1Department of Social and Behavioral Sciences, Yale University School of Public Health, New Haven, CT
2Cancer Outcomes, Public Policy, and Effectiveness Research (COPPER) Center, Yale Cancer Center and Yale University School of Medicine, New Haven, CT
3Section of General Internal Medicine, Department of Internal Medicine, Yale University School of Medicine, New Haven, CT
4Department of Chronic Disease Epidemiology, Yale University School of Public Health, New Haven, CT
Dr Gross has received research grants from Pfizer, Medtronic Inc., Johnson & Johnson, and 21st Century Oncology. This investigation was supported by a Pilot Grant and a P30 Cancer Center Support Grant (CCSG), both from Yale Comprehensive Cancer Center.
J Clin Pathways. 2016;2(7):47-54. Received August 1, 2016; accepted August 17, 2016.
The collection of the California cancer incidence data used in this study was supported by the California Department of Public Health as part of the statewide cancer reporting program mandated by California Health and Safety Code Section 103885; the National Cancer Institute’s Surveillance, Epidemiology and End Results Program under contract N01-PC-35136 awarded to the Northern California Cancer Center, contract N01-PC-35139 awarded to the University of Southern California, and contract N02-PC-15105 awarded to the Public Health Institute; and the Centers for Disease Control and Prevention’s National Program of Cancer Registries, under agreement #U55/CCR921930-02 awarded to the Public Health Institute. The authors of this report are responsible for its content. The ideas and opinions expressed herein are those of the author(s) and endorsement by the State of California, Department of Public Health, the National Cancer Institute, and the Centers for Disease Control and Prevention or their Contractors and Subcontractors is not intended nor should be inferred. The authors acknowledge the efforts of the Applied Research Program, NCI; the Office of Research, Development and Information, CMS; Information Management Services (IMS), Inc.; and the Surveillance, Epidemiology, and End Results (SEER) Program tumor registries in the creation of the SEER-Medicare database. The interpretation and reporting of the SEER-Medicare data are the sole responsibility of the authors.
Kathy Doan, MPH, 60 College Street, Suite 432, New Haven, CT 06510, Phone: 203-737-8096, Fax: 203-785-6980, Email: email@example.com
Abstract: Literature suggests that depression has adverse effects on health outcomes among patients with cancer. However, the associations between depression and end-of-life (EOL) cancer care are inconclusive. In this retrospective study, the authors investigated whether depression diagnoses were associated with EOL care intensity among elderly cancer decedents. The Surveillance, Epidemiology, and End Results (SEER)-Medicare database was used to identify 84,947 Medicare beneficiaries aged 66 years or older diagnosed with cancer from 2004 to 2011 who died of cancer within 3 years. Hierarchical generalized linear models were used to evaluate associations between pre- and post-cancer depression and EOL care intensity. Associations between sociodemographic and other health factors with depression diagnoses and with EOL care intensity were also evaluated. The results suggest that depression is associated with lower EOL care intensity and higher hospice utilization among elderly patients with cancer. More discussions among medical providers about cancer and co-occurring mental illness can lead to improved protocols for patients with depression who need appropriate mental health treatment and end-of-life care.
Key Words: cancer, depression, end-of-life care, care intensity
Citation: Journal of Clinical Pathways. 2016;2(7):47-54.
Received August 1, 2016; accepted August 17, 2016.
Currently, end-of-life (EOL) care constitutes a disproportionate amount of Medicare costs, encompassing over one-fourth of the spending on the elderly in the last year of life.1-3 Moreover, overly aggressive EOL care may not be consistent with patient wishes or associated with better health.4-6 Because EOL care is costly and may not always be cost effective, it is important to identify factors that may be associated with overly aggressive EOL care.
