Massage, Music and Art Therapy in Hospice: Results of a National Survey

Abstract

Context

Complementary and alternative medicine (CAM) provides clinical benefits to hospice patients, including decreased pain and improved quality of life. Yet little is known about the extent to which U.S. hospices employ CAM therapists.

Objectives

To report the most recent national data regarding the inclusion of art, massage, and music therapists on hospice interdisciplinary teams and how CAM therapist staffing varies by hospice characteristics.

Methods

A national cross-sectional survey of a random sample of hospices (n=591; 84% response rate) from September 2008 to November 2009.

Results

Twenty-nine percent of hospices (169 of 591) reported employing an art, massage, or music therapist. Of those hospices, 74% employed a massage therapist, 53% a music therapist, and 22% an art therapist, and 42% expected the therapist to attend interdisciplinary staff meetings, indicating a significant role for these therapists on the patient’s care team. In adjusted analyses, larger hospices compared with smaller hospices had significantly higher odds of employing a CAM therapist (adjusted odds ratio (AOR) = 6.38, 95% CI 3.40, 11.99) and forprofit hospices had lower odds of employing a CAM therapist compared with nonprofit hospices (AOR = 0.52, 95% CI 0.32, 0.85). Forty-four percent of hospices in the Mountain/Pacific region reported employing a CAM therapist versus 17% in the South Central region.

Conclusion

Less than one-third of U.S. hospices employ art, massage, or music therapists despite the benefits these services may provide to patients and families. A higher proportion of large hospices, nonprofit hospices and hospices in the Mountain/Pacific region employ CAM therapists, indicating differential access to these important services.

Introduction

The integration of complementary and alternative medicine (CAM) with traditional medical care has been increasing, with interest from both patients and providers.^1,2 Complementary and alternative therapies focus on quality of life measures and are thus especially relevant in medical domains that focus on comfort care, such as palliative and hospice care.^3–5 For example, music therapy has been shown to improve pain, agitation, depression, and other quality of life measures in nursing home patients⁶ as well as those receiving home hospice care.⁷ Similarly, art therapy has demonstrated improvements in pain, fatigue, anxiety, and depression among palliative care inpatients, and massage has been shown to improve both psychological and physical well-being in patients with a variety of cancers.^8–10

Although interdisciplinary care is both fundamental to the hospice philosophy and a key component of high quality hospice care, CAM therapists are not a required component of the hospice interdisciplinary team under the Medicare Hospice Benefit.¹¹ In addition, because Medicare reimburses hospices on a per diem basis, hospices cannot separately bill Medicare for the services of CAM therapists. Existing evidence regarding the employment of CAM therapists at hospices is sparse and consists of only one national survey of hospices.¹² This study found that 55% of hospices had one or more salaried persons providing complementary therapies; however, the survey response rate was only 56%, the study was conducted more than a decade ago, and there is no information regarding the type of therapists employed. Smaller, regional studies found that the prevalence of CAM services in the hospice setting varied, with a range of 56% of hospices in Texas¹² to 86% and 90% percent in Washington¹³ and Illinois,¹⁴ respectively. In addition, a national report found that an estimated 42% of hospices nationally offered CAM services.¹ These studies, however, have differing definitions of complementary and alternative medicine and focus on the provision of CAM services (which could be by either volunteers or paid and licensed staff), rather than the employment of CAM therapists.

We sought to augment our understanding of the inclusion of CAM therapists on hospice interdisciplinary teams using recent national data on hospice staffing. Staffing is an important indicator of a hospice’s investment in the provision of quality care and has been found to vary across hospices nationwide.¹⁵ Although the impact on patients and families of a hospice’s investment in CAM therapists is not known, having more professionalized staff with more advanced educational and licensure backgrounds has been associated with better patient outcomes in other settings.^16–20

We surveyed a national sample of hospices regarding their use of CAM therapists. Specifically, our aims were to: 1) determine the extent to which hospices nationwide employ massage, music, and art therapists, and 2) estimate the association between the employment of CAM therapists and hospice characteristics including hospice ownership, size, chain status, proportion of at-home hospice care, years since hospice certification, concern over market share, and region. As both CAM and hospice service offerings increase, it is important to elucidate the use and variation in CAM services in hospices nationwide.

