specialty and ovarian cancer patient volume, and multiple aggressive end-of-life care indicators.5,8,13,14
This analysis includes Medicare claims data from years 2000 to 2016 linked with SEER-Registry data.15 The SEER program includes 18 population-based cancer registries collecting patient demographics, tumor characteristics, first course of treatment, and survival on all residents in their geographic catchment areas who are diagnosed with invasive cancer. The SEER-Medicare linkage contains treatment and outcomes for roughly 25% of elderly patients with cancer in the United States.
Women diagnosed between 2000 and 2015 who died between 2000 and 2016 with a first and only primary ovarian cancer (with the exception of basal and squamous cell skin carcinomas) were included in this analysis. The sample was restricted to women older than 66 years at diagnosis who had complete case information captured by a SEER registry (n = 26,012). Women not enrolled in both Part A and Part B Medicare or who were enrolled in a Health Maintenance Organization plan in the 12 months before diagnosis were excluded due to incomplete treatment information (n = 7,220). We excluded cases with unknown diagnosis or death month, death within 30 days of diagnosis or no death date, discrepancies between SEER and Medicare birth and death dates, diagnosis at death or autopsy, noninvasive disease, no claims after diagnosis, and hospice admission date predating diagnosis (n = 6,387). Women who did not have any outpatient ovarian cancer management were excluded (n = 3,393), as were patients seen by physicians with only one patient in the analytic sample (n = 2,724). In total, 6,288 women with ovarian cancer were included in our analytic sample. The Wayne State University Institutional Review Board determined this study was exempt.
Aggressive end-of-life care has been well defined in a validated set of claims-based metrics reflecting inappropriate end-of-life care among populations dying of cancer.5-8,10,13,16 Using these metrics, outcomes included indicators for hospice use, end-of-life treatments, and end-of-life hospital utilization.8,13,14 Hospice use included not enrolling in hospice (no hospice) or hospice enrollment less than 3 days before death (late hospice). Treatment measures included chemotherapy in the last 2 weeks of life, life-extending procedures (ventilation, resuscitation, or feeding tubes), and invasive procedures (surgery requiring anesthesia, placement of arterial or central line, endoscopy, interventional radiology procedure, radiotherapy, or pelvic examination with tissue sampling).7,10,17 Following prior methods, paracentesis, thoracentesis, and venting feeding tubes are considered acceptable palliative procedures in the last 30 days of life and were excluded from these measures.10 Hospital utilization measures included > 1 ED visit or hospital admission, ICU admission, and death in an acute care hospital.13,14,16,18 Independent binary measures for each outcome were created using Medicare claims data from the last 30 days of life on the basis of prior definitions and updated with International Classification of Diseases codes (ICD) 10th revision coding.6,7,10
Women were assigned to the physician they saw most for their ovarian cancer in an outpatient evaluation and management (E/M) setting. Encounters were classified as E/M using current procedural terminology codes beginning with 99 in the National Claims History Medicare claims file, and encounters were restricted to ovarian cancer by diagnosis code (ICD, 9th revision: 183, 1830, 18300, 18309, 1832, 1833, 1834, 1835, 1836, 1837, 1838, and 1839; ICD, 10th revision: C561, C562, and C569). We included physicians who saw at least two patients in the analytic sample, so there are an average of four clustered patients per physician.
Physician characteristics were derived from the entire sample of women with ovarian cancer, regardless of vital status. Women were included who had the same inclusion criteria as the analytic cohort in this analysis, with the exception of having died. Physician characteristics are derived from all women with complete Medicare coverage diagnosed between 2000 and 2016 with invasive ovarian cancer.
Physicians were assigned to the most specialized specialty code they billed across E/M visits.19 If the specialty was not gynecologic (GYN) oncology, oncology, or gynecology care, they were classified as others.
Patient volume was determined by counting the number of women with ovarian cancer who had an E/M encounter with each provider per year.19 Annual patient volume was categorized as < 3, 3 to < 5, 5 to < 10, and 10+.
A priori identified confounders included race or ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic, Asian, or other races), nonurban residence, age of death, years between diagnosis and death, SEER stage at diagnosis (localized, regional, distant, or unknown), Charlson score (0, 1, 2, or 3+), census tract poverty (0% to < 5%, 5 to < 10%, 10 to < 20%, 20% to 100%, or unknown), and marital status (married or unmarried). Race or ethnicity was determined by SEER abstractors, with priority given to self-reported information.20 Charlson score category is derived from a modified Charlson index score for each patient calculated from the 12 months before diagnosis.21,22
We first tabulated patient and provider characteristics and examined bivariate associations between physician characteristics and receipt of aggressive end-of-life care indicators using chi-squared tests. We then fit multilevel logistic models for each outcome with a random intercept for each physician. The interclass correlation coefficient (ICC) was calculated for each model to determine the percent variation in each outcome attributable to the physician level and evaluate how that variation changes with the addition of physician characteristics, including specialty, ovarian cancer patient volume, and region.