There is increasing interest in identifying associations between depression and EOL care intensity for patients with cancer.7,8 Research shows that patients with cancer have a higher risk of depression than the general population.9-13 Indeed, depression is highly prevalent in this population;14,15 between 15% and 50% of patients will experience depressive symptoms.13,15-18 Because depression is among one of the leading causes of disability globally, it has major implications for health outcomes among affected individuals.15 Depression can adversely affect functional status, health-related quality of life, care utilization, and medical costs.19-21 Several studies indicate that co-occurring depression and cancer may affect patient morbidity such that those with depression and cancer are less likely to adhere to treatments and more likely to have other comorbidities and prolonged hospital stays compared with non-depressed patients.7,9,20,22 These factors can lead to increased health care utilization, high expenditures, and aggressive EOL cancer care.6,8,19,21
Studies have examined the relationship between depression and EOL cancer care, but available research is limited in terms of study scope and size9,23-25 or has had poorly defined depression diagnostic criteria.22 One study found that, in adult patients with cancer (≥21 years of age), depression was associated with higher care utilization and expenditures; however, the window of time for the depression diagnosis was omitted.22 Another study examining the prevalence and cumulative expenditures for post-cancer depression in patients with prostate cancer (>65 years of age) suggested that depression was correlated with significant health care utilization, expenses, and mortality.24 Conversely, three studies investigating elderly patients with a certain type of cancer—breast, colorectal, or pancreatic—found that pre-cancer depression was associated with a decrease in EOL care intensity.9,23,25 Therefore, the evidence thus far has been inconclusive. Furthermore, the influence of factors that are likely to affect the intensity of EOL care, including the timing of the patient’s depression (ie, whether diagnosis occurs before or after cancer diagnosis) and the patient’s type of cancer, has not been investigated.
To address this knowledge gap, we investigated whether depression is associated with differences in EOL care intensity patterns in a population-based cohort of Medicare beneficiaries who died from cancer, incorporating a range of cancer types and depression diagnosis criteria. We then analyzed associations between depression and EOL care intensity patterns after controlling for possible confounding variables. We hypothesized that patients with pre- or post-cancer depression would be less likely to receive aggressive EOL care than their non-depressed counterparts. Understanding how mental illness may be correlated with EOL care for cancer patients can help expand knowledge regarding the underlying mechanisms causing these discrepancies, with potential implications for policy and action. Evidence of a relationship between pre- or post-cancer depression and subsequent EOL care intensity for adults with terminal cancer might also identify a select group for whom mental health services may be needed and support earlier communication of palliative care.
For this retrospective cohort analysis, we utilized the Surveillance, Epidemiology, and End Results (SEER)-Medicare database, a unique population-based data source linking Medicare enrollment and claims to cancer registries from the time of a beneficiary’s Medicare eligibility until date of death.26 To date, the SEER cancer registries encompass approximately 30% of the US population, collecting information on cancer incidence and survival.26 We used SEER data to identify patient sociodemographic and cancer tumor characteristics for our sample. Medicare inpatient and outpatient claims were utilized to determine all covered health services and to categorize the depression diagnosis for each participant who died from cancer. The Yale Human Investigation Committee determined this research study did not directly involve human subjects.
We identified elderly Medicare fee-for-service decedents who had breast, prostate, lung, colorectal, pancreatic, liver, kidney, melanoma, or hematological cancer diagnosed during 2004-2011 and who died within 3 years of initial diagnosis as a result of their disease by December 2011. Because Medicare claims were used to identify patients with a pre-cancer depression diagnosis in the year prior to cancer diagnosis, the study population was restricted to patients who were 66 years old or older at cancer diagnosis and continuously enrolled in Medicare Parts A and B for at least 12 months before cancer diagnosis through death. This provided a minimum 12 months of Medicare data to ascertain depression. Patients were excluded from analyses if cancer diagnosis occurred solely through death certificates or autopsy claims, if they lived less than 6 months after cancer diagnosis, or if their income or education by zip code information was unknown.
We identified the presence of depression among participants based on a search for depression diagnoses in all inpatient and outpatient Medicare claims data according to International Classification of Diseases (ICD)-9 diagnostic codes 296.2, 296.3, 296.5, 296.6, 296.7, 298.0, 301.10, 301.12, 301.13, 309.0, 309.1, and 311.9 These include primary and secondary codes for diagnoses of major depressive disorder, bipolar disorder, affective personality disorder, and other medically relevant depressive symptoms. Previous studies on depression and cancer have also used this set of ICD-9 codes.9,23,25 Patients were identified as having a depression diagnosis if they had at least one claim with at least one of these ICD-9 codes.27 We categorized pre-cancer depression as occurring at any point during the 12-month period prior to cancer diagnosis. Individuals with pre-cancer depression can be with or without post-cancer depression. Individuals with post-cancer depression are categorized as only having depression during the 6-month period after cancer diagnosis. Individuals with no depression do not have Medicare claims with the ICD-9 depression codes prior to or after cancer diagnosis.
The sample consisted of 84,947 elderly Medicare beneficiaries with cancer from 2004-2011. Among them, 5072 patients (6.0%) had pre-cancer depression, 6677 patients (7.9%) had post-cancer depression, and 73,198 (86.2%) had no depression.