Methods

Study Design and Sample

We conducted a national cross-sectional study (“The National Hospice Study”) of a simple random sample of 775 hospices operating in the United States from September 2008 to November 2009. As described elsewhere,²¹ we chose our random sample from the 2006 Medicare Provider of Services file (N =3036 active hospices), which includes all hospices that participate in the Medicare program (approximately 93% of all hospices nationally). In addition, when the 2008 Medicare Provider of Services file (N =3306 active hospices) became available, we augmented our sample with hospices that were newly operating between 2006 and 2008. We estimated that 18% of hospices had been operating for two years or less. We then randomly selected 139 hospices (18% of our random sample of 775) from the 2008 Medicare Provider of Services file, to establish a total sample of 914 hospices.

We sent an introductory e-mail letter to each hospice medical director requesting his or her participation and a follow-up e-mail with a link to the web-based survey. Hospice medical directors were instructed to have the survey completed by the individual(s) at their hospice most knowledgeable about the survey questions.

Dependent Variables

To assess CAM staffing, we asked hospice providers, “On average over the past 12 months, how many of each of the following types of full-time equivalent (FTE) staff did your hospice employ or use as contracted staff?” We asked this question about the following types of CAM staff: art therapists, licensed massage therapists, and music therapists. These categories were chosen based on in-depth interviews with hospices during our survey development period. We identified a hospice as employing a CAM therapist if the hospice employed > 0 FTE of at least one of the following: art therapist, licensed massage therapist, and music therapist. We also asked survey respondents to identify which staff, “are expected to attend interdisciplinary team meetings (check all that apply).” The list of staff was: physicians, nurses, speech/occupational/ physical therapists, social workers, psychologists, art/music therapists, pharmacists, pastoral care/chaplains, volunteers, and administrators.

Independent Variables

The survey included questions regarding descriptive characteristics of hospices: size (number of patients per day in the past twelve months, which we categorized as <20, 20–49, 50–99, or 100 or greater to be consistent with existing studies reporting results from this survey^15,21–24 and national estimates of hospice size reported annually by the National Hospice and Palliative Care Organization²⁵), ownership (nonprofit, for-profit, government), whether the hospice was part of a chain of hospices, years since hospice certification (twp or fewer or more than two), proportion of the hospice’s patients receiving care at home, concern regarding market share (measured as: not at all concerned, slightly concerned, somewhat concerned, or very concerned), and census region.

Statistical Analyses

We calculated the proportion of hospices that reported employing > 0 FTE for at least one modality of CAM therapist (i.e., art, massage, and massage). We also calculated the proportion of hospices employing each type of therapist (i.e., art, massage, and music), both within all hospices and for only hospices that employed any CAM therapist. In addition, for hospices that did employ a CAM therapist, we calculated FTE for each individual therapy and all CAM services combined. For hospices that employed art or music therapists, we calculated the proportion of hospices that expected art/music therapists to attend interdisciplinary team meetings.

To test bivariate associations between whether or not a hospice employed CAM therapists and hospice size (measured categorically), ownership, chain membership, years since certification (two or fewer or more than two), concern over market share, and census region, we used Chi-square tests. In addition, for hospice size and market share, we conducted Cochran-Armitage tests for trend across categories. To assess the association between whether or not a hospice employed CAM therapists and hospice size (measured as a continuous variable) and the proportion of a hospice’s patients in home hospice, we used Wilcoxon rank-sum tests because both variables were not normally distributed.

For hospices that reported employment of > 0 FTE CAM therapists, we compared the distribution of FTE number of CAM therapists by categorical hospice characteristics using Wilcoxon rank-sum tests because the FTE number of CAM therapists was not normally distributed. In addition, to test for a linear trend in average FTE number of CAM therapists across hospice size and market share categories, we conducted Cochran-Armitage trend tests.

To estimate the hospice characteristics independently associated with employing CAM therapists, we calculated adjusted odds ratios (AOR) using modified Poisson regression models with a robust error variance. These multivariable regression models adjusted for hospice size, hospice ownership, whether the hospice was part of a chain of hospices, years since hospice certification, the proportion of patients cared for at home, concern over market share, and census region. We performed all analyses using the statistical software SAS, version 9.3.

