This Article Abstract Full Text (PDF) Services Download to citation manager PubMed PubMed Citation

Characterizing Hospice Services in the United States

Correspondence: Address correspondence to Dr. Maureen A. Smith, Department of Population Health Sciences, University of Wisconsin School of Medicine and Public Health, 505 WARF Building, 610 Walnut Street, Madison, WI 53726. E-mail: maureensmith{at}wisc.edu

	Abstract

TOP Abstract Methods Results Discussion References

 
Purpose: Although caregivers desire specific information about hospice programs, there is little descriptive information available. We characterized agencies that provide formal or informal hospice care in the United States according to four types of services considered important by caregivers: medications and treatments; rehabilitative care; emotional, social, and spiritual support; and practical support (e.g., continuous home care). Design and Methods: Data were from the nationally representative 2000 National Home and Hospice Care Survey. We categorized agencies into service mixes reflecting combinations of the four service types by using a grade-of-membership model to score each agency. Results: Of the 11,419 agencies represented in our data, slightly more than half (52%) reflected some mix of services from all four service types. The remaining agencies provided service mixes that reflected relatively few or no services from at least one of the four service types. Specifically, approximately 7% lacked (i.e., had relatively few or no) medications and treatment services, 6% lacked rehabilitative care, 26% lacked emotional/social/spiritual support, and about 10% lacked multiple services. When compared to agencies that reflected a mix of all four services, agencies that lacked multiple services could be distinguished by their lack of formal certification as either a hospice or home health agency. Implications: Few U.S. agencies that provide formal or informal hospice care incorporate a balanced mix of all types of services, but most agencies lack only one type of service. Categorization of agencies based on services that are relevant to caregivers may assist providers in describing hospice care.

Key Words: Hospices • Caregivers • End of life • Spirituality • Social support

A major reason for underutilization of hospice care is late referral by physicians to hospice. The median length of stay of a patient in a hospice is only 15.6 days, and for many it is only several days (Haupt, 2003). This occurs even though the Medicare Hospice Benefit reimburses up to 6 months of care in Medicare-certified hospices (Emanuel et al., 2002; Rickerson, Harrold, Kapo, Carroll, & Casarett, 2005), some states classify patient eligibility as a life expectancy of less than 12 months, and care in noncertified hospices may be available through private insurance or other parts of the Medicare program such as the Medicare Home Health Benefit (Hamilton, 1993). Lack of easily available information about hospice is a major reason why physicians are uncomfortable discussing hospices and tend to bring the issue up very late (Ahmed et al., 2004; Friedman et al., 2002). This may include lack of information about local hospices or variability in how hospice services are provided.

In spite of research describing patients' desire for specific information (Casarett et al., 2005), there is little descriptive information available to physicians about hospice care. It is important to note that the role of agencies not formally certified as hospices in caring for dying patients may be critical. As identified by Kane (1999), dying patients may receive therapies through home care agencies. However, studies have often focused on agencies that have been formally certified as hospices by Medicare or Medicaid. Nonproprietary (i.e., nonprofit) Medicare and Medicaid certified hospices provide more comprehensive care (Carlson, Gallo, & Bradley, 2004) than for-profit certified hospices, but no studies have examined how existing agencies in the United States match what Casarett and colleagues (2005) identified as important to patients and their families as they contemplate hospice enrollment (e.g., availability of medical treatments, emotional and social support). The goal of the present study was to build upon this work and address a major gap in knowledge by characterizing agencies that provide formal and informal hospice, beginning with the Casarett and associates (2005) framework, to identify services that patients and their families consider important to decision making. To conduct this study, we used data from a nationally representative survey of hospices and home health agencies.

	Methods

TOP Abstract Methods Results Discussion References

 
Data
We analyzed data from the 2000 agency file of the National Home and Hospice Care Survey (NHHCS). The National Center for Health Statistics collects data for the NHHCS biannually from a nationally representative probability sample of home health and hospice agencies and their patients (Haupt, 1994). Data are obtained through personal interviews with hospice administrators and staff of selected agencies using a two-stage, stratified sampling design. The participation rate in the 2000 wave was 96%, with 1,425 agencies responding. These 1,425 agencies represented 11,419 agencies in the United States.

Variables
Casarett and colleagues (2005) originally developed eight categories of services based on interviews of 100 bereaved caregivers describing positive and negative features that influenced their decision to enroll in hospice (Casarett et al., 2005; Teno et al., 2004). In that study, interviews were audiotaped, transcribed, and coded. Three investigators grouped the coded information into 40 categories using techniques derived from grounded theory. A consensus process was used to reduce the number of categories to 12, and an iterative process was used to group 3 categories and drop 2, leaving 8 categories of information.

