Cost Differential Associated with Hospice Use among Older Patients with Multiple Myeloma

Abstract

BACKGROUND:

Hospice is an effective end-of-life care approach for patients with incurable illnesses such as multiple myeloma; however, it has been historically underutilized. In addition to improving quality of life, hospice enrollment reduces healthcare spending in many incurable illnesses but this has been unstudied in the myeloma population to date.

MATERIAL AND METHODS:

Retrospective analysis of myeloma cases diagnosed from 2007-2013 in the Surveillance, Epidemiology, and End Results (SEER)-Medicare linked dataset. Included patients were: diagnosed at age 65 or older, received myeloma-directed therapy, had an overall survival greater than 3 months, and were enrolled in Medicare the month preceding death. Costs included those paid by Medicare and patient copays during the 30 days preceding death.

RESULTS:

2,075 patients were included in the analysis. 56% were enrolled in hospice at end of life. Increasing age and female gender were associated with greater odds of hospice enrollment. Non-white race, Medicaid enrollment, and increasing comorbidities were associated with decreased odds. Hospice enrollment was associated with a $13,574 (p < 0.0001) decrease in costs; however, the maximal savings were observed by those enrolled more than 14 days prior to death.

CONCLUSION:

While improving quality of life should be the ultimate reason for increasing hospice utilization among patients with myeloma, there seems to be considerable cost implications as well.

INTRODUCTION

Multiple myeloma is a hematologic cancer diagnosed predominantly among older adults, with a median age of 69 years at diagnosis.¹ Although there has been an unprecedented number of new treatments introduced over the past two decades, myeloma remains incurable and patients experience a substantial symptom burden particularly near the end of life.^2,3

Hospice, a comprehensive care model that focuses on improving quality of life through palliative symptom management and psychosocial support services, is an effective end-of-life care approach for patients with incurable illnesses.⁴ Historically, hospice has been underutilized among patients with multiple myeloma; however, a recent study suggests that enrollment has increased in recent years. Odejide and colleagues recently reported that from 2000 to 2013, the rate of hospice utilization among patients with myeloma nearly doubled, from 29% to 57%.⁵

The clinical and psychosocial benefits of hospice care include improved symptom management, access to needed equipment and additional in-home support, reduction in emotional and spiritual distress, and improved caregiver outcomes, all of which likely improve quality of life at end of life.⁶ In addition to the clinical and psychosocial benefits, hospice utilization has positive implications for healthcare expenditures. Hospice utilization saves the Medicare program approximately $2,300 per beneficiary overall;⁷ however, cost savings are 2-3 fold for those enrolling at least 2 weeks prior to death.⁸

Approximately one-quarter of all Medicare spending is caring for individuals during the final year of life.⁹ The use of intensive end-of-life care measures—intensive care admissions, emergency department visits, and end-stage treatment regimens—are a major contributor to these costs. Debate arises whether intensive end-of-life care contributes to clinical benefit and improved quality of life for patients with advanced cancer. In fact, some studies suggest that patients who receive intensive care experience lower quality of life in comparison to those who enroll in hospice or receive symptom-directed care only at end of life.¹⁰

Among patients with myeloma, who experience high cost of care due to high symptomology and clinical needs, the cost implications are likely even greater than the general population; but, this has been understudied to date. To address this gap in the literature, we compared the costs of end-of-life care for patients with myeloma who were enrolled in hospice at end of life to those patients who were not.

MATERIALS AND METHODS

We performed a retrospective analysis of myeloma cases diagnosed from 2007-2013 in the Surveillance, Epidemiology, and End Results (SEER)-Medicare linked dataset. The SEER-Medicare database has been described in detail elsewhere.¹¹ Briefly, SEER collects demographics, tumor characteristics, and survival data from 18 population-based cancer registries throughout the United States covering approximately 26% of the U.S. Population.¹² In the SEER-Medicare linked database, the SEER registry data is linked to Medicare enrollment and claims data. Of all people 65 years of age or older in the SEER registry, 93% have been matched to their corresponding Medicare data.¹² At the time this study was conducted, the SEER-Medicare linkage included all Medicare-eligible persons appearing in the SEER data through 2013 and their Medicare claims and survival data through 2014. The study was performed under the supervision of the local institutional review board.

To create a more homogenous cohort of older patients with myeloma who received treatment and would likely have had the opportunity to pursue hospice prior to death, those not meeting the following criteria were excluded in the analysis: diagnosed at age 65 or older, received myeloma directed therapy, and had an overall survival greater than 3 months. Those not enrolled in Medicare the month preceding death were also excluded as the care received at end of life and the associated costs were not available in the dataset.

