- Email: [email protected]
Utilization of palliative care and acute care services in older adults with advanced cancer
J O U RN A L OF GE RI A TR IC O N CO LOG Y 7 ( 20 1 6 ) 3 9 –46
Available online at www.sciencedirect.com
ScienceDirect
Utilization of palliative care and acute care services in older adults with advanced cancer Namita Sharmaa,b , Amit M. Sharmaa,b , Martha A. Wojtowyczc , Dongliang Wangc , Ajeet Gajraa,b,⁎ a
SUNY Upstate University, Department of Medicine, Syracuse, NY 13210, USA VA Medical Center, Syracuse, NY, USA c SUNY Upstate University, Department of Public Health and Preventive Medicine, Syracuse, NY 13210, USA b
AR TIC LE I N FO
ABS TR ACT
Article history:
Objectives: There is a gap in knowledge regarding the rates of utilization of palliative care
Received 9 February 2015
services (PCS) and acute care services (ACS) among older patients with advanced cancer
Received in revised form
close to end of life. We analyzed the utilization of these services among older adults
31 October 2015
(65 years and older) and compared them to those in younger adults (40–64 years) with
Accepted 7 December 2015
advanced cancer.
Available online 4 January 2016
Materials and Methods: A retrospective chart review of 567 veterans who died with advanced cancer between 2002 and 2009 and utilized PCS and ACS prior to death was conducted after
Keywords:
IRB approval. To assess PCS utilization, we studied the mean duration between day of
Older patients
hospice referral and time of death (DOR) and the mean length of stay with hospice (LoS).
Cancer
The frequency of emergency room visits (ERVLM), hospital admissions (HALM), and ICU
Palliative care services
admissions (ICULM) in the last month of life was used as a measure for ACS. The differences
Hospice
among older and younger patients were compared using two sample t-tests. Results: Older adults had earlier referral to PCS [mean DOR: 47.3 versus 34.5 days, p = 0.015], longer stay with hospice [mean LoS: 32.5 versus 20.2 days, p = 0.007], fewer hospital [HALM: 0.7 versus 0.9, p = 0.043], and ICU admissions [ICULM: 0.1 versus 0.2, p = 0.030] per patient. The proportion of patients utilizing ER visits [53.5 % versus 59.5%, p = 0.173] and hospital admissions [58.6% versus 65.1%, p = 0.13] in the last month of life was similar in both age groups with fewer older adults utilizing ICU care [13.2% versus 19.5%, p = 0.047]. Conclusion: Older patients with cancer are likely to be referred to PCS earlier than younger patients and spend a longer duration with PCS prior to death. However, there continues to be significant utilization of ACS in all patients with advanced cancer. Better understanding of the goals of care in older adults with cancer and education of oncology providers regarding the need to utilize and integrate palliative care services earlier in the course of disease is imperative. © 2015 Published by Elsevier Ltd.
1. Introduction 1.1. Background ⁎ Corresponding author at: 750 East Adams Street, SUNY Upstate University, Department of Medicine, Syracuse, NY-13210, USA. E-mail address: [email protected] (A. Gajra).
http://dx.doi.org/10.1016/j.jgo.2015.12.004 1879-4068 © 2015 Published by Elsevier Ltd.
Timely utilization of palliative care services (PCS), including hospice care has been shown to improve survival rates in
40
JO U RN A L OF GE RI A TR IC O N COL O G Y 7 (2 0 1 6 ) 3 9 –4 6
patients with advanced cancer and decrease the need for acute medical interventions close to the end of life (EOL).1,2 Early integration of palliative care in the treatment of adults with advanced cancer can lead to measurable improvements in quality of life, less aggressive care at the EOL, improvement in overall survival, as well as longer enrollment in hospice services.3 Family members of patients with home hospice reported higher satisfaction, fewer unmet needs, and a better communication with physicians about medical decision-making.4 The majority of Americans prefer to die at home if they are terminally ill.5 Additionally, it has been noted that the site of death may be associated with perceived quality of life at the end of life. Thus patients who die at home especially with hospice services tend to have a better quality of care than in an institutional setting.6 The role of PCS may be greater in the elderly with advanced cancer. Given that individuals age 65 years and older represent the fastest growing segment of the population in Western countries, there is an urgent need to assess availability, acceptance, and utilization of PCS close to end of life in patients with advanced cancer. With burgeoning health care costs, especially in the United States, and given the high expense incurred close to end of life in patients with advanced cancer, as well as among older adults, this area needs to be studied further. High utilization of acute care settings, including emergency room visits, hospitalizations, and intensive care monitoring, all contribute to these rising costs in addition to taxing the capacity of health care delivery services. In the absence of appropriate use of PCS, greater proportion of adults with advanced cancer rely on acute care services (ACS) at the end of life.7 However, there are limited data regarding the utilization of these services by this specific population of older adults with advanced cancer. A recent study in the Journal of Geriatric Oncology aimed at evaluating the needs in elderly patients affected by cancer and the state of the research in palliative care failed to meet its aim due to a paucity of literature focusing on these issues.8 Hence, to compare the utilization of PCS and ACS in older adults with advanced, incurable cancer, we carried out a retrospective study of adult patients with advanced, incurable cancer at the Syracuse VA medical center (VAMC) with the intent to study the impact of age upon such use. We defined PCS as the referral to and acceptance of palliative care: this was synonymous with formal hospice care in a vast majority of the patients but also included care rendered via a home nursing agency with palliative care goals (without formal hospice designation). The actual PCS could have been rendered in the home setting, at a skilled nursing facility or a hospice unit, as is the case of the VAMC which is housed physically within the hospital itself. Irrespective of the setting, the principles of PCS were adhered to (i.e., focus on maintaining comfort and dignity as well as the discontinuation of all cancer fighting treatments). Information was collected on factors that could potentially impact a patient's decision to accept PCS including demographics (age, gender, and race), tumor factors (type of cancer), and living situation (alone vs accompanied and marital status). It was decided to include patients with solid tumors and lymphoma but not those with leukemia or myeloma since in the latter, the status of remission is not always immediately obvious and these
diseases often require ongoing acute care such as transfusion support and management of infectious complications post chemotherapy. While the prime objective was to study the utilization of PCS and ACS in older adults, we included younger patients from the same time period to compare the differences based on age. Thus, the purpose of this analysis was to assess whether there are differences in patterns of referral, acceptance and ultimately, utilization of PCS based on age. Our hypothesis was that older adults with cancer will likely be referred to PCS earlier than younger patients, accept PCS more often and spend greater time with PCS prior to death than younger patients. As a corollary, older adults will use ACS with decreased frequency in the last month of life compared to younger patients.
2. Methods 2.1. Objectives 2.1.1 To study the utilization of PCS at our institution for adults who died as a consequence of advanced cancer by calculating the mean duration between day of hospice referral and time of death (DoR) and the mean length of stay with hospice (LoS) for each age group. 2.1.2 To compare the utilization of PCS between older (≥ 65 years) and younger adults (40–64 years) with advanced cancer. 2.1.3 To study rates of ACS utilization in the last month of life between older and younger patients by identifying Emergency room visits in the last month of life (ERVLM), Hospital admissions in the last month of life (HALM), and Intensive care unit admissions in the last month of life (ICULM). 2.1.4 To study the association of living situation and marital status with duration of PCS.
2.2. Setting and Population The study setting was Syracuse Veterans Affairs Medical Center (SVAMC). All patients who died as a consequence of advanced cancer at the SVAMC from 2002 to 2009 were included in the cohort. These patients were identified from the VAMC tumor registry. At the SVAMC, all patients with advanced cancer are referred to an “advanced illness care coordinator”, a function served by a licensed social worker. Patients who are referred have the right to decline PCS until they are psychologically ready to make such a transition. Data regarding date of referral, acceptance of PCS and death were collected from the notes entered by the palliative service in the electronic medical record. Admissions to the hospital and emergency room visits were collected from the patient chart. Thus, inclusion criteria as defined for this analysis include: patients (age more than 40 years or less than 100 years) with advanced cancer who died during the stated period, who were referred to PCS irrespective of whether they accepted PCS or continued with acute care for their cancer. The patients who presented with acute illness and early
41
J O U RN A L OF GE RI A TR IC O N CO LOG Y 7 ( 20 1 6 ) 3 9 –46
demise even prior to a referral to PCS were not included. Patients below 40 or above 100 years of age and those with non-metastatic disease were excluded since younger patients and the rare centenarians may have goals different than those of the typical younger and older patients, respectively. For reasons stated in the previous section, patients with leukemia and myeloma were excluded.
2.3. Data Collection After initial IRB approval, a retrospective electronic medical records chart review of 567 patients identified by the tumor registry was performed. We compared 372 older patients (≥65 years) with 195 younger (40–64 years old) adults.