Results

Study Population

Of the total 914 hospices randomly selected for the survey, 208 were excluded because they were no longer providing hospice care or had closed their facility at the time of the survey, resulting in 706 hospices eligible to respond. Of these 706 hospices, 591 completed our survey, for a response rate of 84%. Of those 591 hospices, 9% (53 of 591) did not respond to the CAM-related questions. Survey response rates differed by hospice ownership. Nonprofit hospices had an 89% response rate; government-owned hospices, an 86% response rate; and for-profit hospices, a 79% response rate (P = 0.004 for Chi-square comparison). Characteristics of our sample of 591 hospices are shown in Table 1.

Table 1.

Characteristics of Hospices (N = 591)

Characteristic	N (%)
Size (number of patients per day)
Less than 20	147 (25%)
20–49	154 (26%)
50–99	152 (26%)
100 or greater	127 (21%)
Missing	11 (2%)
Hospice Ownership
Nonprofit	283 (48%)
For-Profit	285 (48%)
Government/Other	23 (4%)
Hospice is a Member of a Chain
No	506 (86%)
Yes	85 (14%)
Years Since Hospice Certification
More than 2	523 (88%)
2 or fewer	68 (12%)
Patients in Home Hospice
Less than 75%	388 (66%)
75% or greater	190 (32%)
Missing	13 (2%)
Hospice concern over losing market share to competitors
Not at all concerned	36 (6%)
Slightly concerned	85 (14%)
Somewhat concerned	210 (36%)
Very concerned	257 (43%)
Missing	3 (0.5%)
Census Region
Mountain Pacific	103 (17%)
South Central	163 (28%)
South Atlantic	96 (16%)
North Central	161 (27%)
New England / Middle Atlantic	68 (12%)

Hospice Complementary and Alternative Medicine Staffing

A minority of hospices (m=169 or 29%) reported employing a CAM therapist (Fig. 1). Twenty-one percent (125 of 591) of hospices hired a massage therapist, 15% (90 of 591) hired an art therapist, 6% (37 of 591) hired a music therapist, and 9% (53 of 591) did not respond to this question in the survey. Of those hospices that did employ a CAM therapist, 74% (125 of 169) reported hiring a massage therapist, 53% (90 of 169) reported hiring a music therapist, and 22% (37 of 169) reported hiring an art therapist.

Figure 1. Proportion of Hospices that Employ Complementary and Alternative Medicinea Staff.

^a Complementary and Alternative Medicine (CAM) staff is defined as > 0 full time equivalent massage, music, or art therapist

For the subgroup of hospices that did employ a CAM therapist, the average number of FTE therapists employed was 1.55 (Table 2). The size of the hospice played a significant role in the number of FTE CAM therapists. The average CAM FTE of a hospice with 100 or more patients per day was 2.5, compared with 1.3 for hospices with 50–99 patients per day, 0.8 for hospices with 20–49 patients per day and 0.7 for hospices with 20 or fewer patients per day (P < 0.001 for Wilcoxon rank-sum test). Of hospices that employed any art or music therapists, 42% (44 of 106) expected the art and music therapists to attend interdisciplinary team meetings.

Table 2.

For Hospices that Employ Complementary and Alternative Medicine (CAM) Therapists: CAM FTE^a By Hospice Characteristics

	Complementary and Alternative Medicine FTE therapists
	Mean	Median	SD
Total	1.55	1.00	2.02
Size (number of patients per day)b
Less than 20	0.68	0.50	0.82
20–49	0.84	0.75	0.62
50–99	1.29	1.00	1.21
100 or greater	2.46	1.25	2.77
Hospice Ownershipb
Nonprofit	1.57	1.00	2.35
For-Profit	1.58	1.00	1.33
Hospice is a Member of a Chain
No	1.54	1.00	2.06
Yes	1.61	1.00	1.79
Years since Hospice Certification
More than 2	1.57	1.00	2.08
2 or fewer	1.33	1.00	0.93
Patients in Home Hospice
Less than 75%	1.77	1.00	2.29
75% or greater	1.01	1.00	0.86
Hospice concern over losing market share to competitors
Not at all concerned	1.61	1.00	1.80
Slightly concerned	1.68	1.00	1.71
Somewhat concerned	1.43	1.00	1.61
Very concerned	1.58	1.00	2.52
Census Region
Mountain Pacific	1.66	1.00	2.08
South Central	1.21	1.00	1.12
South Atlantic	2.77	1.00	4.01
North Central	1.28	1.00	1.37
New England/Middle Atlantic	1.43	1.00	1.36

