Comparing the clinical trial efficacy versus real-world effectiveness of treatments for multiple myeloma: a population-based study.

Not available.

Patterns of Survivorship Follow-Up Care Among Patients With Breast Cancer: A Retrospective Population-Based Cohort Study in Ontario, Canada, Between 2006 and 2016.

PURPOSE: Many cancer survivors have ongoing follow-up with their oncologist(s), despite evidence that this care can be competently managed by primary care and transitioning well survivors could relieve growing pressure on cancer care systems. We analyzed population-based administrative data from Ontario, Canada, to examine rates of transition to primary care-led follow-up care during the survivorship phase, including clinical and demographic predictors associated with being transitioned. METHODS: We conducted a retrospective cohort study to describe the patterns of survivorship follow-up care among all patients with breast cancer in Ontario from 2006 to 2016. Data were derived from the Ontario Cancer Registry and other linked data sets. We defined the survivorship phase of care beginning at 2 years after initial diagnosis. Logistic regression was used to explore factors potentially prognostic of no oncology visits in each of the years after survivorship. RESULTS: Our survivorship cohort was composed of 71,719 patients with breast cancer, 42% of whom were considered to have transitioned from oncology to primary care 2 years after diagnosis. Although the number of patients having oncology visits diminished over time, a quarter of the cohort continued being seen in year 5 of survivorship. Regression analysis found older age, early cancer stage, living farther from a cancer center, not receiving radiation or chemotherapy, and high well-being to be associated with transitioning to primary care. CONCLUSION: Our findings contribute to the development of low-risk profiles among survivors to inform optimal transition from oncology to primary care. Further research examining qualitative perspectives from oncologists, cancer survivors, and primary care is also required to illuminate other sentinel factors to be considered when transitioning during follow-up.

Increasing interprofessional collaboration in community-based palliative care: a pilot study of the CAPACITI education program for primary care providers.

Interprofessional collaboration in palliative care is essential to ensuring high-quality care for seriously ill patients. Education interventions to increase competency in palliative care should incorporate team-building skills to encourage an interprofessional approach. We developed and piloted a virtual educational program named CAPACITI for interprofessional teams to promote a community palliative approach to care. Primary care teams from across Ontario, Canada, participated in CAPACITI which consisted of 10 facilitated sessions that emphasized how to operationalize a palliative care approach as a team. Pre- and post-study questionnaires were completed by each team, including the AITCS-II, a validated instrument that measures interprofessional collaboration. We analyzed individual paired differences in summary scores and in each of three subdomains of the AITCS-II questionnaire: partnership, cooperation, and coordination. Seventeen teams completed the AITCS-II post survey, representing 133 participants. Teams varied demographically and ranged from 5 to 16 members. After CAPACITI, the overall mean AITCS-II summary score among teams increased to 96.0 (SD = 10.0) for a significant paired mean difference increase of 9.4 (p = .03). There were also significant increases in the partnership (p = .01) and in the cooperation subdomains (p = .04). CAPACITI demonstrated the potential for improving collaboration among primary care teams, which can lead to improved provider and patient outcomes in palliative care.

Advance Directives Change Frequently in Nursing Home Residents.

OBJECTIVES: To describe the rate, timing, and pattern of changes in advance directives (ADs) of do not resuscitate (DNR) and do not hospitalize (DNH) orders among new admissions to nursing homes (NHs). DESIGN: A retrospective cohort study. SETTING AND PARTICIPANTS: Admissions to all publicly funded NHs in Ontario, Canada, between January 1, 2013, and December 31, 2017. METHODS: Residents were followed until discharged from incident NH stay, death, or were still present at the end of study (December 31, 2019). They were categorized into 3 mutually exclusive baseline composite AD groups: Full Code, DNR Only, and DNR+DNH. We used Poisson regression models to estimate the incidence rate ratios of AD change between different AD groups and different decision makers for personal care, adjusted for baseline clinical and sociodemographic variables. RESULTS: A total of 102,541 NH residents were eligible for inclusion. Residents with at least 1 AD change accounted for 46% of Full Code, 30% of DNR Only, and 25% of DNR+DNH group. Median time to first AD change ranged between 26 and 55 weeks. For Full Code and DNR Only residents, the most frequent change was to an AD 1 level lower in aggressiveness or intervention, whereas for DNR+DNH residents the most frequent change was to DNR Only. About 16% of residents had 2 or more AD changes during their stay. After controlling for covariates, residents with a DNR-only order or DNR+DNH orders at admission and those with a surrogate decision maker were associated with lower AD change rates. CONCLUSIONS AND IMPLICATIONS: Measuring AD adherence rates that are documented only at a particular time often underestimates the dynamics of AD changes during a resident's stay and results in an inaccurate measure of the effectiveness of AD on resident care. There should be more frequent reviews of ADs as they are quite dynamic. Mandatory review after an acute change in a resident's health would ensure that ADs are current.