For this study, we excluded one category suggested by Casarett and associates (2005; source of payment) as it was a characteristic of the individual patient, not of the hospice. We excluded a second category because no NHHCS variables indicated whether "alternative locations were available if the patient could not stay at home." The six remaining categories were medication and treatments, seeing the same doctors, emotional and social support, spiritual support, sending someone to visit, and practical support and assistance (Casarett et al., 2005; Teno et al., 2004).

We grouped 26 service variables recorded in the NHHCS (Haupt, 1994) into these six categories of information. We excluded one NHHCS variable, skilled nursing services, because these services were available through almost all agencies (approximately 93%) and therefore it was not a useful variable on which to classify agencies. Because only one NHHCS variable, physician services, indicated whether patients would likely see the same doctors, we combined this variable into the medications and treatments category. Five categories remained: medications and treatments, emotional and social support, spiritual support, sending someone to visit, and practical support. Because medications and treatments had a large number of items, we separated it into two categories (medications and treatments, and rehabilitative care). Finally, because spiritual support had only two items, we combined it with emotional and social support. It was difficult to distinguish services that identified the Casarett and colleagues (2005) categories sending someone to visit and practical support, so we combined these two categories. The result was four main groups of services that agencies could offer.

Once we had categorized agencies by their mix of services provided, we examined the association between the agency service mix categories and organizational and other descriptive characteristics. These included type of ownership (proprietary or not); whether the agency was operated by a hospital, nursing home, or health maintenance organization (HMO); the agency's Medicare/Medicaid hospice and home health certification status (certified as neither, certified as home health agency only, certified as hospice agency only, certified as both); whether the agency currently had computerized medical records or planned to have them; total number of active patients of any kind; geographic region; and whether the agency was in a metropolitan service area. A metropolitan service area is a U.S. Census Bureau-defined urbanized area of at least 50,000 inhabitants with a total metropolitan population of at least 100,000.

Analysis
We used a two-step approach to categorize agencies. First, we used a grade of membership (GOM) model (Manton, Woodbury, & Tolley, 1994) to summarize the service mix of each agency. We achieved this by initially defining four "idealized" agencies comprising all individual services within one (and only one) of the four types of services (medications and treatments; rehabilitative therapy; emotional, social, and spiritual support; and practical support). Specifically, we defined each of the four idealized agencies as having all of the services of one service type, and none of the services from any other service type (e.g., all of the medications and treatment services, and no rehabilitative therapy, etc.). These idealized agencies served as reference points against which we could compare each actual agency and motivated the categorization of actual agencies (Casarett et al., 2005; Teno et al., 2004), as described previously.

We then used the GOM model to estimate four scores for each agency in the data set, such that each score represented the similarity of a given agency to each of the four idealized reference agencies. According to the GOM model, each score was between 0 and 1, and together all four scores were constrained to sum to 1 for a given agency (see Manton et al., 1994, for details of GOM model estimation). For example, an agency that provided some emotional, social, and spiritual services; some practical support services; and no other types of services (i.e., provided services from the third and fourth service categories and none from the first and second service categories) could receive a score set of, for example, (0, 0, 0.43, 0.57). Alternatively, an agency that provided several services from all four service categories could receive a score set of, for example (0.30, 0.26, 0.22, 0.22). Thus, in contrast to a traditional cluster analysis, the GOM model did not assign each agency to a single group but instead allowed membership in multiple groups by scoring each agency according to the relative number of services it provided from each of the service categories. All GOM analyses used the software developed by Charpentier (1996).

The second step of our analysis categorized agencies into service mixes based on their four GOM scores (e.g., reflected all four services, had relatively few or no rehabilitative services, etc.). We examined the distribution of the scores and identified the service mix combinations based on the approach developed by Berkman, Singer, and Manton (1989) and Singer (1989). Briefly, we identified an agency as essentially reflecting a single service type if the single score corresponding to that service type was greater than or equal to 0.9. Similarly, we identified an agency as reflecting a particular combination of two service types if the sum of the two scores corresponding to the two service types was greater than or equal to 0.9. We defined combinations of three service types or, ultimately, all four service types in a similar fashion. Less than 1% of agencies had scores that reflected mixes of more than one 3-service type—these were categorized as four service type mixes. Note that classification of a given agency as "lacking" a particular service type (or multiple types) did not preclude that agency from providing any services of that type, only that such services composed few (or none) of their offerings relative to the total number of services offered from the other service types. After GOM estimation, we applied the survey design weights for the NHHCS to the distribution of scores to generate population frequencies of each service mix (Manton et al., 1994).