^aCAM FTE is the full-time equivalent art, massage, and music therapists employed by a hospice

^bP<0.05 for a Wilcoxon Rank Sum Test

SD, standard deviation

Hospice Characteristics Associated with Employment of Complementary and Alternative Medicine Staff

In bivariate analyses, hospices that employed a CAM therapist had significantly higher odds of being larger and nonprofit, and CAM therapist employment also varied by region, extent of reported market share concern, and years since hospice certification. Specifically, having a CAM therapist was positively associated with hospice size (measured categorically, P<0.001 for Chi-square test). The Cochran-Armitage trend test across size categories also was significant (P < 0.001). The average size of hospices employing a CAM therapist was 153 patients per day, compared with 70 patients per day for hospices not employing a CAM therapist (P < 0.001 for Wilcoxon rank-sum test).

Significant ownership differences were apparent in the odds of a hospice employing a CAM therapist. Specifically, 40% of nonprofit hospices compared with only 24% of for-profit hospices reported hiring a CAM therapist (P < 0.001 for Chi-square test). Whether or not a hospice employed a CAM therapist also varied by a hospice’s concern over market share (P=0.018 for Chi-square test; P=0.003 for Cochran-Armitage trend test). For example, 27% of hospices that reported being “very concerned” about losing market share to competition employed a CAM therapist compared with 47% of those that reported being “not at all concerned.” Bivariate logistic regression results for employing a CAM therapist by hospice characteristics are shown in Table 3.

Table 3.

Unadjusted and Adjusted Associations Between the Presence of a Hired Complementary and Alternative Medicine Therapist and Hospice Characteristics

		Unadjusted		Adjustedd
	% Hiring CAM Therapistsa	Odds Ratio	95% CI	Odds Ratio	95% CI
Size (number of patients per day)
Less than 20	19%	1.00		1.00
20–49	27%	1.52	(0.86, 2.69)	1.86	(0.99, 3.50)
50–99	28%	1.67	(0.95, 2.94)	2.41	(1.28, 4.54)c
100 or greater	55%	5.03	(2.88, 8.79)c	6.38	(3.40, 11.99)c
Hospice Ownership
Nonprofit	40%	1.00		1.00
For Profit	24%	0.47	(0.32, 0.69)c	0.52	(0.32, 0.85)b
Hospice is a Member of a Chain
No	27%	1.00		1.00
Yes	32%	1.04	(0.62, 1.73)	1.07	(0.58, 1.96)
Years Since Hospice Certification
More than 2	34%	1.00		1.00
2 or fewer	20%	0.48	(0.29, 0.81)c	0.84	(0.39, 1.81)
Hospice concern over losing market share to competitors
Not at all concerned	47%	1.00		1.00
Slightly concerned	43%	0.85	(0.36, 1.96)	0.91	(0.33, 2.48)
Somewhat concerned	30%	0.49	(0.22, 1.07)	0.46	(0.18, 1.19)
Very concerned	27%	0.42	(0.20, 0.92)b	0.35	(0.14, 0.89)b
Census Region
Mountain Pacific	44%	1.00		1.00
South Central	17%	0.25	(0.14, 0.45)c	0.33	(0.17, 0.64)c
South Atlantic	24%	0.39	(0.20, 0.76)c	0.31	(0.15, 0.64)c
North Central	39%	0.81	(0.48, 1.35)	0.91	(0.49, 1.66)
New England / Middle Atlantic	38%	0.76	(0.40, 1.45)	0.77	(0.36, 1.61)

^aComplementary and Alternative Medicine (CAM) therapist is defined as > 0 full time equivalent art, massage, and music therapists

^bp value < 0.05

^cp value < 0.01

^dMultivariable model includes all variables shown in this table simultaneously entered as predictors in addition to the log transformed proportion of a hospice’s patients receiving care at home.

CI, confidence interval

In multivariable analyses, the significant associations found in the bivariate analyses generally persisted. Larger hospices had significantly higher odds of employing a CAM therapist compared with smaller hospices (Table 3). Nonprofit hospices had nearly twice the odds as forprofit hospices of employing a CAM therapist (AOR of for-profit compared with nonprofit hospices was 0.52, 95% CI 0.32, 0.85). Hospices that were very concerned with losing market share to competitors had substantially lower odds of employing a CAM therapist compared with hospices that were not concerned with losing market share (AOR 0.35, 95% CI 0.14, 0.89). Results from multivariable models that included hospice size and concern regarding market share measured as ordinal variables did not materially differ from those presented in Table 3. The AOR for hospice size measured as a categorical ordinal variable was 1.81 (95%CI 1.49, 2.21) and the AOR for market share as a categorical ordinal variable was 0.67 (95%CI 0.53, 0.84).