A Prognostic Survival Model Incorporating Patient-Reported Outcomes for Transplant-Ineligible Patients With Multiple Myeloma.

Developing prognostic tools specifically for patients themselves represents an important step in empowering patients to engage in shared decision-making. Incorporating patient-reported outcomes may improve the accuracy of these prognostic tools. We conducted a retrospective population-based study of transplant-ineligible (TIE) patients with multiple myeloma (MM) diagnosed between January 2007 and December 2018. A multivariable Cox regression model was developed to predict the risk of death within 1-year period from the index date. We identified 2356 patients with TIE MM. The following factors were associated with an increased risk of death within 1 year: age > 80 (HR 1.11), history of heart failure (HR 1.52), "CRAB" at diagnosis (HR 1.61), distance to cancer center (HR 1.25), prior radiation (HR 1.48), no proteosome inhibitor/immunomodulatory therapy usage (HR 1.36), recent emergency department (HR 1.55) or hospitalization (HR 2.13), poor performance status (ECOG 3-4 HR 1.76), and increasing number of severe symptoms (HR 1.56). Model discrimination was high with C-statistic of 0.74, and calibration was very good. To our knowledge, this represents one of the first prognostic models developed in MM incorporating patient-reported outcomes. This survival prognostic tool may improve communication regarding prognosis and shared decision-making among older adults with MM and their health care providers.

Treatment Pattern, Healthcare Resource Utilization and Symptom Burden Among Patients with Triple Class Exposed Multiple Myeloma: A Population-Based Cohort Study.

PURPOSE: This study aims to describe the treatment patterns, outcomes, health care utilization and symptom burden of triple class exposed (TCE) relapsed/refractory patents with multiple myeloma (MM) receiving a subsequent line of treatment (LOT). METHODS: This is a retrospective observational cohort study using administrative databases in Ontario, Canada. Outcomes were captured for TCE patients receiving a subsequent LOT and included: treatment regimen details, time to next treatment (TTNT), overall survival (OS), health care utilization, palliative care referral, and patient reported symptoms. RESULTS: Of the 16,777 patients diagnosed with MM between 2007-2021 in Ontario, 1358 (8%) patients were classified as TCE. Among the TCE MM patients, 489 (36%) received a subsequent LOT. The two most commonly administered therapies post TCE were carfilzomib/dexamethasone (n = 111, 22%) and pomalidomide/dexamethasone(n = 95, 19%). Median TTNT was 1.7 months (95%CI 1.2-2.4 months) and median OS 12.8 months (95%CI 10.8-16.5). Healthcare utilization was high with 276 (56%) of patients evaluated in an emergency department (ED) or admitted to hospital. There was high symptom burden as reported by patients with moderate-severe impairment in well-being, fatigue, pain and drowsiness noted in greater than 25% of the cohort. Palliative care referrals rates were low with only 10% (n = 48) patients referred to palliative care. Among the patients that died during study follow up, the majority died in hospital (n = 147,44%). CONCLUSION: Our study reports one of the largest series of real-world TCE patients published and demonstrates the poor outcomes of TCE patients receiving a subsequent LOT.

Sex-Based Analysis of Quality Indicators of End-of-Life Care in Gastrointestinal Malignancies.

Indices of aggressive or supportive end-of-life (EOL) care are used to evaluate health services quality. Disparities according to sex were previously described, with studies showing that male sex is associated with aggressive EOL care. This is a secondary analysis of 69,983 patients who died of a GI malignancy in Ontario between 2006 and 2018. Quality indices from the last 14-30 days of life and aggregate measures for aggressive and supportive EOL care were derived from administrative data. Hospitalizations, emergency department use, intensive care unit admissions, and receipt of chemotherapy were considered indices of aggressive care, while physician house call and palliative home care were considered indices of supportive care. Overall, a smaller proportion of females experienced aggressive care at EOL (14.3% vs. 19.0%, standardized difference = 0.13, where ≥0.1 is a meaningful difference). Over time, rates of aggressive care were stable, while rates of supportive care increased for both sexes. Logistic regression showed that younger females (ages 18-39) had increased odds of experiencing aggressive EOL care (OR 1.71, 95% CI 1.30-2.25), but there was no such association for males. Quality of EOL care varies according to sex, with a smaller proportion of females experiencing aggressive EOL care.