We evaluated associations between baseline patient sociodemographic characteristics and pre- and post-cancer depression diagnoses. Sociodemographic variables included race, age, gender, marital status, and metropolitan status of residence. Additionally, the SEER-Medicare database allows for census-based estimates of the median household income and the percentage of adults with a high school education or less by patients’ location of residence at the zip code level.
We also evaluated associations between pre- and post-cancer depression diagnoses and other patient health characteristics, including comorbidities, disability status, and tumor characteristics. We used modified Charlson comorbidity conditions created for Medicare claims data to determine the degree of comorbidity per patient.28 This adaptation is based on an approach that requires the diagnostic code to appear on either inpatient or outpatient Medicare claims in the year prior to the cancer diagnosis.23,25,29 Depression was excluded from the Charlson comorbidity index analysis because it is our exposure variable of interest. Patients were categorized as having 0, 1-2, or more than 3 comorbidities. A disability index, which serves as a multivariate claim-based indicator for services commonly needed by older patients with poor or functional performance, was also used to evaluate the presence of disability.30 Tumor characteristics evaluated included tumor site; advanced stage of cancer (yes or no); multiple cancers (yes or no); and duration between cancer diagnosis and death (6 months to 1 year, 1 to 2 years, 2 to 3 years), as reported by the SEER cancer registries.
EOL Care Intensity
To determine EOL care intensity, we searched patients’ Medicare claims data for the presence of 6 EOL care intensity measures, developed by Earle et al:31,32 >1 hospitalization within 30 days of death; >1 emergency department (ED) visit within 30 days of death; ≥1 intensive care unit (ICU) admission within 30 days of death; in-hospital death; any hospice use within 180 days of death; and chemotherapy received within 14 days of death.
Sociodemographic and tumor characteristics of cancer decedents were summarized using means (SD) for continuous variables and percentages for categorical variables. We used t-tests to compare the continuous age variable, and Pearson chi-square tests to compare categorical sociodemographic and tumor characteristics, between decedents with pre-cancer depression and those without depression and between decedents with post-cancer depression and those without depression.
To examine unadjusted associations between depression diagnosis and EOL care intensity measures of interest, we carried out Pearson chi-square tests.
To investigate whether diagnoses of pre- or post-cancer depression might significantly account for variations in EOL care intensity versus cancer decedents without depression, we constructed 2-level hierarchical generalized linear models (HGLMs), clustering patients by hospital referral region, to perform multivariate analyses. Conservative HGLM results were used to examine adjusted associations between depression and EOL care intensity. Adjusted odds ratios (AORs) and 95% confidence intervals (CIs) were estimated after adjustment for confounding variables such as patient demographics, clinical factors, and market factors.
Additionally, we tested for any significant interactions of depression with cancer type, race, and income on EOL care intensity. Statistical significance was set at P < .05. All analyses were performed using SAS version 9.4 (SAS Institute, Inc, Cary, NC).
Depression and Sociodemographic and Health Characteristics
Bivariate associations between patient sociodemographic characteristics and diagnoses of pre- or post-cancer depression versus no depression are shown in Table 1. The mean baseline age of decedents without depression was slightly lower than that of decedents with pre-cancer depression (78.3 vs 78.6 years, P < .05) but slightly higher than that of participants with post-cancer depression (78.3 vs 77.8 years, P < .001). Compared with decedents without depression, those with pre- or post-cancer depression were significantly more likely to be non-Hispanic white, male, and unmarried (P < .001).
Bivariate associations between patient health characteristics and diagnoses of pre- or post-cancer depression versus no depression are shown in Table 2. Decedents with pre- or post-cancer depression tended to have more comorbidities and a worse disability index than those without depression (P < .001). However, decedents with pre- or post-cancer depression were significantly less likely to have multiple cancers than those without depression (P < .001).
Depression and EOL Care Intensity
Unadjusted associations were found between depression diagnosis and EOL care intensity measures (Figure 1). Within our sample of cancer decedents, those with pre- or post-cancer depression were significantly less likely to receive ICU services in the last 30 days or late chemotherapy in the last 14 days of life than those without depression (P < .001). Decedents with pre- or post-cancer depression were also significantly less likely to experience in-hospital death than those without depression (pre-cancer depression, P < .001; post-cancer depression, P = .001). Decedents with pre- or post-cancer depression had lower rates of repeat ED visits (pre-cancer depression, P < .05; post-cancer depression, P = .215) and were significantly more likely to use hospice care in the last 180 days of life than those without depression (P < .001). Compared with decedents without depression, decedents with pre-cancer depression were less likely to have repeat hospitalizations, whereas decedents with post-cancer depression were more likely to have repeat hospitalizations; however, these results were not statistically significant (P = .126 and P = .074, respectively).