Discussion

We found that less than one-third of U.S. hospices employed an art, music, or massage therapist either part-time or full-time. Of hospices that did employ a therapist, the average full-time equivalent staff across art, massage, and music therapy was only 1.6. This result is surprising given the focus of hospice on interdisciplinary care and the symptom management that CAM therapy has been found to improve.

One likely reason that CAM therapists are not employed more widely in hospices is cost. Hospices are currently paid by Medicare on a per diem basis ($153/day in 2013).²⁶ Many hospices report that the current per diem reimbursement is not adequate to cover their cost of care²⁶ and hospices have indicated that limited funding is a significant barrier to offering CAM services.^14,27,28 This is despite the fact that research suggests that CAM therapies may be cost-effective care modalities.^29,30 Our results are consistent with studies of acute care hospitals that have found that those with greater financial resources are significantly more likely to offer CAM services to their patients.³¹ Our finding that larger hospices had significantly higher odds of employing CAM therapists than smaller hospices is likely because of economies of scale at larger hospices that enable them to spread the cost of FTE CAM therapists across a large patient population. Similar results suggest that economies of scale at large hospices enable them to provide higher quality hospice care as measured by greater implementation of preferred practices,²¹ a more comprehensive bereavement program,²⁴ lower patient disenrollment rates,^32–34 and more open enrollment policies for complex and potentially costly patient populations.²³ Although our survey did not measure the use of volunteers to provide CAM services, this may be a way for hospices to address the financial barriers of hiring CAM staff, particularly for smaller hospices.

Our finding that for-profit compared with nonprofit hospices had lower odds of employing CAM therapists is consistent with a growing body of evidence indicating important variation by hospice ownership in hospice care processes and patient outcomes including the range of services received by patients,³⁵ staffing ratios,¹⁵ patient disenrollment rates,¹⁵ and bereavement services.¹⁵ Our results differ from a recent analysis of hospice staffing,¹⁵ which found that forprofit hospices employed a wider range of hospice staff; however, that study did not include CAM therapists.

Consistent with findings in the literature,³¹ we anticipated that hospices operating in more competitive environments would be more likely to offer non-required services, such as CAM therapies, to their patients. In contrast, our results showed an inverse relationship between a hospice’s concern over market share and the odds that it employed CAM therapists. This relationship may be caused by the increased financial strain placed on hospices in more competitive markets. A case study of a large Midwestern hospice found that increased competition in a market coincided with increased economic pressure for hospices, resulting in a more narrow scope of services delivered to patients and families.³⁶

Our findings should be interpreted in light of several limitations. First, data were self-reported by hospices, and we were not able to validate the staffing of CAM therapists. Second, although the overall survey response rate was high (84%), for-profit hospices were significantly less likely to respond than non-profit hospices. Given we found that for-profit hospices had lower odds of employing CAM therapists compared with nonprofit hospices, our result that only 29% of hospices employ a CAM therapist may be overestimated. We could not test for differences in survey response rate by hospice size; however, the distribution of our sample by hospice size is similar to the distribution nationally.²⁵ We also do not have information on the specific responder to the hospice survey although hospice medical directors were instructed to have the survey completed by the individual(s) at their hospice most knowledgeable about the survey questions. Fourth, our survey provides data on only the hiring of CAM therapists, not the provision of CAM services, which would be helpful in future studies. Finally, our survey did not ask if licensed massage therapists attended interdisciplinary team meetings and future research regarding hospice interdisciplinary team meetings should include this important group.

Research on the quality of hospice care has demonstrated that innovative practices, which might include CAM therapy, are a major determinant of improved overall quality in hospice care.^37,38 Our finding that less than one-third of hospices in the U.S. employ CAM therapists suggests that patients and families have differential access to these therapies, despite their potential to improve pain, symptoms, and other measures of quality of life for patients and families. Future studies to elucidate why some hospices employ CAM therapists and others do not and the impact of such therapists on patient and family experiences at the end of life are warranted.