2.4. Measures Independent variables identified were: age of the patient at the time of diagnosis of advanced cancer, race of the patient (Caucasian, African-American and others), marital status of the subjects (divorced, married, single, widowed and unknown), living condition of the patient at the time of diagnosis (home with family, home alone and nursing home), location of the cancer (head & neck, lung & pleura, gastrointestinal including liver & pancreas, genitourinary including bladder & prostate, lymph nodes and others). Utilization of PCS was identified as mean number of days of referral to hospice before death (DOR) and mean length of stay in hospice before death (LoS). Utilization of hospital resources: ACS was measured in terms of ER visits (ERVLM), hospital admissions (HALM) and ICU admissions (ICULM) in the last month of life.
2.5. Statistical Analysis Numeric variables were summarized using simple descriptive statistics such as the number of observations, mean and standard deviation. Frequencies and relative frequencies were computed for all categorical variables. The utilization of PCS and ACS between younger and older groups was compared using both unadjusted two-sample tests and multiple regression analyses adjusting the potential confounding effects of other factors. For DOR and LoS, two sample t-tests were performed and Satterthwaite approximation was used if the variances were statistically different between the two groups. For ERVLM, HALM and ICULM, two sample tests for Poisson rate were performed. Multiple linear regressions (for DOR and LoS) and multiple Poisson regressions (for ERVLM, HALM and ICULM) were also performed in order to adjust the impact of cancer location (lung vs gastrointestinal vs head & neck vs others), marital status (married vs others) and living condition (home with family vs others). Overdispersion in the Poisson regression was adjusted by Pearson chi-square statistic. The absolute percentages of patients who utilized acute care (ER visits, hospital and ICU admissions) in the last month of life and who died in the hospital were compared by Pearson chi-square tests, if less than 50% of the cells have expected counts less than 5, or Fisher's exact test otherwise. Furthermore, multivariate linear regression for the utilization of PCS and ACS and multivariate logistic regression for the absolute percentages of patients who utilized acute care were performed in order to adjust the impact of location of cancer (lung vs gastrointestinal vs head & neck vs others) and social support, including marital status (married vs others) and living situation (home with family vs others).
Table 1 – Sample characteristics. Variable
Age
Race
Sex Location of cancer
Marital status
Living condition
Mean (std) Median (min, max) White African American Others Male Lung & pleura Gastrointestinal Head & neck Lymph nodes Genito-urinary Others Married Divorced Widowed Single Unknown Home with family Home alone Nursing home
Young group (40–64) N = 195
Old group (65 and above) N = 372
p-Value[a]
57.7 (4.76) 58 (44, 64) 177 (90.8%) 15 (7.7%) 3 (1.5%) 187 (95.9%) 92 (47.2%) 46 (23.6%) 22 (11.3%) 13 (6.7%) 8 (4.1%) 14 (7.2%) 70 (35.9%) 78 (40.0%) 9 (4.6%) 38 (19.5%)
75.8 (6.69) 75 (65, 100) 358 (96.2%) 13 (3.5%) 1 (0.3%) 367 (98.7%) 182 (48.9%) 84 (22.6%) 37 (9.9%) 16 (4.3%) 38 (10.2%) 15 (4.0%) 173 (46.5%) 85 (22.8%) 77 (20.7%) 33 (8.9%) 4 (1.1%) 207 (55.6%) 140 (37.6%) 25 (6.7%)
<0.001
104 (53.3%) 86 (44.1%) 5 (2.6%)
0.061
0.037 0.072
<0.0001
0.059
[a] p values are calculated from Pearson chi-square tests, if less than 50% of the cells have expected counts less than 5; otherwise Fisher's exact test is used.
42
JO U RN A L OF GE RI A TR IC O N COL O G Y 7 (2 0 1 6 ) 3 9 –4 6
Fig. 1 – Distribution of age: Histograms with density plots (top) and boxplots (bottom) of age by group. The bold line represents a normal density estimate and the dashed line represents a kernel density estimate.
3. Results
the older group). More than half of the patients across both the groups were living at home with family.
3.1. Sample Characteristics
3.2. Utilization of Palliative Care Services
Out of the 567 patients who met the eligibility criteria, 195 patients were in the younger group and 372 in the older group (Table 1, Fig. 1). More than 90% of the sample comprised of Caucasian males. Lung cancer was the most commonly encountered tumor type (47% in the young group and 48% in
Earlier referral to PCS was observed in the older group [mean DOR: 47.3 days versus 34.5 days, p = 0.015]. Older patients spent more time in hospice [mean LoS: 32.5 days versus 20.2 days, p = 0.007] than the younger group. There were 65 patients (16.9%) in the older group and 35 (17.9%) in the
Table 2 – Comparison of outcomes between two groups. Variable
DOR LoS ERVLM HALM ICULM LoS[b]
Mean STD Mean STD Mean STD Mean STD Mean STD N Mean STD
days days number of visits number of admissions number of admissions
Young group (40–64) N = 195
Old group (65 and above) N = 372
p-Value ⁎ (p-Value &)
34.5 51.98 20.2 43.25 0.8 0.83 0.86 0.76 0.22 0.47 160 24.7 46.61
47.3 70.88 32.5 63.67 0.7 0.79 0.72 0.72 0.14 0.36 307 39.1 68.03
0.015 (0.027) 0.007 (0.013) 0.138 (0.119) 0.043 (0.035) 0.030 (0.023) 0.016 (0.012)
Abbreviations: DOR, the mean duration between day of hospice referral and time of death; LoS, length of stay; ERVLM, emergency room visits in the last month of life; HALM, hospital admissions in the last month of life; ICULM, intensive care unit admissions in the last month of life; STD, standard deviation. [b] 0 and missing values excluded. The 0 values are considered as missing values and excluded from analysis, assuming that all 0 values are as a result that the patient refused hospice. Including the 0 values in the analysis will underestimate the true value. ⁎ p Values from univariate two-sample tests. & p Values from multivariate regression analyses.
J O U RN A L OF GE RI A TR IC O N CO LOG Y 7 ( 20 1 6 ) 3 9 –46
younger group who did not accept transition to PCS and for this analysis, they were considered to have LoS of 0 days. When these patients were excluded from LoS analysis, the difference between two groups continued to be significant [mean LoS: 39.1 versus 24.7 days, p = 0.016] (Table 2). We also conducted multivariate linear regression analysis that adjusted for type of cancer as well as marital status and living situation. These factors were found to not be significantly associated with the outcomes variables and did not confound the association between age and outcome variables. After the adjustment, mean DOR and mean LoS for older patients are 13.5 days and 12.8 days more than younger patients, respectively. When the patients who did not accept PCS were excluded from LoS analysis, mean LoS for older patients are 15.2 days longer.
3.3. Utilization of Acute Care Services There were fewer hospital admissions [HALM: 0.72 versus 0.86, p = 0.043] and ICU admissions [ICULM: 0.14 versus 0.22, p = 0.030] per patient in the last month of life in older adults when compared to younger adults (Table 2). There was no significant difference between the numbers of ER visits for older versus younger patients (p = 0.138). After the adjustment, mean HALM, mean ICULM and mean ERVLM for younger veterans are 0.18 days (p = 0.035), 0.48 days (p = 0.023) and 0.14 visits (p = 0.119) more than older veterans, respectively. However, the absolute percentage of patients utilizing acute care (ER visits and hospital admissions) in last month of life was similar in both groups with fewer older adults receiving ICU care (Table 3). We conducted multivariate logistic regression analysis that adjusted for type of cancer, marital status and living situation. These factors were not statistically significantly associated with the outcomes variables and did not confound the association between age and outcome variables. After the adjustment, the odds ratio between younger and older patients with respects to death in hospital, having at least one ER visit, at least one hospital admission and at least one ICU visit were 0.770 (95% CI: 0.470–1.261), 0.771 (95% CI: 0.540–1.101), 0.762 (95% CI: 0.530–1.096) and 0.612 (95% CI: 0.383–0.980) of the odds for older patients, respectively. Table 3 – Utilization of ACS by two groups. Variable Young group Old group (40–64) (65 and above) N = 195 N = 372 Death In hospital ERVLM
31 (15.9%)
49 (13.2%)
116 (59.5%)
199 (53.5%)
HALM
127 (65.1%)
218 (58.6%)
ICULM
38 (19.5%)
49 (13.2%)
OR, 95%CI, p-value ⁎ (p-value &) 0.770, 0.470 to 1.261, 0.376 (0.300) 0.771, 0.540 to 1.101, 0.173 (0.152) 0.762, 0.530 to 1.096, 0.130 (0.142) 0.612, 0.383 to 0.980, 0.047 (0.041)
Abbreviations: ERVLM, emergency room visits in the last month of life; HALM, hospital admissions in the last month of life; ICULM, intensive care unit admissions in the last month of life. ⁎ p Values are calculated from Pearson chi-square tests. & p Values, adjusted OR and 95% CI from multivariate logistic regression analyses.