Multivariate analyses showed that pre- and post-cancer depression were associated with significantly lower rates of ICU use, in-hospital death, and late chemotherapy (Table 3). Compared with decedents without depression, decedents with pre-cancer depression were significantly less likely to have repeated hospitalizations or ED visits, experience in-hospital death, and receive ICU services or late chemotherapy. However, decedents with pre-cancer depression were more likely to use hospice than decedents without depression. Decedents with post-cancer depression had similar EOL care intensity patterns to those with pre-cancer depression but were marginally more likely to have repeat hospitalizations than their non-depressed counterparts.
No significant interactions of depression with cancer type, race, or income on EOL care intensity were found (P > .05).
Associations between depression diagnoses before or after cancer diagnoses and EOL care intensity among Medicare fee-for-service beneficiaries who died as a result of cancer were examined. Across various EOL care intensity measures, patients with pre- or post-cancer depression were less likely to receive aggressive EOL care and more likely to utilize hospice than patients without depression. These results are consistent with studies that correlated depression with lower odds of receiving definitive treatment among cancer patients.9,23
Our findings build upon previous work in a few important ways. First, our research links depression, be it a pre- or post-cancer depression diagnosis, with EOL care aggressiveness using comprehensive measures. To our knowledge, there have been no studies conducted examining both pre- and post-cancer depression diagnoses within the same study population. Prior research investigated EOL care patterns and expenditures with pre-cancer depression alone or post-cancer depression alone, but these findings were inconsistent. Therefore, our results provide a comprehensive view of EOL care intensity patterns and depression among cancer decedents. Additionally, while our research identifies significant differences in EOL care intensity patterns between participants with pre- or post-cancer depression and those without depression, we did not find important interactions between depression and race or income. In our population-based study, after controlling for these characteristics, which often serve as proxies for socioeconomic mobility and access to care, pre- or post-cancer depression remained a significant predictor of care intensity in decedents.
Instead of concentrating on a specific type of cancer, our research incorporates a range of cancer types, offering an enhanced understanding of the effects of depression on older patients with terminal cancer. Our results differ from studies that demonstrated associations between depression and greater EOL care aggressiveness with higher health care utilization.22,24 Different cancers have distinct prevalence rates of depression, which may impact EOL care intensity. For example, depression prevalence was reported as 38% in patients with pancreatic cancer, 14–40% in patients with breast cancer, and 4.7–33% in patients with lung cancer.22 While we did not find significant interactions between depression and cancer type, these differences are plausible given that the onset of depression may vary by age, gender, or cancer. Additionally, prior studies investigated different cohorts and diagnostic criteria, which do not allow for accurate comparisons. Certain cancers may also lend themselves to different pathways and symptoms that contribute to varying care trends. Further understanding of the social or biological mechanisms that lead to these care patterns may enhance EOL care quality for all cancer patients.
A possible limitation to our study is that accurate diagnosis of patients with depression is frequently underreported in administrative claims data.33 However, the depression rate in our sample (13.9%) falls within the recognized prevalence range of 5–20% for patients with cancer diagnosed with major depressive disorder and depressive symptoms that are comorbid with cancer,16 supporting our findings.
Our study is strengthened by a large sample size and is adequately powered to detect statistically significant differences between cancer patients with pre- or post-cancer depression and those with no depression. However, because our analyses are based on the SEER-Medicare fee-for-service population and our study sample was 82.4% non-Hispanic white, our findings may not characterize all Medicare beneficiaries. Our findings may also not apply to a younger, non-Medicare-based cohort due to possible differences in sociodemographic and tumor characteristics. Finally, because our population-based cohort was comprised of retrospective cancer decedents, participants with pre- or post-cancer depression who did not die during the window of time we designated were not included in our analyses. Prospective research to confirm our findings is necessary.