Separate multivariable multilevel logistic models with a random intercept for physician were used to evaluate the association of physician characteristics (volume, specialty, and region) with each outcome. Models were adjusted for patient-level age at death, Charlson score, SEER stage at diagnosis, survival time, patient race or ethnicity, urbanicity of residence, marital status, and poverty. All analyses were conducted in SAS version 9.4, and a two-tailed P value < .05 was considered statistically significant.
We conducted sensitivity analyses to evaluate the robustness of the patient-physician clustering and our exclusion of single-patient physicians. The analysis was also run with restriction to physicians seeing at least 3 patients. The results were similar in each case.
Table 1 presents patient demographic and clinical characteristics and physician characteristics. The majority of women included in this analytic sample were non-Hispanic White (86.1%), were not currently married (56.2%), and had no comorbid conditions (59.9%). Most women lived in urban (98.0%) census tracts with < 20% poverty (84.9%). The majority of women were diagnosed with distant spread disease (80.8%) that had serous histology (76%). The mean survival was 2.7 years with a standard deviation of 2.5 years. The mean age at death was 79.1 years. Most women received the majority of their ovarian cancer care from medical oncologists (53.4%) or GYN oncologists (34.8%). Women were seen by 1,467 physicians totally. Most physicians were medical oncologists (67%) or GYN oncologists (21%). Physicians had a mean 4.3 patients overall, and most (46.6%) had an annual ovarian cancer volume of < 3 (Table 1).
Figure 1 displays the distribution of aggressive end-of-life care by physician specialty. In this sample, 51% of women received at least one type of aggressive end-of-life care. Most common were no hospice enrollment (28.9%), ICU stay (18.6%), invasive procedure (20.7%), and terminal hospitalization (17.2%). The proportion of women not enrolling in hospice, staying in the ICU, receiving a life-extending procedure, and receiving an invasive procedure varied by physician specialty. Compared with patients of other specialty physicians, fewer women (nearly 20%) managed by GYN oncologists did not enroll in hospice (Fig 1). A smaller proportion (23%) of women with high-volume providers did not enroll in hospice (P < .01). By contrast, a larger proportion of women with high-volume providers received an invasive procedure (22.5%, P < .03) or a life-extending procedure (11.4%, P = .02; results not shown).
Table 2 displays the ICC for each type of aggressive end-of-life care. The ICC is the proportion of variance in each outcome variable that is explained by the grouping of patients under their most-seen physician.23 Chemotherapy had the highest physician variation (12.4%) with about 4% of this variation explained by physician specialty, volume and region, and no change with addition of patient characteristics (P = .02). In the hospice use group, 9.9% of variation in not enrolling in hospice was accounted for by physician clustering (P < .01), and 17% of this variation was explained by physician characteristics. After accounting for patient characteristics, around 8% of the variation at the physician level remained.
Among hospital utilization measures, having an ICU stay, > 1 ED visit, and terminal hospitalization varied significantly at the physician level (Table 2). About 6% of variation in terminal hospitalization is explained at the physician level, and 8% of that variation is explained by physician characteristics. Similarly, 5% of variation in ICU stays is explained at the physician level, and 33% of that variation is explained by physician characteristics. ED variation was no longer statistically significant at the physician level when accounting for different patient characteristics. All intensive treatments had significant variation because of physician clustering. After adjustment for physician specialty and volume, variation in life-extending procedures and invasive procedures was no longer significant (Table 2).
In a sensitivity analysis including physicians with only one patient, results did not change meaningfully for outcomes other than chemotherapy. Sensitivity results for chemotherapy found 36% variation in receipt of chemotherapy at the physician level (P < .01), with about 8% of that variation explained by adding physician and patient characteristics and 32.8% variation (P < .01) at the physician level remaining (Appendix Table A1, online only).
Table 3 displays the adjusted odds ratios (ORs) and 95% CIs for the associations between physician characteristics and aggressive end-of-life care indicators. In fully adjusted models, compared with the other specialty physicians, patients of oncologists and GYN oncologists had lower odds of no hospice enrollment (OR 0.61; CI, 0.42 to 0.88; OR 0.64; CI, 0.42 to 0.96). Compared with those with a low-volume provider, women seeing a physician with a high annual volume also had lower odds of no hospice enrollment (OR 0.67; CI, 0.52 to 0.87). Compared with patients of other specialty physicians, women seeing GYN oncologists and oncologists were less likely to receive a life-extending procedure (OR 0.54; CI, 0.32 to 0.90 and OR 0.43; CI, 0.27 to 0.68; Table 3).