43
4. Discussion In this study, we found that older patients were referred to PCS earlier and spent a longer period with PCS compared to the younger cohort. However, a similar proportion of older and younger patients (17% and 18%, respectively) declined PCS. Further, more than half of older patients continued to have ER visits and hospital admissions in the last month of life and this proportion was not remarkably different than younger patients.
4.1. PCS Utilization Some unique features of this retrospective review include the fact that we were able to assess mean duration of referral to hospice services which may convey the intent of the medical provider to initiate formal PCS. While it may appear that DOR and LoS are inter-dependent, these variables can be discordant. Patients were offered to transition to PCS via a referral but may elect not to accept PCS for a while. In fact, 17% of the older and 18% of the younger patients did not accept such a transition and hence their LoS is considered a zero for statistical purposes albeit not the DOR. However, we did not find discordance between these variables likely due to the fact that similar proportion of older and younger patients declined a transition to PCS. Older adults with advanced cancer were referred to PCS earlier and spend a longer duration with hospice than younger patients with similar diagnoses. While this may potentially raise a concern for age bias, given that ultimately all these patients succumbed to their disease, a longer duration with PCS would reflect appropriate clinical practices in patients close to end of life. This is a move in the right direction especially since older adults may have goals which differ from younger patients, such as maintaining QoL with lower acceptance of therapy associated toxicity and less focus on prolongation of life. In a report from a US-based community based cancer center, there were no age differences among patients referred to hospice versus those that were not referred to hospice prior to death.9 This study was similar to ours in that it includes patients with a variety of cancers encountered in the community setting. However, it did not delve into the age-based differences in outcomes as attempted in our analysis. Investigators from Asia have reported similar associations between older age and a longer length of stay with PCS.10,11 In another US based retrospective review utilizing two large population databases, fewer patients with advanced lung cancer under age 65 received hospice services as compared to patients above 65 years of age.12 In a Canadian study limited to patients with advanced pancreatic cancer, earlier PCS consultation was associated with less aggressive care near death: decreased use of chemotherapy (odds ratio [OR]: 0.34, 95% confidence interval [CI]: 0.25–0.46); lower risk of ICU admission (OR: 0.12, 95% CI: 0.08–0.18); multiple ED visits (OR: 0.19, 95% CI: 0.16–0.23); multiple hospitalizations near death (OR: 0.24, 95% CI: 0.19 to 0.31).13 In another retrospective study limited to patients with ovarian cancer, older women received hospice care for a longer duration over a 10 year study period. However, it was also observed that these women had statistically significant increases in
44
JO U RN A L OF GE RI A TR IC O N COL O G Y 7 (2 0 1 6 ) 3 9 –4 6
intensive care unit admissions, hospitalizations, repeated emergency department visits, and health care transitions (all P ≤ .01). Inpatients referred to hospice were more likely to enroll in hospice within 3 days of death than outpatients 14 (adjusted odds ratio: 1.36; 95% CI: 1.12–1.66). A recent report found that predictors of a short hospice stay (defined as ≤ 3 days) include hematologic malignancies, male gen15 der, married status and younger age (< 65 years). While the former two do not apply to our cohort, these findings are similar in terms of age. We did not find marital status to be associated with LoS with PCS in our population. This may be a consequence of lack of representation of women in the sample. Overall, our findings are similar to the above studies in that older patients were referred to PCS earlier and spent longer duration with PCS compared to younger patients. Unlike many of these studies which are limited to patients with single cancer type, our study included patients with variety of advanced solid tumors.
4.2. ACS Utilization Use of ACS is associated with enhanced costs close to end of life and the timely use of PCS can limit the costs incurred by the health care system. A recent report of Medicare beneficiaries compared ACS utilization and costs of care in patients with poor prognosis cancer enrolled with hospice versus a matched cohort that died without hospice. In patients utilizing hospice, hospitalizations, hospital admissions, invasive procedures were all reduced significantly leading to a diminished cost of care versus those that were not enrolled to hospice.16 A variable which is often not reported upon, even in large population database analyses, are the emergency room visits. Emergency room visits can add significant burden to the healthcare system and can play a role in compromising the perceived quality of life for the patients and their caregivers. In our cohort, older patients received less aggressive care close to EOL including fewer ER visits, hospital admissions, ICU admissions as compared to younger patients. However, more than half of older adults with advanced cancer required ER visits and hospital admissions in the last month of life. The high rates of ACS utilization (ER visits and hospital admission) in > 50% of patients with relatively short LOS with hospice suggest that PCS could have been initiated sooner. The rates of ACS utilization indicate that the treatment team needs to better anticipate a transition to palliative care. In a study by Mack et al., that included only patients with advanced lung and colon cancer, nearly half of patients received at least one marker of aggressive EOL care, including chemotherapy in the last 14 days of life (16%), intensive care unit care in the last 30 days of life (9%), and acute hospital-based care in the last 30 days of life (40%) with no association with age noted in that study.7 The latter rates in our cohort are somewhat higher for the older patients at 13% and 59% for ICULM and HALM respectively suggesting room for improvement. In another retrospective review of the Ontario registry, factors associated with aggressive care close to EOL included male gender, younger age as well as tumor type (lung, breast and hematological malignancies).17 In our cohort, we did not find
any association with tumor type and use of ACS in the last month of life. We are not able to study the association between gender and ACS utilization due to preponderance of males in our cohort. In a recently reported pooled analysis of a retrospective cohort study, the use of community based palliative care teams reduced the number of ER visits and hospital deaths at the end of life.18 Such community based palliative care teams may be effective in reducing ACS utilization at the EOL.
4.3. Barriers Accessing PCS Qualitative evidence in the literature highlighted numerous barriers to accessing palliative care. Some barriers are a consequence of the lack of education of physicians and other health care professionals, regarding end of life and its management. Heath care providers' personal and religious beliefs may also impact their view on this matter. Clear and honest communication between the health care provider and the patient is essential for patients to understand their prognosis and accept a transition to PCS at the appropriate time. In a retrospective study looking at patients with metastatic lung cancer, around half of patients had not discussed hospice with a provider within 4 to 7 months after diagnosis.19 Other barriers are the result of patient/family misinformation and aspects of the hospice system that limit the number and types of patients deemed appropriate for hospice care.20 As per a recent study, many patients do not understand that the chemotherapy in a metastatic cancer is only palliative and not curative.21 Better understanding about the goals of treatment would probably allow them to choose PCS over chemotherapy. In another study, patients and caregivers found that palliative care was often provided sub-optimally during the “out-of-hours” that is weekends and public holidays.22 It was perceived by caregivers that the “service responses” during the “out-of-hours” were designed for acute medical rather than palliative care needs. These systemic issues in the adequate delivery of PCS are vital for such services to be more accessible and hence, acceptable to the patient and the family. Good anticipatory care with the timely provision of information to patients and caregivers so as to set the right expectations is therefore critical for the success of PCS.22 In a recent retrospective review of over 100,000 patients with cancer, the cost of medical care in the last month of life was 43% higher in those receiving aggressive care compared to those receiving non-aggressive care ($18,131 vs $12,678, P < .0001).23 Communicating these and similar data to healthcare providers and possibly to patients and advocacy groups may facilitate the wider acceptance of PCS close to EOL.
4.4. Study Limitations & Scope for Future Work Limitations of our study would be retrospective chart review, small sample size, single center data, predominantly male patients and single payer insurance system. Most of this study was completed before the impact of Affordable Care Act (ACA) of 2010 which gave several incentives to health care providers for initiating end-of-life discussions with their patients. It would be interesting to study health outcomes and physician
J O U RN A L OF GE RI A TR IC O N CO LOG Y 7 ( 20 1 6 ) 3 9 –46
perceptions before and after the advent of ACA. Secondly, multi-center data could be accessed to compare our results with the national average and advocate for policy changes based on the results. Further, we did not look specifically at the subset of patients who declined palliative care altogether. So, the impact upon their utilization of ACS in the last month cannot be discerned.
4. 5.
5. Conclusion 6.
Older patients with cancer are likely to be referred to PCS earlier and spend a longer duration with PCS. However, with more than half of older adults with cancer reporting ER visits and admissions in the last month of life, there continues to be significant utilization of ACS by this population. Earlier end of life discussions and referral to PCS in patients with advanced cancer and limited life expectancy can decrease the use of aggressive care, increase the use of hospice services and improve the quality of life during the last days of their life. Physician–patient communication is essential for patients to understand their prognosis and the concept of hospice, which can ultimately lead to an increase in the proportion of patients accepting hospice. Better understanding of the goals of care in older adults and education of oncology providers regarding the need to integrate PCS earlier in the course of disease is imperative.