We did not include patient preferences in our analyses because they are not known from Medicare claims. However, research suggests patient wishes play a role in EOL expenditure variation.6,34 We have reason to believe patient preferences may play some role in these EOL care patterns. In a study examining the effects of pre-cancer depression on older women with breast cancer, researchers arrived at two hypotheses for increased mortality patterns among depressed participants: (1) depression causes the individual to be less capable of functioning properly in society, and (2) depression may be an indicator of global brain dysfunction.23 Because patients with depression may have less social support and are less likely to seek care when ill or to adhere to medical schedules, these factors may contribute to lower rates of survival and health care utilization.9,23
Feelings of hopelessness and vulnerability may influence depressed patients’ EOL care patterns. For example, patients with cancer and their providers may adopt nihilistic attitudes toward their illness due to depression or other negative aspects of their life, which could influence their likelihood of receiving specialized care.9,23,35 Patients with pre- or post-cancer depression may also prefer a quality-enhancing approach that incorporates palliative care options like hospice. This preference aligns with family members’ wishes for hospice to be introduced earlier, with benefits for patients and caregivers.27,36 Our findings indicate that not all depressed patients with cancer are therapeutic minimalists, as they were more likely to utilize hospice. We believe physician behaviors and supply elements also may be driving factors in associations observed between depression and EOL cancer care intensity. Research examining patient experiences, physician behaviors, and supply factors is needed.
Our study results suggest that pre-cancer depression and post-cancer depression are associated with lower EOL care intensity and higher hospice utilization among elderly patients with cancer. While the significantly greater rates of hospice care among participants with pre- or post-cancer depression are promising, our findings that EOL care among patients with depression is less aggressive should be interpreted with caution. Improving physician awareness of the high prevalence of depression comorbid with cancer and establishing accurate screening and treatment for depressed patients can lessen the detrimental influence of depression on cancer health outcomes.9 Earlier communication and initiation of hospice can also increase palliative care rates. More discussions among medical providers about cancer and co-occurring mental illness can lead to improved protocols for patients with depression who need appropriate mental health treatment and EOL care, especially among the elderly patient population. Such actions can minimize unnecessary care utilization among elderly adults with cancer and promote better, more empathetic care that aligns with patient preferences.
1. Hogan C, Lunney J, Gabel J, Lynn J. Medicare beneficiaries’ costs of care in the last year of life. Health Affairs (Millwood). 2001;20(4):188-195.
2. Riley GF, Lubitz JD. Long-term trends in Medicare payments in the last year of life. Health Serv Res. 2010;45(2):565-576.
3. Hogan C. Direct Research, LLC. Spending in the last year of life and the impact of hospice on Medicare outlays. Published June 22, 2015. Accessed August 22, 2016.
4. Morden NE, Chang CH, Jacobson JO, et al. End-of-life care for Medicare beneficiaries with cancer is highly intensive overall and varies widely. Health Aff (Millwood). 2012;31(4):786-796.
5. Wang SY, Hall J, Pollack CE, et al. Trends in end-of-life cancer care in the Medicare program. J Geriatr Oncol. 2016;7(2):116-125.
6. Barnato AE, Herndon MB, Anthony DLG, et al. Are regional variations in end-of-life care intensity explained by patient preferences?: a study of the US Medicare population. Med Care. 2007;45(5):386-393.
7. Meyer F, Fletcher K, Prigerson HG, Braun IM, Maciejewski PK. Advanced cancer as a risk for major depressive episodes. Psychooncology. 2015;24(9):1080-1087.
8. Hinz A, Krauss O, Hauss JP, et al. Anxiety and depression in cancer patients compared with the general population. Eur J Cancer Care (Engl). 2010;19(4):522-529.
9. Boyd CA, Benarroch-Gampel J, Sheffield KM, Han Y, Kuo YF, Riall TS. The effect of depression on stage at diagnosis, treatment, and survival in pancreatic adenocarcinoma. Surgery. 2012;152(3):403-413.
10. Mitchell AJ, Chan M, Bhatti H, et al. Prevalence of depression, anxiety, and adjustment disorder in oncological, haematological, and palliative-care settings: a meta-analysis of 94 interview-based studies. Lancet Oncol. 2011;12(2):160-174.
11. Parpa E, Tsilika E, Gennimata V, Mystakidou K. Elderly cancer patients’ psychopathology: a systematic review: aging and mental health. Arch Gerontol Geriatr. 2015;60(1):9-15.
12. Krishnan KR, Delong M, Kraemer H, et al. Comorbidity of depression with other medical diseases in the elderly. Bio Psychiatry. 2002;52(6):559-588.
13. Honda K, Goodwin RD. Cancer and mental disorders in a national community sample: findings from the national comorbidity survey. Psychother Psychosom. 2004;73(4):235-242.