The following Medicare claims files were used for this analysis: Hospice, which includes all claims for hospice services [Medicare Part A]; Medicare Provider Analysis and Review (MEDPAR), which includes all hospital short-stay, long-stay, and skilled nursing facility bills [Part A]; National Claims History (NCH), which includes all physician/supplier bills [Part B]; Outpatient, which includes all bills from institutional outpatient providers [Part B]; Home Health Agency (HHA), which includes all claims for home health services [Part B]; and Durable Medical Equipment (DME), which includes all claims for durable medical equipment such as oxygen tanks [Part B], The Part D file, which includes claims for prescription drugs, were not included as they are only available on a subset of patients in the dataset [Part D].

Hospice utilization was defined as one or more claims for hospice services at end-of-life. All reported costs, including those paid by Medicare and patient copays, during the 30 days preceding death were captured and adjusted for inflation using the medical services section of the Consumer Price Index. In a secondary analysis, we compared the costs at myeloma diagnosis (+/− 30 days) to better elucidate if any differences in costs were associated with patient characteristics other than hospice enrollment. Costs were compared between those enrolled in hospice and those who weren’t using bivariate analyses and multivariate linear regression models. Sub-analyses were then performed comparing the costs of individuals with earlier versus later hospice enrollment.

RESULTS

A cohort of 2,075 patients were included in the analysis, the median age was 76 at myeloma diagnosis and 52% were male. Cohort selection is detailed in Figure 1. Seventy-nine percent were white, 16% African-American or black, and 5% another race. The median overall survival of the cohort was 18.5 months (95% CI 17.4-19.2).

Figure 1:

Cohort Selection

Fifty-six percent of patients enrolled in hospice at end of life. Of those that did, 18% enrolled within 3 days prior to death, which is generally considered late enrollment. The characteristics of patients enrolled in hospice versus those who were not are compared in Table 1. Interestingly, the median overall survival, from time of diagnosis, was longer among those enrolled in hospice (19.6 months) in comparison to those who did not (16.6 months; p = 0.0045). In multivariate analysis, we found that increasing age and female gender were associated with increased odds of hospice utilization, while non-white race, Medicaid dual-enrollment (in addition to Medicare), and increasing comorbidities were associated with decreased odds [Table 2].

Table 1:

Patient Characteristics

	All (n = 2075)	No Hospice Enrollment (n =917)	Hospice Enrollment (n =1158)
Age at myeloma diagnosis (median, range)	75 (65-97)	75 (65-96)	77 (65-97)
Gender
Male	52%	57 %	48%
Female	48%	43%	52%
Race
White	79%	75 %	82%
Non-White	21%	25%	18%
Medicaid dual-enrollment
Yes	29%	35 %	24%
No	71%	65%	76%
Poor Performance Status Indicators20
Yes	21%	22%	20%
No	79%	78%	80%
Charlson Comorbidity Index Score21 (median, range)	2 (0-11)	2 (0-9)	1 (0-11)
Overall Survival (median, 95% CI)	18.5 (14.4-19.2)	16.6 (15.2-17.8)	19.6 (18.6-21.5)

²⁰- Griffiths, et al, Blood, 2011.

²¹- Klabunde, et al .Journal of clinical epidemiology, 2000.

Table 2:

Multivariate Analyses of Hospice Utilization Hospice Enrollment

	aOR	95% CI	P
Age (per year)	1.04	1.02-1.05	<0.0001
Female Gender	1.52	1.26-1.82	<0.0001
None-White Race	0.77	0.61-0.97	0.0239
Medicaid dual-enrollment	0.63	0.51-0.77	<0.0001
Poor Performance Status Indicators	0.96	0.76-1.22	0.7492
Charlson Comorbidity Index Score (per point)	0.95	0.91-1.00	0.0468

Overall, the median cost of care at end of life was $15,870 (range 0-$215,048). The median cost of care for patients who were enrolled in hospice was $9,175, compared to $22,921 for those patients not enrolled (p <0.0001). After controlling for covariates, hospice enrollment overall was associated with a $13,574 (SE $832; p<0.0001) decrease in costs [Table 2]. We performed a sub-analysis to determine if earlier hospice enrollment was associated with less expenditures. The median cost of care for patients with enrollment within 3 days of end of life was $19,582 compared to $15,914 for those enrolled within 3-14 days, and $4,508 for those enrolled more than 14 days (p < 0.0001) [Figure 2]. The costs at myeloma diagnosis were similar between the cohorts (p = 0.8464), suggesting that costs difference at end of life are related to hospice use and not to other differences between the cohorts.

Figure 2: Median Cost of Care (2014 US dollars).

The median cost of care during the final 30 days of life was $22,291 for those not enrolled onto hospice compared to $9,175 for those enrolled. Those enrolled more than 14 days prior to end of life had a $4,508 median cost of care compared to $19,582 for those enrolled within 3 days.

DISCUSSION

In the current study, we found that hospice enrollment among patients with myeloma is associated with decreased healthcare expenditures at end of life. In a similar study of Medicare beneficiaries with leukemia, those not enrolled in hospice had nearly twice the expenditures within the last 30 days of life than those enrolled.¹³ While improving quality of life should be the ultimate reason for increasing hospice utilization among patients with myeloma, there seems to be considerable cost implications as well.