Disclosures and Conflict of Interest Statements Study concepts: A. Gajra, N. Sharma. Study design: A. Gajra, N. Sharma, A.M. Sharma. Data acquisition: N. Sharma, A.M. Sharma. Quality control of data and algorithms: A. Gajra, N. Sharma, A.M. Sharma. Data analysis and interpretation: A. Gajra, N. Sharma, A.M. Sharma. Statistical analysis: D. Wang, M.A. Wojtowycz. Manuscript preparation: N. Sharma, A.M. Sharma. Manuscript editing: A. Gajra Manuscript review: A. Gajra.
REFERENCES
1. Keating Nancy L, Landrum Mary Beth, Lamont Elizabeth B, Earle Craig C, Bozeman Samuel R, Mcneil Barbara J. End-of-life care for older cancer patients in the veterans health administration versus the private sector. Cancer 2010: 3732–3739. 2. Gonsalves Wilson I, Tashi Tsewang, Krishnamurthy Jairam, Davies Tracy, Ortman Stephanie, Thota Ramya, Aldoss Ibrahim, Ganta Ashwin, Kalaiah Mudappa, Didwaniya Neha, Eberle Catherine, Ganti Apar K, Silberstein Peter T, Subbiah Shanmuga. Effect of Palliative Care Services on the Aggressiveness of End-of-Life Care in the Veteran's Affairs Cancer Population. J. Palliat Med. 2011;14(11):1231–1235. 3. Temel Jennifer S, Greer Joseph A, Muzikansky Alona, Gallagher Emily R, Admane Sonal, Jackson Vicki A, Dahlin
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
45
Constance M, Blinderman Craig D, Jacobsen Juliet, Pirl William F, Andrew Billings J, Lynch Thomas J. Early palliative care for patients with metastatic non-small-cell lung cancer. N Engl J Med 2010:733–742. Teno JM. Family perspectives on end-of-life care at the last place of care. JAMA 2004;1(291):88–93. Goodman David, Morden Nancy, Chang Chiang-Hua, Fisher Elliott, Wennberg John, Bronner Kristina. Trends in Cancer Care Near the End of Life A Dartmouth Atlas of Health Care Brief. The Dartmouth Atlas of Health Care; September 4 2013 [Accessed August 24, 2014. , http://www.dartmouthatlas.org/ downloads/reports/Cancer_brief_090413.pdf]. Teno JM, Gozalo PL, Bynum JP, Leland NE, Miller SC, Morden NE, Scupp T, Goodman DC, Mor V. Change in end-of-life care for Medicare beneficiaries: site of death, place of care, and health care transitions in 2000, 2005, and 2009. JAMA Feb 6 2013;309(5):470–477 [PMID: 23385273]. Mack JW, Cronin A, Keating NL, Taback N, Huskamp HA, Malin JL, Earle CC, Weeks JC. Associations between End-of-life discussion characteristics and care received near death: a prospective cohort study. J Clin Oncol 2012:4387–4395. Brighi N, Balducci L, Biasco G. Cancer in the elderly: is it time for palliative care in geriatric oncology? J Geriatr Oncol 2014: 197–203. Cowall DE, Yu BW, Heineken SL, Lewis EN, Chaudhry V, Daugherty JM. End-of-life care at a community cancer center. J Oncol Pract Jul 2012;8(4):e40–e44, http://dx.doi.org/10.1200/JOP.2011.000451 [Epub 2012 May 22. PubMed PMID: 23180997; PubMed Central PMCID: PMC3396828]. Chang TS, Su YC, Lee CC. Determinants for aggressive end-of-life care for oral cancer patients: a population-based study in an Asian country. Medicine (Baltimore) Jan 2015;94(4): e460, http://dx.doi.org/10.1097/MD.0000000000000460 [PubMed PMID: 25634186]. Tang ST, Liu TW, Shyu YI, Huang EW, Koong SL, Hsiao SC. Impact of age on end-of-life care for adult Taiwanese cancer decedents, 2001–2006. Palliat Med Jan 2012;26(1):80–88, http: //dx.doi.org/10.1177/0269216311406989 [Epub 2011 May 23. PubMed PMID: 21606128]. Mack JW, Chen K, Boscoe FP, Gesten FC, Roohan PJ, Weeks JC, Schymura MJ, Schrag D. Underuse of hospice care by Medicaid-insured patients with stage IV lung cancer in New York and California. J Clin Oncol Jul 10 2013;31(20):2569–2579, http://dx.doi.org/10.1200/JCO.2012.45.9271 [Epub 2013 Jun 3. PubMed PMID: 23733768; PubMed Central PMCID: PMC3699723]. Jang RW, Krzyzanowska MK, Zimmermann C, Taback N, Alibhai SM. Palliative care and the aggressiveness of end-of-life care in patients with advanced pancreatic cancer. J Natl Cancer Inst Jan 20 2015;107(3), http://dx.doi.org/10.1093/jnci/dju424 [pii: dju424., Print 2015 Mar. PubMed PMID: 25609233]. Wright AA, Hatfield LA, Earle CC, Keating NL. End-of-life care for older patients with ovarian cancer is intensive despite high rates of hospice use. J Clin Oncol Nov 1 2014;32(31): 3534–3539, http://dx.doi.org/10.1200/JCO.2014.55.5383 [Epub 2014 Oct 6. PubMed PMID: 25287831; PubMed Central PMCID: PMC4209104]. O'Connor NR, Hu R, Harris PS, Ache K, Casarett DJ. Hospice admissions for cancer in the final days of life: independent predictors and implications for quality measures. J Clin Oncol Oct 1 2014;32(28):3184–3189, http://dx.doi.org/10.1200/JCO.2014.55.8817 [Epub 2014 Aug 25. PubMed PMID: 25154824; PubMed Central PMCID: PMC4171361]. Obermeyer Z, Makar M, Abujaber S, Dominici F, Block S, Cutler DM. Association between the Medicare hospice benefit and health care utilization and costs for patients with poor-prognosis cancer. JAMA Nov 12 2014;312(18):1888–1896, http://dx.doi.org/10.1001/jama.2014.14950 [PubMed PMID: 25387186; PubMed Central PMCID: PMC4274169].
46
JO U RN A L OF GE RI A TR IC O N COL O G Y 7 (2 0 1 6 ) 3 9 –4 6
17. Ho TH, Barbera L, Saskin R, Lu H, Neville BA, Earle CC. Trends in the aggressiveness of end-of-life cancer care in the universal health care system of Ontario, Canada. J Clin Oncol Apr 20 2011;29(12):1587–1591, http://dx.doi.org/10.1200/JCO.2010.31.9897 [Epub 2011 Mar 14. PubMed PMID: 21402603; PubMed Central PMCID: PMC3082976]. 18. Seow H, Brazil K, Sussman J, Pereira J, Marshall D, Austin PC, Husain A, Rangrej J, Barbera L. Impact of community based, specialist palliative care teams on hospitalisations and emergency department visits late in life and hospital deaths: a pooled analysis. BMJ Jun 6 2014;348:g3496, http://dx.doi.org/ 10.1136/bmj.g3496 [PubMed PMID: 24906901; PubMed Central PMCID: PMC4048125]. 19. Huskamp HA, Keating NL, Malin JL, Zaslavsky AM, Weeks JC, Earle CC, Teno JM, Virnig BA, Kahn KL, He Y, Ayanian JZ. Discussions with physicians about hospice among patients with metastatic lung cancer. Arch Intern Med May 25 2009;169(10): 954–962, http://dx.doi.org/10.1001/archinternmed.2009.127 [PubMed PMID: 19468089; PubMed Central PMCID: PMC2689617].
20. Friedman Betsy T, Harwood M Kay, Shields Matthew. Barriers and enablers to hospice referrals: an expert overview. J Palliat Med 2002:73–84. 21. Weeks Jane C, Catalano Paul J, Cronin Angel, Finkelman Matthew D, Mack Jennifer W, Keating Nancy L, Schrag Deborah. Patients' expectations about effects of chemotherapy for advanced cancer. N Engl J Med 2012:1616–1625. 22. Worth A, Boyd K, Kendall M, Heaney D, Macleod U, Cormie P, Hockley J, Murray S. Out-of-hours palliative care: a qualitative study of cancer patients, carers and professionals. Br J Gen Pract Jan 2006;56(522):6–13 [PMID: 16438809]. 23. Cheung MC, Earle CC, Rangrej J, Ho TH, Liu N, Barbera L, Saskin R, Porter J, Seung SJ, Mittmann N. Impact of aggressive management and palliative care on cancer costs in the final month of life. Cancer May 29 2015, http://dx.doi.org/10.1002/cncr.29485 [[Epub ahead of print] PubMed PMID: 26031241].