14. Widera EW, Block SD. Managing grief and depression at the end of life. Am Fam Physician. 2012;86(3):259-264.
15. Massie MJ. Prevalence of depression in patients with cancer. J Natl Cancer Inst Monogr. 2004(32):57-71.
16. Rosenstein DL. Depression and end-of-life care for patients with cancer. Dialogues Clin Neurosci. 2011;13(1):101-108.
17. Weinberger MI, Bruce ML, Roth AJ, Breitbart W, Nelson CJ. Depression and barriers to mental health care in older cancer patients. Int J Geriatr Psychiatry. 2011;26(1):21-26.
18. Pirl WF. Evidence report on the occurrence, assessment, and treatment of depression in cancer patients. J Natl Cancer Inst Monogr. 2004(32):32-39.
19. Crystal S, Sambamoorthi U, Walkup JT, Akincigil A. Diagnosis and treatment of depression in the elderly Medicare population: Predictors, disparities, and trends. J Am Geriatr Soc. 2003;51(12):1718-1728.
20. Walker J, Hansen CH, Martin P, et al. Prevalence, associations, and adequacy of treatment of major depression in patients with cancer: a cross-sectional analysis of routinely collected clinical data. Lancet Psychiatry. 2014;1(5):343-350.
21. Mystakidou K, Parpa E, Tsilika E, et al. Geriatric depression in advanced cancer patients: the effect of cognitive and physical functioning. Geriatr Gerontol Int. 2013;13(2):281-288.
22. Pan X, Sambamoorthi U. Health care expenditures associated with depression in adults with cancer. J Community Support Oncol. 2015;13(7):240-247.
23. Goodwin JS, Zhang DD, Ostir GV. Effect of depression on diagnosis, treatment, and survival of older women with breast cancer. J Am Geriatr Soc. 2004;52(1):106-111.
24. Jayadevappa R, Malkowicz SB, Chhatre S, Johnson JC, Gallo JJ. The burden of depression in prostate cancer. Psychooncology. 2012;21(12):1338-1345.
25. Baillargeon J, Kuo YF, Lin YL, Raji MA, Singh A, Goodwin JS. Effect of mental disorders on diagnosis, treatment, and survival of older adults with colon cancer. J Am Geriatr Soc. 2011;59(7):1268-1273.
26. SEER-Medicare: Brief description of the SEER-Medicare Database. National Institutes of Health Web site. http://healthcaredelivery.cancer.gov/seermedicare/overview/. Updated March 2, 2015. Accessed August 22, 2016.
27. Noyes K, Liu H, Lyness JM, Friedman B. Medicare beneficiaries with depression: Comparing diagnoses in claims data with the results of screening. Psychiatr Serv. 2011;62(10):1159-1166.
28. Klabunde CN, Potosky AL, Legler JM, Warren JL. Development of a comorbidity index using physician claims data. J Clin Epidemiol. 2000;53(12):1258-1267.
29. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40(5):373-383.
30. Davidoff AJ, Zuckerman IH, Pandya N, et al. A novel approach to improve health status measurement in observational claims-based studies of cancer treatment and outcomes. J Geriatr Oncol. 2013;4(2):157-165.
31. Earle CC, Park ER, Lai B, Weeks JC, Ayanian JZ, Block S. Identifying potential indicators of the quality of end-of-life cancer care from administrative data. J Clin Oncol. 2003;21(6):1133-1138.
32. Earle CC, Neville BA, Landrum MB, et al. Evaluating claims-based indicators of the intensity of end-of-life cancer care. Int J Qual Health Care. 2005;17(6):505-509.
33. Spettell CM, Wall TC, Allison J, et al. Identifying physician-recognized depression from administrative data: Consequences for quality measurement. Health Serv Res. 2003;38(4):1081-1102.
34. Baker LC, Bundorf MK, Kessler DP. Patients’ preferences explain a small but significant share of regional variation in medicare spending. Health Aff (Millwood). 2014;33(6):957-963.
35. Periyakoil VS, Hallenbeck J. Identifying and managing preparatory grief and depression at the end of life. Am Fam Physician. 2002;65(5):883-890.
36. Dionne-Odom JN, Azuero A, Lyons KD, et al. Benefits of early versus delayed palliative care to informal family caregivers of patients with advanced cancer: Outcomes from the ENABLE III randomized controlled trial. J Clin Oncol. 2015;33(13):1446-1452.