Among this national sample of women dying of ovarian cancer, a significant amount of variation in not enrolling patients in hospice, inappropriate end-of-life hospital utilization, and aggressive treatments were explained at the physician level. Although chemotherapy in the last 2 weeks of life was the least frequent outcome, it had the largest variation explained at the physician level, and this variation was not further explained by accounting for physician specialty, volume, or patient mix. Notably, physicians' influence on variation in no hospice enrollment and chemotherapy was much higher than physicians' influence on the variation in any aggressive care, indicating that physicians' influence may be most important for these specific aspects of end-of-life care.
Our findings suggest that physicians had significant influence on the variation in patients with ovarian cancer not enrolling in hospice at the end of life. Prior research suggests that physician recommendations and engagement in end-of-life care discussions influence enrollment in hospice11 and women receiving less aggressive care.24 Prior studies of regional variation have found that physicians drive the regional variation in end-of-life care intensity,12,25 and physician end-of-life practices are influenced by physicians' peers both within and outside of practice.26 However, in our study, physicians' influence on variation in the use of hospice and other end-of-life care services was not meaningfully explained by region, patient volume, or their specialty. Our findings build on prior studies to quantify the extent to which physicians influence the receipt of hospice and aggressive end-of-life care services among women dying of ovarian cancer. As end-of-life care remains aggressive for women with ovarian cancer,7,23,27 despite guidelines targeted toward increasing hospice enrollment and reducing overuse of intensive treatments, additional efforts are warranted to better support physicians in engaging in end-of-life and goals of care conversations with their patients.
In our sample, the largest amount of physician variation was found in receipt of chemotherapy in the last two weeks of life. Very little of this variation was explained by physician specialty, annual ovarian cancer patient volume, physician region, or patient mix. End-of-life chemotherapy is considered low-quality care because it reduces quality of life without improving survival, and it can increase risk for receipt of other aggressive care such as ventilation and dying in an ICU.28-30 A relevant exception for women with ovarian cancer is clinical trial participation. Women with ovarian cancer who participate in clinical trials are more likely to receive chemotherapy in the last 2 weeks of life, but clinical trial enrollment remains low.31-33 In our sensitivity analysis that included single-patient providers, we saw a substantial increase in physician-level variation for chemotherapy. This is in line with a prior study of variation in end-of-life chemotherapy use in patients with lung cancer. Green et al34 found that physicians in small independent practices were more likely to order chemotherapy in the last month of life. Taken together, these findings suggest that physicians are an important target for reducing end-of-life chemotherapy among women with ovarian cancer.
This study included a large, diverse, population-based sample of women with ovarian cancer and used a validated set of aggressive end-of-life care indicators that reflect distinct elements of care. However, there are potential limitations that warrant comment. First, we assigned women to the provider they saw most for their outpatient ovarian cancer E/M visits. Given our data source, we were unable to ascertain whether this was indeed the most influential physician for the end-of-life care received. However, we expect our results are a conservative estimate of physician-level variation. We also selected this assignment method as outpatient visits are the optimal setting for appropriate goals of care discussions, as has been done in prior studies. Second, we used Medicare ovarian cancer patient claims data to assign provider specialty and ovarian cancer patient volume. Although Medicare volume is only a portion of true patient mix, the median age of ovarian cancer diagnosis is 63 years, so older women likely reflect a substantial proportion of patient mix. Third, our sample only includes older women and does not reflect Medicare HMO recipients; thus, our generalizability is limited to this group. However, unlike the prior single-institution studies evaluating ovarian cancer end-of-life care, using Medicare claims provides a more representative sample and ensures we have full clinical details no matter which hospital or hospice a patient visited which gives us a complete picture of the end-of-life care received. Fourth, it is possible our measure of invasive and life-extending procedures captured some appropriate palliative procedures. We suspect this would be a small number as we have excluded procedures that were clearly appropriate for palliation at the end of life, but not all are identifiable. Finally, we were unable to account for hospital characteristics in this analysis, and they may influence physician practice patterns. Future work clarifying this relationship is warranted.
In conclusion, in this population-based sample of women with ovarian cancer, variation in receipt of end-of-life care was in part due to physician influence. Our results suggest that physicians may have end-of-life practices that are contributing to the receipt of aggressive care at the end of life and patients not enrolling in hospice. Efforts to improve the quality of end-of-life care should therefore include interventions targeted toward improving patient-provider decision making about end-of-life care and better supporting providers to engage in goals of care and end-of-life care discussions.