Available online at www.sciencedirect.com
ScienceDirect
Utilization of palliative care and acute care services in older adults with advanced cancer Namita Sharmaa,b , Amit M. Sharmaa,b , Martha A. Wojtowyczc , Dongliang Wangc , Ajeet Gajraa,b,⁎ a
SUNY Upstate University, Department of Medicine, Syracuse, NY 13210, USA VA Medical Center, Syracuse, NY, USA c SUNY Upstate University, Department of Public Health and Preventive Medicine, Syracuse, NY 13210, USA b
AR TIC LE I N FO
ABS TR ACT
Article history:
Objectives: There is a gap in knowledge regarding the rates of utilization of palliative care
Received 9 February 2015
services (PCS) and acute care services (ACS) among older patients with advanced cancer
Received in revised form
close to end of life. We analyzed the utilization of these services among older adults
31 October 2015
(65 years and older) and compared them to those in younger adults (40–64 years) with
Accepted 7 December 2015
advanced cancer.
Available online 4 January 2016
Materials and Methods: A retrospective chart review of 567 veterans who died with advanced cancer between 2002 and 2009 and utilized PCS and ACS prior to death was conducted after
Keywords:
IRB approval. To assess PCS utilization, we studied the mean duration between day of
Older patients
hospice referral and time of death (DOR) and the mean length of stay with hospice (LoS).
Cancer
The frequency of emergency room visits (ERVLM), hospital admissions (HALM), and ICU
Palliative care services
admissions (ICULM) in the last month of life was used as a measure for ACS. The differences
Hospice
among older and younger patients were compared using two sample t-tests. Results: Older adults had earlier referral to PCS [mean DOR: 47.3 versus 34.5 days, p = 0.015], longer stay with hospice [mean LoS: 32.5 versus 20.2 days, p = 0.007], fewer hospital [HALM: 0.7 versus 0.9, p = 0.043], and ICU admissions [ICULM: 0.1 versus 0.2, p = 0.030] per patient. The proportion of patients utilizing ER visits [53.5 % versus 59.5%, p = 0.173] and hospital admissions [58.6% versus 65.1%, p = 0.13] in the last month of life was similar in both age groups with fewer older adults utilizing ICU care [13.2% versus 19.5%, p = 0.047]. Conclusion: Older patients with cancer are likely to be referred to PCS earlier than younger patients and spend a longer duration with PCS prior to death. However, there continues to be significant utilization of ACS in all patients with advanced cancer. Better understanding of the goals of care in older adults with cancer and education of oncology providers regarding the need to utilize and integrate palliative care services earlier in the course of disease is imperative. © 2015 Published by Elsevier Ltd.
1. Introduction 1.1. Background ⁎ Corresponding author at: 750 East Adams Street, SUNY Upstate University, Department of Medicine, Syracuse, NY-13210, USA. E-mail address: [email protected] (A. Gajra).
http://dx.doi.org/10.1016/j.jgo.2015.12.004 1879-4068 © 2015 Published by Elsevier Ltd.
Timely utilization of palliative care services (PCS), including hospice care has been shown to improve survival rates in
40
JO U RN A L OF GE RI A TR IC O N COL O G Y 7 (2 0 1 6 ) 3 9 –4 6
patients with advanced cancer and decrease the need for acute medical interventions close to the end of life (EOL).1,2 Early integration of palliative care in the treatment of adults with advanced cancer can lead to measurable improvements in quality of life, less aggressive care at the EOL, improvement in overall survival, as well as longer enrollment in hospice services.3 Family members of patients with home hospice reported higher satisfaction, fewer unmet needs, and a better communication with physicians about medical decision-making.4 The majority of Americans prefer to die at home if they are terminally ill.5 Additionally, it has been noted that the site of death may be associated with perceived quality of life at the end of life. Thus patients who die at home especially with hospice services tend to have a better quality of care than in an institutional setting.6 The role of PCS may be greater in the elderly with advanced cancer. Given that individuals age 65 years and older represent the fastest growing segment of the population in Western countries, there is an urgent need to assess availability, acceptance, and utilization of PCS close to end of life in patients with advanced cancer. With burgeoning health care costs, especially in the United States, and given the high expense incurred close to end of life in patients with advanced cancer, as well as among older adults, this area needs to be studied further. High utilization of acute care settings, including emergency room visits, hospitalizations, and intensive care monitoring, all contribute to these rising costs in addition to taxing the capacity of health care delivery services. In the absence of appropriate use of PCS, greater proportion of adults with advanced cancer rely on acute care services (ACS) at the end of life.7 However, there are limited data regarding the utilization of these services by this specific population of older adults with advanced cancer. A recent study in the Journal of Geriatric Oncology aimed at evaluating the needs in elderly patients affected by cancer and the state of the research in palliative care failed to meet its aim due to a paucity of literature focusing on these issues.8 Hence, to compare the utilization of PCS and ACS in older adults with advanced, incurable cancer, we carried out a retrospective study of adult patients with advanced, incurable cancer at the Syracuse VA medical center (VAMC) with the intent to study the impact of age upon such use. We defined PCS as the referral to and acceptance of palliative care: this was synonymous with formal hospice care in a vast majority of the patients but also included care rendered via a home nursing agency with palliative care goals (without formal hospice designation). The actual PCS could have been rendered in the home setting, at a skilled nursing facility or a hospice unit, as is the case of the VAMC which is housed physically within the hospital itself. Irrespective of the setting, the principles of PCS were adhered to (i.e., focus on maintaining comfort and dignity as well as the discontinuation of all cancer fighting treatments). Information was collected on factors that could potentially impact a patient's decision to accept PCS including demographics (age, gender, and race), tumor factors (type of cancer), and living situation (alone vs accompanied and marital status). It was decided to include patients with solid tumors and lymphoma but not those with leukemia or myeloma since in the latter, the status of remission is not always immediately obvious and these
diseases often require ongoing acute care such as transfusion support and management of infectious complications post chemotherapy. While the prime objective was to study the utilization of PCS and ACS in older adults, we included younger patients from the same time period to compare the differences based on age. Thus, the purpose of this analysis was to assess whether there are differences in patterns of referral, acceptance and ultimately, utilization of PCS based on age. Our hypothesis was that older adults with cancer will likely be referred to PCS earlier than younger patients, accept PCS more often and spend greater time with PCS prior to death than younger patients. As a corollary, older adults will use ACS with decreased frequency in the last month of life compared to younger patients.
2. Methods 2.1. Objectives 2.1.1 To study the utilization of PCS at our institution for adults who died as a consequence of advanced cancer by calculating the mean duration between day of hospice referral and time of death (DoR) and the mean length of stay with hospice (LoS) for each age group. 2.1.2 To compare the utilization of PCS between older (≥ 65 years) and younger adults (40–64 years) with advanced cancer. 2.1.3 To study rates of ACS utilization in the last month of life between older and younger patients by identifying Emergency room visits in the last month of life (ERVLM), Hospital admissions in the last month of life (HALM), and Intensive care unit admissions in the last month of life (ICULM). 2.1.4 To study the association of living situation and marital status with duration of PCS.
2.2. Setting and Population The study setting was Syracuse Veterans Affairs Medical Center (SVAMC). All patients who died as a consequence of advanced cancer at the SVAMC from 2002 to 2009 were included in the cohort. These patients were identified from the VAMC tumor registry. At the SVAMC, all patients with advanced cancer are referred to an “advanced illness care coordinator”, a function served by a licensed social worker. Patients who are referred have the right to decline PCS until they are psychologically ready to make such a transition. Data regarding date of referral, acceptance of PCS and death were collected from the notes entered by the palliative service in the electronic medical record. Admissions to the hospital and emergency room visits were collected from the patient chart. Thus, inclusion criteria as defined for this analysis include: patients (age more than 40 years or less than 100 years) with advanced cancer who died during the stated period, who were referred to PCS irrespective of whether they accepted PCS or continued with acute care for their cancer. The patients who presented with acute illness and early
41
J O U RN A L OF GE RI A TR IC O N CO LOG Y 7 ( 20 1 6 ) 3 9 –46
demise even prior to a referral to PCS were not included. Patients below 40 or above 100 years of age and those with non-metastatic disease were excluded since younger patients and the rare centenarians may have goals different than those of the typical younger and older patients, respectively. For reasons stated in the previous section, patients with leukemia and myeloma were excluded.
2.3. Data Collection After initial IRB approval, a retrospective electronic medical records chart review of 567 patients identified by the tumor registry was performed. We compared 372 older patients (≥65 years) with 195 younger (40–64 years old) adults.