Are We Consulting, Sharing Care, or Taking Over? A Conceptual Framework.

Background: Primary- and specialist-level palliative care services are needed. They should work collaboratively and synergistically. Although several service models have been described, these remain open to different interpretations and deployment. Aim: This article describes a conceptual framework, the Consultation-Shared Care-Takeover (C-S-T) Framework, its evolution and its applications. Design: An iterative process informed the development of the Framework. This included a symposium, literature searches, results from three studies, and real-life applications. Results: The C-S-T Framework represents a spectrum anchored by the Consultation model at one end, the Takeover model at the other end, and the Shared Care model in the center. Indicators, divided into five domains, help differentiate one model from the other. The domains are (1) Scope (What aspects of care are addressed by the palliative care clinician?); (2) Prescriber (Who prescribes the treatments?); (3) Communication (What communication occurs between the palliative care clinician and the patient's attending clinician?); (4) Follow-up (Who provides the follow-up visits and what is their frequency?); and (5) Most responsible practitioner (MRP) (Who is identified as MRP?). Each model demonstrates strengths, limitations, uses, and roles. Conclusions: The C-S-T Framework can be used to better describe, understand, assess, and monitor models being used by specialist palliative care teams in their interactions with primary care providers and other specialist services. Large studies are needed to test the application of the Framework on a broader scale in health care systems.

Socioeconomic Status, Palliative Care, and Death at Home Among Patients With Cancer Before and During COVID-19.

Importance: The COVID-19 pandemic had a profound impact on the delivery of cancer care, but less is known about its association with place of death and delivery of specialized palliative care (SPC) and potential disparities in these outcomes. Objective: To evaluate the association of the COVID-19 pandemic with death at home and SPC delivery at the end of life and to examine whether disparities in socioeconomic status exist for these outcomes. Design, Setting, and Participants: In this cohort study, an interrupted time series analysis was conducted using Ontario Cancer Registry data comprising adult patients aged 18 years or older who died with cancer between the pre-COVID-19 (March 16, 2015, to March 15, 2020) and COVID-19 (March 16, 2020, to March 15, 2021) periods. The data analysis was performed between March and November 2023. Exposure: COVID-19-related hospital restrictions starting March 16, 2020. Main Outcomes and Measures: Outcomes were death at home and SPC delivery at the end of life (last 30 days before death). Socioeconomic status was measured using Ontario Marginalization Index area-based material deprivation quintiles, with quintile 1 (Q1) indicating the least deprivation; Q3, intermediate deprivation; and Q5, the most deprivation. Segmented linear regression was used to estimate monthly trends in outcomes before, at the start of, and in the first year of the COVID-19 pandemic. Results: Of 173 915 patients in the study cohort (mean [SD] age, 72.1 [12.5] years; males, 54.1% [95% CI, 53.8%-54.3%]), 83.7% (95% CI, 83.6%-83.9%) died in the pre-COVID-19 period and 16.3% (95% CI, 16.1%-16.4%) died in the COVID-19 period, 54.5% (95% CI, 54.2%-54.7%) died at home during the entire study period, and 57.8% (95% CI, 57.5%-58.0%) received SPC at the end of life. In March 2020, home deaths increased by 8.3% (95% CI, 7.4%-9.1%); however, this increase was less marked in Q5 (6.1%; 95% CI, 4.4%-7.8%) than in Q1 (11.4%; 95% CI, 9.6%-13.2%) and Q3 (10.0%; 95% CI, 9.0%-11.1%). There was a simultaneous decrease of 5.3% (95% CI, -6.3% to -4.4%) in the rate of SPC at the end of life, with no significant difference among quintiles. Patients who received SPC at the end of life (vs no SPC) were more likely to die at home before and during the pandemic. However, there was a larger immediate increase in home deaths among those who received no SPC at the end of life vs those who received SPC (Q1, 17.5% [95% CI, 15.2%-19.8%] vs 7.6% [95% CI, 5.4%-9.7%]; Q3, 12.7% [95% CI, 10.8%-14.5%] vs 9.0% [95% CI, 7.2%-10.7%]). For Q5, the increase in home deaths was significant only for patients who did not receive SPC (13.9% [95% CI, 11.9%-15.8%] vs 1.2% [95% CI, -1.0% to 3.5%]). Conclusions and Relevance: These findings suggest that the COVID-19 pandemic was associated with amplified socioeconomic disparities in death at home and SPC delivery at the end of life. Future research should focus on the mechanisms of these disparities and on developing interventions to ensure equitable and consistent SPC access.