The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or the National Cancer Institute.
M.A.M. received research support from the National Cancer Institute institutional training grant T32-CA-236621.
Conception and design: Megan A. Mullins, Shitanshu Uppal, Michele L. Cote, Lauren P. Wallner
Administrative support: Julie J. Ruterbusch, Michele L. Cote, Lauren P. Wallner
Collection and assembly of data: Julie J. Ruterbusch, Michele L. Cote
Data analysis and interpretation: All authors
Manuscript writing: All authors
Final approval of manuscript: All authors
Accountable for all aspects of the work: All authors
The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/op/authors/author-center.
Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).
Lauren P. Wallner
Honoraria: Kaiser Permanente
No other potential conflicts of interest were reported.
This study used the linked SEER-Medicare database. The interpretation and reporting of these data are the sole responsibility of the authors. The authors acknowledge the efforts of the National Cancer Institute; the Office of Research, Development and Information, CMS; Information Management Services (IMS) Inc; and the SEER Program tumor registries in the creation of the SEER-Medicare database.
|1.||Wright AA, Keating NL, Ayanian JZ, et al: Family perspectives on aggressive cancer care near the end of life. JAMA 315:284-292, 2016 Crossref, Medline, Google Scholar|
|2.||Wright AA, Zhang B, Ray A, et al: Associations between end-of-life discussions, patient mental health, medical care near death, and caregiver bereavement adjustment. JAMA 300:1665-1673, 2008 Crossref, Medline, Google Scholar|
|3.||von Gruenigen V, Daly B, Gibbons H, et al: Indicators of survival duration in ovarian cancer and implications for aggressiveness of care. Cancer 112:2221-2227, 2008 Crossref, Medline, Google Scholar|
|4.||National Quality Forum: NQF-endorsed palliative care and end of life care endorsement maintenance standards. https://www.qualityforum.org/Projects/n-r/Palliative_Care_and_End-of-Life_Care/Table_of_Measures.aspx Google Scholar|
|5.||Jacobson JO, Neuss MN, Hauser R: Measuring and improving value of care in oncology practices: ASCO programs from quality oncology practice initiative to the rapid learning system. Am Soc Clin Oncol Ed Book:e70-e76, 2012 Link, Google Scholar|
|6.||Wright AA, Hatfield LA, Earle CC, et al: End-of-life care for older patients with ovarian cancer is intensive despite high rates of hospice use. J Clin Oncol 32:3534-3539, 2014 Link, Google Scholar|
|7.||Mullins M, Ruterbusch JJ, Clarke P, et al: Trends and racial disparities in aggressive end of life care for a national sample of women with ovarian cancer. Cancer 127:2229-2237, 2021 Crossref, Medline, Google Scholar|
|8.||National Quality Forum: NQF: National voluntary consensus standards for quality of cancer care. http://www.qualityforum.org/publications/2009/05/National_voluntary_consensus_standards_for_Quality_of_Cancer_Care.aspx Google Scholar|
|9.||Taylor JS, Brown AJ, Prescott LS, et al: Dying well: How equal is end of life care among gynecologic oncology patients? Gynecol Oncol 140:295-300, 2016 Crossref, Medline, Google Scholar|
|10.||Taylor JS, Rajan SS, Zhang N, et al: End-of-Life racial and ethnic disparities among patients with ovarian cancer. J Clin Oncol 35:1829-1835, 2017 Link, Google Scholar|
|11.||Brown AJ, Sun CC, Prescott LS, et al: Missed opportunities: Patterns of medical care and hospice utilization among ovarian cancer patients. Gynecol Oncol 135:244-248, 2014 Crossref, Medline, Google Scholar|
|12.||Keating NL, Huskamp HA, Kouri E, et al: Factors contributing to geographic variation in end-of-life expenditures for cancer patients. Health Aff (Millwood) 37:1136-1143, 2018 Crossref, Medline, Google Scholar|
|13.||Earle CC, Park ER, Lai B, et al: Identifying potential indicators of the quality of end-of-life cancer care from administrative data. J Clin Oncol 21:1133-1138, 2003 Link, Google Scholar|
|14.||Earle CC, Landrum MB, Souza JM, et al: Aggressiveness of cancer care near the end of life: Is it a quality-of-care issue? J Clin Oncol 26:3860-3866, 2008 Link, Google Scholar|