2.4. Measures Independent variables identified were: age of the patient at the time of diagnosis of advanced cancer, race of the patient (Caucasian, African-American and others), marital status of the subjects (divorced, married, single, widowed and unknown), living condition of the patient at the time of diagnosis (home with family, home alone and nursing home), location of the cancer (head & neck, lung & pleura, gastrointestinal including liver & pancreas, genitourinary including bladder & prostate, lymph nodes and others). Utilization of PCS was identified as mean number of days of referral to hospice before death (DOR) and mean length of stay in hospice before death (LoS). Utilization of hospital resources: ACS was measured in terms of ER visits (ERVLM), hospital admissions (HALM) and ICU admissions (ICULM) in the last month of life.
2.5. Statistical Analysis Numeric variables were summarized using simple descriptive statistics such as the number of observations, mean and standard deviation. Frequencies and relative frequencies were computed for all categorical variables. The utilization of PCS and ACS between younger and older groups was compared using both unadjusted two-sample tests and multiple regression analyses adjusting the potential confounding effects of other factors. For DOR and LoS, two sample t-tests were performed and Satterthwaite approximation was used if the variances were statistically different between the two groups. For ERVLM, HALM and ICULM, two sample tests for Poisson rate were performed. Multiple linear regressions (for DOR and LoS) and multiple Poisson regressions (for ERVLM, HALM and ICULM) were also performed in order to adjust the impact of cancer location (lung vs gastrointestinal vs head & neck vs others), marital status (married vs others) and living condition (home with family vs others). Overdispersion in the Poisson regression was adjusted by Pearson chi-square statistic. The absolute percentages of patients who utilized acute care (ER visits, hospital and ICU admissions) in the last month of life and who died in the hospital were compared by Pearson chi-square tests, if less than 50% of the cells have expected counts less than 5, or Fisher's exact test otherwise. Furthermore, multivariate linear regression for the utilization of PCS and ACS and multivariate logistic regression for the absolute percentages of patients who utilized acute care were performed in order to adjust the impact of location of cancer (lung vs gastrointestinal vs head & neck vs others) and social support, including marital status (married vs others) and living situation (home with family vs others).
Table 1 – Sample characteristics. Variable
Age
Race
Sex Location of cancer
Marital status
Living condition
Mean (std) Median (min, max) White African American Others Male Lung & pleura Gastrointestinal Head & neck Lymph nodes Genito-urinary Others Married Divorced Widowed Single Unknown Home with family Home alone Nursing home
Young group (40–64) N = 195
Old group (65 and above) N = 372
p-Value[a]
57.7 (4.76) 58 (44, 64) 177 (90.8%) 15 (7.7%) 3 (1.5%) 187 (95.9%) 92 (47.2%) 46 (23.6%) 22 (11.3%) 13 (6.7%) 8 (4.1%) 14 (7.2%) 70 (35.9%) 78 (40.0%) 9 (4.6%) 38 (19.5%)
75.8 (6.69) 75 (65, 100) 358 (96.2%) 13 (3.5%) 1 (0.3%) 367 (98.7%) 182 (48.9%) 84 (22.6%) 37 (9.9%) 16 (4.3%) 38 (10.2%) 15 (4.0%) 173 (46.5%) 85 (22.8%) 77 (20.7%) 33 (8.9%) 4 (1.1%) 207 (55.6%) 140 (37.6%) 25 (6.7%)
<0.001
104 (53.3%) 86 (44.1%) 5 (2.6%)
0.061
0.037 0.072
<0.0001
0.059
[a] p values are calculated from Pearson chi-square tests, if less than 50% of the cells have expected counts less than 5; otherwise Fisher's exact test is used.
42
JO U RN A L OF GE RI A TR IC O N COL O G Y 7 (2 0 1 6 ) 3 9 –4 6
Fig. 1 – Distribution of age: Histograms with density plots (top) and boxplots (bottom) of age by group. The bold line represents a normal density estimate and the dashed line represents a kernel density estimate.
3. Results
the older group). More than half of the patients across both the groups were living at home with family.
3.1. Sample Characteristics
3.2. Utilization of Palliative Care Services
Out of the 567 patients who met the eligibility criteria, 195 patients were in the younger group and 372 in the older group (Table 1, Fig. 1). More than 90% of the sample comprised of Caucasian males. Lung cancer was the most commonly encountered tumor type (47% in the young group and 48% in
Earlier referral to PCS was observed in the older group [mean DOR: 47.3 days versus 34.5 days, p = 0.015]. Older patients spent more time in hospice [mean LoS: 32.5 days versus 20.2 days, p = 0.007] than the younger group. There were 65 patients (16.9%) in the older group and 35 (17.9%) in the
Table 2 – Comparison of outcomes between two groups. Variable
DOR LoS ERVLM HALM ICULM LoS[b]
Mean STD Mean STD Mean STD Mean STD Mean STD N Mean STD
days days number of visits number of admissions number of admissions
Young group (40–64) N = 195
Old group (65 and above) N = 372
p-Value ⁎ (p-Value &)
34.5 51.98 20.2 43.25 0.8 0.83 0.86 0.76 0.22 0.47 160 24.7 46.61
47.3 70.88 32.5 63.67 0.7 0.79 0.72 0.72 0.14 0.36 307 39.1 68.03
0.015 (0.027) 0.007 (0.013) 0.138 (0.119) 0.043 (0.035) 0.030 (0.023) 0.016 (0.012)
Abbreviations: DOR, the mean duration between day of hospice referral and time of death; LoS, length of stay; ERVLM, emergency room visits in the last month of life; HALM, hospital admissions in the last month of life; ICULM, intensive care unit admissions in the last month of life; STD, standard deviation. [b] 0 and missing values excluded. The 0 values are considered as missing values and excluded from analysis, assuming that all 0 values are as a result that the patient refused hospice. Including the 0 values in the analysis will underestimate the true value. ⁎ p Values from univariate two-sample tests. & p Values from multivariate regression analyses.
J O U RN A L OF GE RI A TR IC O N CO LOG Y 7 ( 20 1 6 ) 3 9 –46
younger group who did not accept transition to PCS and for this analysis, they were considered to have LoS of 0 days. When these patients were excluded from LoS analysis, the difference between two groups continued to be significant [mean LoS: 39.1 versus 24.7 days, p = 0.016] (Table 2). We also conducted multivariate linear regression analysis that adjusted for type of cancer as well as marital status and living situation. These factors were found to not be significantly associated with the outcomes variables and did not confound the association between age and outcome variables. After the adjustment, mean DOR and mean LoS for older patients are 13.5 days and 12.8 days more than younger patients, respectively. When the patients who did not accept PCS were excluded from LoS analysis, mean LoS for older patients are 15.2 days longer.
3.3. Utilization of Acute Care Services There were fewer hospital admissions [HALM: 0.72 versus 0.86, p = 0.043] and ICU admissions [ICULM: 0.14 versus 0.22, p = 0.030] per patient in the last month of life in older adults when compared to younger adults (Table 2). There was no significant difference between the numbers of ER visits for older versus younger patients (p = 0.138). After the adjustment, mean HALM, mean ICULM and mean ERVLM for younger veterans are 0.18 days (p = 0.035), 0.48 days (p = 0.023) and 0.14 visits (p = 0.119) more than older veterans, respectively. However, the absolute percentage of patients utilizing acute care (ER visits and hospital admissions) in last month of life was similar in both groups with fewer older adults receiving ICU care (Table 3). We conducted multivariate logistic regression analysis that adjusted for type of cancer, marital status and living situation. These factors were not statistically significantly associated with the outcomes variables and did not confound the association between age and outcome variables. After the adjustment, the odds ratio between younger and older patients with respects to death in hospital, having at least one ER visit, at least one hospital admission and at least one ICU visit were 0.770 (95% CI: 0.470–1.261), 0.771 (95% CI: 0.540–1.101), 0.762 (95% CI: 0.530–1.096) and 0.612 (95% CI: 0.383–0.980) of the odds for older patients, respectively. Table 3 – Utilization of ACS by two groups. Variable Young group Old group (40–64) (65 and above) N = 195 N = 372 Death In hospital ERVLM
31 (15.9%)
49 (13.2%)
116 (59.5%)
199 (53.5%)
HALM
127 (65.1%)
218 (58.6%)
ICULM
38 (19.5%)
49 (13.2%)
OR, 95%CI, p-value ⁎ (p-value &) 0.770, 0.470 to 1.261, 0.376 (0.300) 0.771, 0.540 to 1.101, 0.173 (0.152) 0.762, 0.530 to 1.096, 0.130 (0.142) 0.612, 0.383 to 0.980, 0.047 (0.041)
Abbreviations: ERVLM, emergency room visits in the last month of life; HALM, hospital admissions in the last month of life; ICULM, intensive care unit admissions in the last month of life. ⁎ p Values are calculated from Pearson chi-square tests. & p Values, adjusted OR and 95% CI from multivariate logistic regression analyses.