Evaluating equity of access and predictors of minimally invasive hysterectomy for endometrial and cervical cancer from 2000 to 2017 in Ontario, Canada: A population-based cohort study.

INTRODUCTION: We sought to assess the uptake of minimally invasive hysterectomy among patients with endometrial and cervical cancer in Ontario, Canada, and assess the equity of access to minimally invasive surgery (MIS) by evaluating associations with patient, disease, institutional, and provider factors. METHODS: This is a retrospective population-based cohort study of hysterectomy for endometrial and cervical cancer in Ontario (2000-2017). Surgical approach, clinicopathologic, sociodemographic, institutional, and provider factors were identified through administrative databases. Fisher's exact, χ2 , Wilcoxon rank sum, logistic regression, and Cox proportional hazards modeling were used to explore factors associated with MIS. RESULTS: A total of 27 652 patients were included. In total, 6199/24 264 (26%) endometrial and 842/3388 (25%) cervical cancer patients received MIS. The proportion of MIS to open surgeries increased from <0.1% in 2000 to over 55% in 2017 (odds ratio [OR] = 1.31, confidence interval [CI] = 1.28-1.34). Low-income quintile, rurality, low hospital volume, nonacademic hospital, nongynecologic oncology surgeon, and earlier year of surgeon graduation were associated with reduced odds of MIS (OR < 1). CONCLUSIONS: The uptake of MIS hysterectomy increased steadily over the time period. Receipt of MIS is dependent upon multiple social determinants, provider variables, and systems factors. These disparities raise concern for health equity in Ontario and have significant implications for health systems planning and resource allocation.

Education modalities for serious illness communication training: A scoping review on the impact on clinician behavior and patient outcomes.

BACKGROUND: Several clinician training interventions have been developed in the past decade to address serious illness communication. While numerous studies report on clinician attitudes and confidence, little is reported on individual education modalities and their impact on actual behavior change and patient outcomes. AIM: To examine what is known about the education modalities used in serious illness communication training and their impact on clinician behaviors and patient outcomes. DESIGN: A scoping review using the Joanna Briggs Methods Manual for Scoping Reviews was conducted to examine studies measuring clinician behaviors or patient outcomes. DATA SOURCES: Ovid MEDLINE and EMBASE databases were searched for English-language studies published between January 2011 and March 2023. RESULTS: The search identified 1317 articles: 76 met inclusion criteria describing 64 unique interventions. Common education modalities used were: single workshop (n = 29), multiple workshops (n = 11), single workshop with coaching (n = 7), and multiple workshops with coaching (n = 5); though they were inconsistently structured. Studies reporting improved clinician skills tended to be in simulation settings with neither clinical practice nor patient outcomes explored. While some studies reported behavior changes or improved patient outcomes, they did not necessarily confirm improvements in clinician skills. As multiple modalities were commonly used and often embedded within quality improvement initiatives, the impact of individual modalities could not be determined. CONCLUSION: This scoping review of serious illness communication interventions found heterogeneity among education modalities used and limited evidence supporting their effectiveness in impacting patient-centered outcomes and long-term clinician skill acquisition. Well-defined educational modalities and consistent measures of behavior change and standard patient-centered outcomes are needed.

Factors Associated With Adolescent and Young Adult Females Attending a Fertility Consultation Within 30 Days of a Cancer Diagnosis in Ontario, Canada.