It is important to note that only a modest reduction in costs was observed among patients enrolled within 3 days of death. Late enrollment and the use of intensive end-of-life interventions remains a consistent trend among patients with myeloma although it has been relatively stable. Odejide and colleagues indicated that late enrollment only marginally increased (12.2% in 2000 to 16.3% in 2013) while enrollment overall nearly doubled.⁵ Poorly defined patient-reported goals of care at end of life may contribute to late or no enrollment in hospice.¹⁴ As the disease progresses, patients with myeloma quickly transition through multiple treatment regimens and experience a rapid decline. As a result, proactive conversations addressing goals of care at end of life may not occur. Consequently, lack of clarity regarding patient goals may lead to intensive medical care at end of life. This awareness reinforces the importance of timely advance care planning, an intervention now reimbursable by Medicare.¹⁵

For patients interested in hospice, access can be highly variable across the US. Overall, hospice utilization has increased dramatically over the past few decades, similarly to what was observed in the myeloma population.⁵ Facilities have been slow to meet the increasing demand. The American Academy of Hospice and Palliative Medicine has made it a strategic objective to foster rapid growth in training capacity in response to the shortage.¹⁶ One study estimated that the number of annual hospice and palliative medicine fellowship graduates would need to nearly double by 2030 (from the current level of 325 to 600 per year) to keep pace with demand.¹⁷ Therefore, hospice enrollment may not fully reflect the goals and wishes of the patient or the recommendations of their provider(s).

Although, the specific drivers of cost of end-of-life care were not analyzed in this study, we speculate that the use of costly end-stage treatment regimens (ie: chemotherapy, radiation, etc) and supportive care interventions (ie; dialysis, blood transfusions, etc), contributed to the higher costs for patients not enrolled in hospice and those with later enrollment. Odejide and colleagues reported that 75% of patients not enrolled in hospice had at least one indicator of medically intensive care at end of life (ie: ICU admission, multiple ED visits, etc) compared to 36% of enrollees.⁵ Further they showed that the chronic dependence on supportive interventions, such as dialysis and transfusions, was negatively associated with timely hospice enrollment. Similarly, among patients with leukemia, transfusion dependence is associated with a 51% shortened duration in hospice at end of life.¹³ Hospice enrollment in the USA typically involves discontinuation of chemotherapy, but this is not the case for many hospice programs in European countries.¹⁸ The ability to continue supportive care interventions during hospice have been proposed as a means to increase enrollment; however, it is important to note that doing so may reduce/eliminate the cost savings associated with hospice use.

An incidental finding of the current study is that the median overall survival was longer among those who enrolled in hospice in comparison to those who did not. This data should not be viewed as a suggestion that hospice prolonged survival compared to non-hospice care; however, this has been observed in other studies of Medicare beneficiaries with incurable diseases.¹⁹ Survival was measured from myeloma diagnosis rather than an index period such as time of disease progression which potentially bias the results. Many patients with myeloma will undergo a number of successive lines of therapy and have several progressions/relapses of their disease; thus, defining a universal index period is not feasible with the level of data included in the SEER-Medicare dataset. While it is possible that hospice enrollment at end of life prolongs survival compared to non-hospice care, it is also possible that patients who have had a longer duration from diagnosis to end-stage myeloma are more likely to enroll in hospice at end of life.

Additionally, the median overall survival was 19.6 months and 16.6 months, respectively, for those enrolled into hospice versus not; both are much lower than what is expected in this population. As the analysis only included decedents, the survival results will not accurately reflect the overall myeloma population. The median overall survival including patients surviving at data-cut-off was 33.1 months (95% CI 31.6-35.1) which is similar to what is expected for older patients with myeloma diagnosed during the period under study.

The SEER-Medicare dataset includes a large, nationally representative sample. However, there are several limitations to our analysis. First, the study was limited to older adults with Medicare coverage. Although the average age of diagnosis is 69, approximately 21.7% of myeloma-related deaths are of individuals age 64 or younger and the exclusion of younger patients, of whom many are uninsured or underinsured, may limit the generalizability of the findings.¹ Secondly, hospice use was determined by Medicare claims. We not have information regarding patient preferences, physician recommendations or appropriateness of care. Thirdly, we could not include expenditures unless they were billed to Medicare; therefore, the cost analysis was from the payer prospective only. The lack of data on privately incurred expenses may limit the generalizability of the study. Lastly, payment structures vary greatly from country to country, so cost implications may differ outside of the US.

Despite these limitations, this study further suggests that hospice utilization should be encouraged among patients with myeloma at end of life due to the clinical, psychosocial, and potential cost benefits. To maximize the benefits, earlier enrollment at least 3 days prior to death is optimal.