43
4. Discussion In this study, we found that older patients were referred to PCS earlier and spent a longer period with PCS compared to the younger cohort. However, a similar proportion of older and younger patients (17% and 18%, respectively) declined PCS. Further, more than half of older patients continued to have ER visits and hospital admissions in the last month of life and this proportion was not remarkably different than younger patients.
4.1. PCS Utilization Some unique features of this retrospective review include the fact that we were able to assess mean duration of referral to hospice services which may convey the intent of the medical provider to initiate formal PCS. While it may appear that DOR and LoS are inter-dependent, these variables can be discordant. Patients were offered to transition to PCS via a referral but may elect not to accept PCS for a while. In fact, 17% of the older and 18% of the younger patients did not accept such a transition and hence their LoS is considered a zero for statistical purposes albeit not the DOR. However, we did not find discordance between these variables likely due to the fact that similar proportion of older and younger patients declined a transition to PCS. Older adults with advanced cancer were referred to PCS earlier and spend a longer duration with hospice than younger patients with similar diagnoses. While this may potentially raise a concern for age bias, given that ultimately all these patients succumbed to their disease, a longer duration with PCS would reflect appropriate clinical practices in patients close to end of life. This is a move in the right direction especially since older adults may have goals which differ from younger patients, such as maintaining QoL with lower acceptance of therapy associated toxicity and less focus on prolongation of life. In a report from a US-based community based cancer center, there were no age differences among patients referred to hospice versus those that were not referred to hospice prior to death.9 This study was similar to ours in that it includes patients with a variety of cancers encountered in the community setting. However, it did not delve into the age-based differences in outcomes as attempted in our analysis. Investigators from Asia have reported similar associations between older age and a longer length of stay with PCS.10,11 In another US based retrospective review utilizing two large population databases, fewer patients with advanced lung cancer under age 65 received hospice services as compared to patients above 65 years of age.12 In a Canadian study limited to patients with advanced pancreatic cancer, earlier PCS consultation was associated with less aggressive care near death: decreased use of chemotherapy (odds ratio [OR]: 0.34, 95% confidence interval [CI]: 0.25–0.46); lower risk of ICU admission (OR: 0.12, 95% CI: 0.08–0.18); multiple ED visits (OR: 0.19, 95% CI: 0.16–0.23); multiple hospitalizations near death (OR: 0.24, 95% CI: 0.19 to 0.31).13 In another retrospective study limited to patients with ovarian cancer, older women received hospice care for a longer duration over a 10 year study period. However, it was also observed that these women had statistically significant increases in
44
JO U RN A L OF GE RI A TR IC O N COL O G Y 7 (2 0 1 6 ) 3 9 –4 6
intensive care unit admissions, hospitalizations, repeated emergency department visits, and health care transitions (all P ≤ .01). Inpatients referred to hospice were more likely to enroll in hospice within 3 days of death than outpatients 14 (adjusted odds ratio: 1.36; 95% CI: 1.12–1.66). A recent report found that predictors of a short hospice stay (defined as ≤ 3 days) include hematologic malignancies, male gen15 der, married status and younger age (< 65 years). While the former two do not apply to our cohort, these findings are similar in terms of age. We did not find marital status to be associated with LoS with PCS in our population. This may be a consequence of lack of representation of women in the sample. Overall, our findings are similar to the above studies in that older patients were referred to PCS earlier and spent longer duration with PCS compared to younger patients. Unlike many of these studies which are limited to patients with single cancer type, our study included patients with variety of advanced solid tumors.
4.2. ACS Utilization Use of ACS is associated with enhanced costs close to end of life and the timely use of PCS can limit the costs incurred by the health care system. A recent report of Medicare beneficiaries compared ACS utilization and costs of care in patients with poor prognosis cancer enrolled with hospice versus a matched cohort that died without hospice. In patients utilizing hospice, hospitalizations, hospital admissions, invasive procedures were all reduced significantly leading to a diminished cost of care versus those that were not enrolled to hospice.16 A variable which is often not reported upon, even in large population database analyses, are the emergency room visits. Emergency room visits can add significant burden to the healthcare system and can play a role in compromising the perceived quality of life for the patients and their caregivers. In our cohort, older patients received less aggressive care close to EOL including fewer ER visits, hospital admissions, ICU admissions as compared to younger patients. However, more than half of older adults with advanced cancer required ER visits and hospital admissions in the last month of life. The high rates of ACS utilization (ER visits and hospital admission) in > 50% of patients with relatively short LOS with hospice suggest that PCS could have been initiated sooner. The rates of ACS utilization indicate that the treatment team needs to better anticipate a transition to palliative care. In a study by Mack et al., that included only patients with advanced lung and colon cancer, nearly half of patients received at least one marker of aggressive EOL care, including chemotherapy in the last 14 days of life (16%), intensive care unit care in the last 30 days of life (9%), and acute hospital-based care in the last 30 days of life (40%) with no association with age noted in that study.7 The latter rates in our cohort are somewhat higher for the older patients at 13% and 59% for ICULM and HALM respectively suggesting room for improvement. In another retrospective review of the Ontario registry, factors associated with aggressive care close to EOL included male gender, younger age as well as tumor type (lung, breast and hematological malignancies).17 In our cohort, we did not find
any association with tumor type and use of ACS in the last month of life. We are not able to study the association between gender and ACS utilization due to preponderance of males in our cohort. In a recently reported pooled analysis of a retrospective cohort study, the use of community based palliative care teams reduced the number of ER visits and hospital deaths at the end of life.18 Such community based palliative care teams may be effective in reducing ACS utilization at the EOL.
4.3. Barriers Accessing PCS Qualitative evidence in the literature highlighted numerous barriers to accessing palliative care. Some barriers are a consequence of the lack of education of physicians and other health care professionals, regarding end of life and its management. Heath care providers' personal and religious beliefs may also impact their view on this matter. Clear and honest communication between the health care provider and the patient is essential for patients to understand their prognosis and accept a transition to PCS at the appropriate time. In a retrospective study looking at patients with metastatic lung cancer, around half of patients had not discussed hospice with a provider within 4 to 7 months after diagnosis.19 Other barriers are the result of patient/family misinformation and aspects of the hospice system that limit the number and types of patients deemed appropriate for hospice care.20 As per a recent study, many patients do not understand that the chemotherapy in a metastatic cancer is only palliative and not curative.21 Better understanding about the goals of treatment would probably allow them to choose PCS over chemotherapy. In another study, patients and caregivers found that palliative care was often provided sub-optimally during the “out-of-hours” that is weekends and public holidays.22 It was perceived by caregivers that the “service responses” during the “out-of-hours” were designed for acute medical rather than palliative care needs. These systemic issues in the adequate delivery of PCS are vital for such services to be more accessible and hence, acceptable to the patient and the family. Good anticipatory care with the timely provision of information to patients and caregivers so as to set the right expectations is therefore critical for the success of PCS.22 In a recent retrospective review of over 100,000 patients with cancer, the cost of medical care in the last month of life was 43% higher in those receiving aggressive care compared to those receiving non-aggressive care ($18,131 vs $12,678, P < .0001).23 Communicating these and similar data to healthcare providers and possibly to patients and advocacy groups may facilitate the wider acceptance of PCS close to EOL.
4.4. Study Limitations & Scope for Future Work Limitations of our study would be retrospective chart review, small sample size, single center data, predominantly male patients and single payer insurance system. Most of this study was completed before the impact of Affordable Care Act (ACA) of 2010 which gave several incentives to health care providers for initiating end-of-life discussions with their patients. It would be interesting to study health outcomes and physician
J O U RN A L OF GE RI A TR IC O N CO LOG Y 7 ( 20 1 6 ) 3 9 –46
perceptions before and after the advent of ACA. Secondly, multi-center data could be accessed to compare our results with the national average and advocate for policy changes based on the results. Further, we did not look specifically at the subset of patients who declined palliative care altogether. So, the impact upon their utilization of ACS in the last month cannot be discerned.
4. 5.
5. Conclusion 6.
Older patients with cancer are likely to be referred to PCS earlier and spend a longer duration with PCS. However, with more than half of older adults with cancer reporting ER visits and admissions in the last month of life, there continues to be significant utilization of ACS by this population. Earlier end of life discussions and referral to PCS in patients with advanced cancer and limited life expectancy can decrease the use of aggressive care, increase the use of hospice services and improve the quality of life during the last days of their life. Physician–patient communication is essential for patients to understand their prognosis and the concept of hospice, which can ultimately lead to an increase in the proportion of patients accepting hospice. Better understanding of the goals of care in older adults and education of oncology providers regarding the need to integrate PCS earlier in the course of disease is imperative.