Purpose: To examine the factors associated with females attending a fertility consultation within 30 days of cancer diagnosis. Methods: This is a retrospective cohort study, including females, 15 to 39 years of age, diagnosed with cancer in Ontario, Canada. Administrative data were used from the Institute of Clinical and Evaluative Sciences for the period 2006 to 2019. A backward selection multivariate logistic regression was performed, with a primary outcome of fertility consultation within 30 days of diagnosis. Results: A total of 20,556 females were included in the study, with 7% having attended a fertility visit within 30 days of diagnosis. Factors associated with being more likely to attend included: not currently having children (odds ratio [OR] = 4.3; confidence interval [95% CI 3.6-5.1]), later years of diagnosis (OR = 3.2; 95% CI [2.8-3.8]), having undergone chemotherapy (OR = 3.6; 95% CI [3.0-4.3]) or radiation therapy (OR = 1.9; 95% CI [1.6-2.2]), and less marginalization within dependency quintiles (OR 1.4; 95% CI [1.1-1.7]). Having a cancer with lower risk to fertility (OR = 0.3; 95% CI [0.2-0.3]), death within a year of diagnosis (OR = 0.4; 95% CI [0.3-0.6]), and residing in a northern region of Ontario (OR = 0.3; 95% CI [0.2-0.4]) were associated with being less likely to attend. For sociodemographic factors, lower levels of income (OR = 0.5; 95% CI [0.4-0.6]) and marginalization with residential instability (OR = 0.6; 95% CI [0.5-0.8]) were associated with being less likely to attend a fertility consultation. Conclusions: Rates for attendance of female fertility consultations after a cancer diagnosis remain low, with disparities by both clinical and demographic factors.

Implementing palliative care education into primary care practice: a qualitative case study of the CAPACITI pilot program.

BACKGROUND: CAPACITI is a virtual education program that teaches primary care teams how to provide an early palliative approach to care. After piloting its implementation, we conducted an in-depth qualitative study with CAPACITI participants to assess the effectiveness of the components and to understand the challenges and enablers to virtual palliative care education. METHODS: We applied a qualitative case study approach to assess and synthesize three sources of data collected from the teams that participated in CAPACITI: reflection survey data, open text survey data, and focus group transcriptions. We completed a thematic analysis of these responses to gain an understanding of participant experiences with the intervention and its application in practice. RESULTS: The CAPACITI program was completed by 22 primary care teams consisting of 159 participants across Ontario, Canada. Qualitative data was obtained from all teams, including 15 teams that participated in focus groups and 21 teams that provided reflection survey data on CAPACITI content and how it translated into practice. Three major themes arose from cross-analysis of the data: changes in practice derived from involvement in CAPACITI, utility of specific elements of the program, and barriers and challenges to enacting CAPACITI in practice. Importantly, participants reported that the multifaceted approach of CAPACITI was helpful to them building their confidence and competence in applying a palliative approach to care. CONCLUSIONS: Primary care teams perceived the CAPACITI facilitated program as effective towards incorporating palliative care into their practices. CAPACITI warrants further study on a national scale using a randomized trial methodology. Future iterations of CAPACITI need to help mitigate barriers identified by respondents, including team fragmentation and system-based challenges to encourage interprofessional collaboration and knowledge translation.

Intensity of end-of-life care among children with life-threatening conditions: a national population-based observational study.

BACKGROUND: Children with life-threatening conditions frequently experience high intensity care at the end of life, though most of this research only focused on children with cancer. Some research suggests inequities in care provided based on age, disease type, socioeconomic status, and distance that the child lives from a tertiary hospital. We examined: 1) the prevalence of indicators of high intensity end-of-life care (e.g., hospital stays, intensive care unit [ICU] stays, death in ICU, use of cardiopulmonary resuscitation [CPR], use of mechanical ventilation) and 2) the association between demographic and diagnostic factors and each indicator for children with any life-threatening condition in Canada. METHODS: We conducted a population-based retrospective cohort study using linked health administrative data to examine care provided in the last 14, 30, and 90 days of life to children who died between 3 months and 19 years of age from January 1, 2008 to December 31, 2014 from any underlying life-threatening medical condition. Logistic regression was used to model the association between demographic and diagnostic variables and each indicator of high intensity end-of-life care except number of hospital days where negative binomial regression was used. RESULTS: Across 2435 child decedents, the most common diagnoses included neurology (51.1%), oncology (38.0%), and congenital illness (35.9%), with 50.9% of children having diagnoses in three or more categories. In the last 30 days of life, 42.5% (n = 1035) of the children had an ICU stay and 36.1% (n = 880) died in ICU. Children with cancer had lower odds of an ICU stay (OR = 0.47; 95% CI = 0.36-0.62) and ICU death (OR = 0.37; 95%CI = 0.28-0.50) than children with any other diagnoses. Children with 3 or more diagnoses (vs. 1 diagnosis) had higher odds of > 1 hospital stay in the last 30 days of life (OR = 2.08; 95%CI = 1.29-3.35). Living > 400 km (vs < 50 km) from a tertiary pediatric hospital was associated with higher odds of multiple hospitalizations (OR = 2.09; 95%CI = 1.33-3.33). CONCLUSION: High intensity end of life care is prevalent in children who die from life threatening conditions, particularly those with a non-cancer diagnosis. Further research is needed to understand and identify opportunities to enhance care across disease groups.