|15.||Warren J, Klabunde C, Schrag D, et al: Overview of the SEER-medicare data: Content, research applications, and generalizability to the United States elderly population. Med Care 40:3-18, 2002 Crossref, Google Scholar|
|16.||Earle CC, Neville BA, Landrum MB, et al: Trends in the aggressiveness of cancer care near the end of life. J Clin Oncol 22:315-321, 2004 Link, Google Scholar|
|17.||Wu E, Rogers A, Ji L, et al: Escalation of oncologic services at the end of life among patients with gynecologic cancer at an urban, public hospital. J Oncol Pract 11:e163-e169, 2015 Link, Google Scholar|
|18.||Morden NE, Chang C-H, Jacobson JO, et al: End-of-life care for Medicare beneficiaries with cancer is highly intensive overall and varies widely. Health Aff (Millwood) 31:786-796, 2012 Crossref, Medline, Google Scholar|
|19.||Baldwin L-MM, Adamache W, Klabunde CN, et al: Linking physician characteristics and Medicare claims data: Issues in data availability, quality, and measurement. Med Care 40:IV-82-IV-95, 2002 Crossref, Google Scholar|
|20.||Adamo M, Dickie L, Ruhl J: SEER Program Coding and Staging Manual 2018. National Cancer Institute, Bethesda, MD 20892. U.S. Department of Health and Human Services, National Institutes of Health National Cancer Institute, 2018. https://seer.cancer.gov/archive/manuals/2018/SPCSM_2018_maindoc.pdf Google Scholar|
|21.||Klabunde CN, Potosky AL, Legler JM, et al: Development of a comorbidity index using physician claims data. J Clin Epidemiol 53:1258-1267, 2000 Crossref, Medline, Google Scholar|
|22.||Charlson ME, Pompei P, Ales KL, et al: A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis 40:373-383, 1987 Crossref, Medline, Google Scholar|
|23.||Merlo J, Chaix B, Yang M, et al: A brief conceptual tutorial of multilevel analysis in social epidemiology: Linking the statistical concept of clustering to the idea of contextual phenomenon. J Epidemiol Community Health 59:443-449, 2005 Crossref, Medline, Google Scholar|
|24.||Lopez-Acevedo M, Havrilesky LJ, Broadwater G, et al: Timing of end-of-life care discussion with performance on end-of-life quality indicators in ovarian cancer. Gynecol Oncol 130:156-161, 2013 Crossref, Medline, Google Scholar|
|25.||Gallo JJ, Andersen MS, Hwang S, et al: Physician preferences for aggressive treatment at the end of life and area-level health care spending: The Johns Hopkins precursors study. Gerontol Geriatr Med 3:2333721417722328, 2017 Crossref, Google Scholar|
|26.||Keating NL, O'Malley AJ, Onnela J-P, et al: Influence of peer physicians on intensity of end-of-life care for cancer decedents. Med Care 57:468-474, 2019 Crossref, Medline, Google Scholar|
|27.||Wang S-Y, Hall J, Pollack CE, et al: Trends in end-of-life cancer care in the Medicare program. J Geriatr Oncol 7:116-125, 2016 Crossref, Medline, Google Scholar|
|28.||Prigerson HG, Bao Y, Shah MA, et al: Chemotherapy use, performance status, and quality of life at the end of life. JAMA Oncol 1:778-784, 2015 Crossref, Medline, Google Scholar|
|29.||Blanke CD, Fromme EK: Chemotherapy near the end of life: First–and third and fourth (line)—Do no harm. JAMA Oncol 1:785-786, 2015 Crossref, Medline, Google Scholar|
|30.||Wright AA, Zhang B, Keating NL, et al: Associations between palliative chemotherapy and adult cancer patients' end of life care and place of death: Prospective cohort study. BMJ 348:g1219, 2014 Crossref, Medline, Google Scholar|
|31.||Nitecki R, Bercow AS, Gockley AA, et al: Clinical trial participation and aggressive care at the end of life in patients with ovarian cancer. Int J Gynecol Cancer 30:201-206, 2020 Crossref, Medline, Google Scholar|
|32.||Murthy VH, Krumholz HM, Gross CP: Participation in cancer clinical trials: Race-, sex-, and age-based disparities. JAMA 291:2720-2726, 2004 Crossref, Medline, Google Scholar|
|33.||Greenwade MM, Moore KN, Gillen JM, et al: Factors influencing clinical trial enrollment among ovarian cancer patients. Gynecol Oncol 146:465-469, 2017 Crossref, Medline, Google Scholar|
|34.||Green JB, Shapiro MF, Ettner SL, et al: Physician variation in lung cancer treatment at the end of life. Am J Manag Care 23:216-223, 2017 Medline, Google Scholar|