Disclosures and Conflict of Interest Statements Study concepts: A. Gajra, N. Sharma. Study design: A. Gajra, N. Sharma, A.M. Sharma. Data acquisition: N. Sharma, A.M. Sharma. Quality control of data and algorithms: A. Gajra, N. Sharma, A.M. Sharma. Data analysis and interpretation: A. Gajra, N. Sharma, A.M. Sharma. Statistical analysis: D. Wang, M.A. Wojtowycz. Manuscript preparation: N. Sharma, A.M. Sharma. Manuscript editing: A. Gajra Manuscript review: A. Gajra.
REFERENCES
1. Keating Nancy L, Landrum Mary Beth, Lamont Elizabeth B, Earle Craig C, Bozeman Samuel R, Mcneil Barbara J. End-of-life care for older cancer patients in the veterans health administration versus the private sector. Cancer 2010: 3732–3739. 2. Gonsalves Wilson I, Tashi Tsewang, Krishnamurthy Jairam, Davies Tracy, Ortman Stephanie, Thota Ramya, Aldoss Ibrahim, Ganta Ashwin, Kalaiah Mudappa, Didwaniya Neha, Eberle Catherine, Ganti Apar K, Silberstein Peter T, Subbiah Shanmuga. Effect of Palliative Care Services on the Aggressiveness of End-of-Life Care in the Veteran's Affairs Cancer Population. J. Palliat Med. 2011;14(11):1231–1235. 3. Temel Jennifer S, Greer Joseph A, Muzikansky Alona, Gallagher Emily R, Admane Sonal, Jackson Vicki A, Dahlin
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
45
Constance M, Blinderman Craig D, Jacobsen Juliet, Pirl William F, Andrew Billings J, Lynch Thomas J. Early palliative care for patients with metastatic non-small-cell lung cancer. N Engl J Med 2010:733–742. Teno JM. Family perspectives on end-of-life care at the last place of care. JAMA 2004;1(291):88–93. Goodman David, Morden Nancy, Chang Chiang-Hua, Fisher Elliott, Wennberg John, Bronner Kristina. Trends in Cancer Care Near the End of Life A Dartmouth Atlas of Health Care Brief. The Dartmouth Atlas of Health Care; September 4 2013 [Accessed August 24, 2014. , http://www.dartmouthatlas.org/ downloads/reports/Cancer_brief_090413.pdf]. Teno JM, Gozalo PL, Bynum JP, Leland NE, Miller SC, Morden NE, Scupp T, Goodman DC, Mor V. Change in end-of-life care for Medicare beneficiaries: site of death, place of care, and health care transitions in 2000, 2005, and 2009. JAMA Feb 6 2013;309(5):470–477 [PMID: 23385273]. Mack JW, Cronin A, Keating NL, Taback N, Huskamp HA, Malin JL, Earle CC, Weeks JC. Associations between End-of-life discussion characteristics and care received near death: a prospective cohort study. J Clin Oncol 2012:4387–4395. Brighi N, Balducci L, Biasco G. Cancer in the elderly: is it time for palliative care in geriatric oncology? J Geriatr Oncol 2014: 197–203. Cowall DE, Yu BW, Heineken SL, Lewis EN, Chaudhry V, Daugherty JM. End-of-life care at a community cancer center. J Oncol Pract Jul 2012;8(4):e40–e44, http://dx.doi.org/10.1200/JOP.2011.000451 [Epub 2012 May 22. PubMed PMID: 23180997; PubMed Central PMCID: PMC3396828]. Chang TS, Su YC, Lee CC. Determinants for aggressive end-of-life care for oral cancer patients: a population-based study in an Asian country. Medicine (Baltimore) Jan 2015;94(4): e460, http://dx.doi.org/10.1097/MD.0000000000000460 [PubMed PMID: 25634186]. Tang ST, Liu TW, Shyu YI, Huang EW, Koong SL, Hsiao SC. Impact of age on end-of-life care for adult Taiwanese cancer decedents, 2001–2006. Palliat Med Jan 2012;26(1):80–88, http: //dx.doi.org/10.1177/0269216311406989 [Epub 2011 May 23. PubMed PMID: 21606128]. Mack JW, Chen K, Boscoe FP, Gesten FC, Roohan PJ, Weeks JC, Schymura MJ, Schrag D. Underuse of hospice care by Medicaid-insured patients with stage IV lung cancer in New York and California. J Clin Oncol Jul 10 2013;31(20):2569–2579, http://dx.doi.org/10.1200/JCO.2012.45.9271 [Epub 2013 Jun 3. PubMed PMID: 23733768; PubMed Central PMCID: PMC3699723]. Jang RW, Krzyzanowska MK, Zimmermann C, Taback N, Alibhai SM. Palliative care and the aggressiveness of end-of-life care in patients with advanced pancreatic cancer. J Natl Cancer Inst Jan 20 2015;107(3), http://dx.doi.org/10.1093/jnci/dju424 [pii: dju424., Print 2015 Mar. PubMed PMID: 25609233]. Wright AA, Hatfield LA, Earle CC, Keating NL. End-of-life care for older patients with ovarian cancer is intensive despite high rates of hospice use. J Clin Oncol Nov 1 2014;32(31): 3534–3539, http://dx.doi.org/10.1200/JCO.2014.55.5383 [Epub 2014 Oct 6. PubMed PMID: 25287831; PubMed Central PMCID: PMC4209104]. O'Connor NR, Hu R, Harris PS, Ache K, Casarett DJ. Hospice admissions for cancer in the final days of life: independent predictors and implications for quality measures. J Clin Oncol Oct 1 2014;32(28):3184–3189, http://dx.doi.org/10.1200/JCO.2014.55.8817 [Epub 2014 Aug 25. PubMed PMID: 25154824; PubMed Central PMCID: PMC4171361]. Obermeyer Z, Makar M, Abujaber S, Dominici F, Block S, Cutler DM. Association between the Medicare hospice benefit and health care utilization and costs for patients with poor-prognosis cancer. JAMA Nov 12 2014;312(18):1888–1896, http://dx.doi.org/10.1001/jama.2014.14950 [PubMed PMID: 25387186; PubMed Central PMCID: PMC4274169].
46
JO U RN A L OF GE RI A TR IC O N COL O G Y 7 (2 0 1 6 ) 3 9 –4 6
17. Ho TH, Barbera L, Saskin R, Lu H, Neville BA, Earle CC. Trends in the aggressiveness of end-of-life cancer care in the universal health care system of Ontario, Canada. J Clin Oncol Apr 20 2011;29(12):1587–1591, http://dx.doi.org/10.1200/JCO.2010.31.9897 [Epub 2011 Mar 14. PubMed PMID: 21402603; PubMed Central PMCID: PMC3082976]. 18. Seow H, Brazil K, Sussman J, Pereira J, Marshall D, Austin PC, Husain A, Rangrej J, Barbera L. Impact of community based, specialist palliative care teams on hospitalisations and emergency department visits late in life and hospital deaths: a pooled analysis. BMJ Jun 6 2014;348:g3496, http://dx.doi.org/ 10.1136/bmj.g3496 [PubMed PMID: 24906901; PubMed Central PMCID: PMC4048125]. 19. Huskamp HA, Keating NL, Malin JL, Zaslavsky AM, Weeks JC, Earle CC, Teno JM, Virnig BA, Kahn KL, He Y, Ayanian JZ. Discussions with physicians about hospice among patients with metastatic lung cancer. Arch Intern Med May 25 2009;169(10): 954–962, http://dx.doi.org/10.1001/archinternmed.2009.127 [PubMed PMID: 19468089; PubMed Central PMCID: PMC2689617].
20. Friedman Betsy T, Harwood M Kay, Shields Matthew. Barriers and enablers to hospice referrals: an expert overview. J Palliat Med 2002:73–84. 21. Weeks Jane C, Catalano Paul J, Cronin Angel, Finkelman Matthew D, Mack Jennifer W, Keating Nancy L, Schrag Deborah. Patients' expectations about effects of chemotherapy for advanced cancer. N Engl J Med 2012:1616–1625. 22. Worth A, Boyd K, Kendall M, Heaney D, Macleod U, Cormie P, Hockley J, Murray S. Out-of-hours palliative care: a qualitative study of cancer patients, carers and professionals. Br J Gen Pract Jan 2006;56(522):6–13 [PMID: 16438809]. 23. Cheung MC, Earle CC, Rangrej J, Ho TH, Liu N, Barbera L, Saskin R, Porter J, Seung SJ, Mittmann N. Impact of aggressive management and palliative care on cancer costs in the final month of life. Cancer May 29 2015, http://dx.doi.org/10.1002/cncr.29485 [[Epub ahead of print] PubMed PMID: 26031241].