Characterizing the Palliative Care Physician Workforce: A Retrospective Cross-Sectional Study With Population-Based Data in Ontario, Canada.

OBJECTIVES: Because of an increasing need to build capacity for end-of-life care, improving access to palliative care is a priority. Where a physician practices (eg, hospital, outpatient clinic, home) directly relates to the type of service and the stage of illness at which care is provided. In this study, we describe the physician palliative care specialist workforce and the settings of care within which they practice. DESIGN: A retrospective cohort. SETTING AND PARTICIPANTS: All physicians with palliative care billing codes who were practicing between April 1, 2018, and March 31, 2019, in Ontario, Canada. METHODS: Descriptive statistics of physician billing location and frequency using linked population-based health administrative data. RESULTS: We identified 8883 physicians who provided palliative care during the study period. Of those, 723 (8.1%) were classified as palliative care specialists (>10% of their billings encounters were palliative care). The majority (57.4%) of palliative care specialists worked in 1 setting more than 90% of their time, across home visits (27.1%), indirect care (22.4%), and office (7.9%). There were 61 palliative care specialists practicing in mixed locations who provided home visits, meaning 310 (42.9%) of the palliative care specialists delivered some home-based care. CONCLUSIONS AND IMPLICATIONS: This research provides a comprehensive description of the current palliative care specialist physician workforce that can support efforts to build capacity for high-quality end-of-life care.

Cancer-specific mortality in multiple myeloma: a population-based retrospective cohort study.

Survival has improved in patients diagnosed with multiple myeloma (MM) over the last two decades; however, there remains a paucity of data on the causes of death in MM patients and whether causes of death change during the disease trajectory. We conducted a retrospective population-based study to evaluate the rates of MM-specific versus non-MM cause of death and to identify factors associated with cause-specific death in MM patients, stratified into autologous stem cell transplant (ASCT) and non-ASCT cohorts. A total of 6,677 patients were included, 2,576 in the ASCT group and 4,010 in the non-ASCT group. Eight hundred and seventy-three (34%) ASCT patients and 2,787 (68%) non-ASCT patients died during the follow-up period. MM was the most frequent causes of death, causing 74% of deaths in the ASCT group and 67% in the non-ASCT group. Other cancers were the second leading causes of death, followed by cardiac and infectious diseases. Multivariable analysis demonstrated that a more recent year of diagnosis and novel agent use within 1 year of diagnosis were associated with a decreased risk of MM-specific death, whereas a history of previous non-MM cancer, older age, and the presence of CRAB criteria at diagnosis increased the risk of non-MM death. Our data suggests that despite improvement in MM outcomes in recent years, MM remains the greatest threat to overall survival for patients. Further advances in the development of effective MM therapeutic agents in both ASCT and non-ASCT populations and patient access to them is needed to improve outcomes.

Palliative care utilization across health sectors for patients with gynecologic malignancies in Ontario, Canada from 2006 to 2018.

OBJECTIVE: Early palliative care (PC) is associated with improved patient quality of life, less aggressive end-of-life care, and prolonged survival. We evaluated patterns of PC delivery in gynecologic oncology. METHODS: We conducted a population-based, retrospective cohort study of gynecologic cancer decedents in Ontario from 2006 to 2018 using linked administrative health care data. RESULTS: The cohort included 16,237 decedents; 51.1% died of ovarian cancer, 30.3% uterine cancer, 12.1% cervical cancer, and 6.5% vulvar/vaginal cancers. Palliative care was most often delivered in the hospital inpatient setting in 81%, and 53% received specialist PC. PC was first received during hospital admission in 53%, and by outpatient physician care in only 23%. Palliative care was initiated a median 193 days prior to death, with the lowest two quintiles initiating care ≤70 days before death. The average user of PC resources (third quintile) received 68 days of PC. While cumulative use of community PC gradually increased over the final year of life, institutional palliative care use exponentially rose from 12 weeks until death. On multivariable analyses, predictors of initiating palliative care during a hospital admission included age ≥70 years at death, ≤3 month cancer survival, having cervical or uterine cancer, not having a primary care provider, or being in the lowest 3 income quintiles. CONCLUSION: Most palliative care is initiated and delivered during hospital admission, and is initiated late in a significant proportion. Strategies to increase access to anticipatory and integrated palliative care may improve the quality of the disease course and the end of life.