Lastly, we identified frequent hospice types and grouped the remainder as "other." We used multinomial logistic regression to estimate the relationship between hospice organizational characteristics and hospice type, controlling for other descriptive characteristics. We calculated 95% confidence intervals. The multinomial logit estimation employed the survey design weights for the NHHCS. After we excluded observations that were missing responses on variables included in our multinomial logistic regression, the survey contained information on 1,310 agencies representing a weighted population of 10,464 agencies. We performed all analyses in the second step using Stata statistical software (StataCorp, 2003).

	Results

TOP Abstract Methods Results Discussion References

 
Description of Services Provided
As shown in Table 1, there was a wide range of services offered by certified agencies providing formal and informal hospice care in the United States in 2000. One third of agencies offered durable medical equipment, and a similar proportion offered medications. The large majority of agencies offered intravenous therapy, whereas about half offered dietary and nutritional services and enterostomal therapy. About three quarters offered occupational, physical, and speech therapies. One third to one half of agencies offered counseling, bereavement care, pastoral care, and spiritual care. Slightly less than half of agencies offered continuous home care, whereas almost all offered personal care.

	Discussion

TOP Abstract Methods Results Discussion References

Our approach is distinct from other approaches used to classify hospice services in that it offers additional information about the structure of formal and informal hospice care from the perspective of patients and families. For example, Medicare classifies hospice services into core and noncore services: Core services (e.g., skilled nursing, counseling) are provided directly by hospice staff, whereas noncore services (e.g., continuous home care, physical therapy, high-tech services) can be outsourced. Previous studies using the Medicare classification have concluded that patients of proprietary (for-profit) Medicare-certified hospices receive a significantly narrower range of services when compared to patients of nonprofit Medicare-certified hospices, specifically those services designated as noncore or discretionary (Carlson et al., 2004). We did not confirm this finding in this larger sample of agencies that provided both formal and informal hospice services; in this larger sample, the primary distinguishing characteristic of agencies that lacked services was a lack of certification as either a hospice or home health agency.

Agencies lacking multiple services were more likely to lack certification as a hospice (e.g., were certified as a home health agency only). Previous studies of home health agencies have suggested that the multiple tasks inherent in home care complicate general home care goal setting. In particular, the goal of therapeutic improvement may conflict with the goal of compensation for functional impairments (Kane, 1999). Both of these home care goals may overlap with the goals of hospice-provided palliative care, but hospice may have additional objectives, such as a comfortable death. These hospice-specific objectives may be more difficult to achieve in a setting that must develop and maintain resources and systems to achieve multiple goals for a variety of clients.

Iwashyna, Chang, Zhang, and Christakis (2002) found substantial variation in the use of hospices across markets, but health care infrastructure did not explain this variation. The authors concluded that their findings might be explained by important local factors such as differences in the mix of services provided by local hospices, local preferences, or community leadership, although their data did not allow them to distinguish among these hypotheses. Although our data also do not allow us to distinguish among these hypotheses, we do find evidence that agencies differ significantly in service mix and that these differences go beyond the for-profit/nonprofit distinction that researchers have previously studied. In particular, agencies lacking multiple services are much more likely to lack certification as a hospice. If significant care for the dying is provided through these noncertified settings, it raises questions about the availability of important services for these patients.

It is important to note that agencies may match the services provided to the needs of their communities and that lacking certain services may not indicate a deficit in desired care from the patient or family perspective. Alternatively, although there is evidence that patients and families value emotional/social and spiritual support at the end of life (Teno et al., 2004), information about these services appears to be less relevant to the initial decision to enter hospice (Casarett et al., 2005). This may suggest that hospices are more likely to focus on offering services critical to the initial decision to enroll.

Some have suggested that a major barrier to physician-initiated discussions of hospice is the lack of easily available information on hospices, which leads to increased physician discomfort and very late referrals to hospice care (Ahmed et al., 2004; Friedman et al., 2002). This may be due to the lack of information about local hospices or the variability in how hospice services are provided. Additional factors that impede discussion of hospice are discomfort on the part of physicians with discussing the terminal phase of illness, and reluctance of the patient and/or family to accept that the terminal phase has been reached (Ahmed et al., 2004). The availability of descriptive information and familiarity with the types of services provided through formal or informal hospice care may reduce this discomfort on the part of physicians and patients/families.

During hospice discussions, patients often do not recall careful consideration of specific hospice services, and many services are appreciated only after enrollment in the hospice (Casarett, Crowley, & Hirschman, 2004). In general, a consideration of hospice benefits is less important than the perceived lack of alternative options (Casarett et al., 2004). Because the hospice decision is complex and requires that patients and their families process large amounts of information in a short period of time, simplified descriptive information on agencies providing formal and informal hospice services may assist providers in describing hospice care options. The descriptive model presented in this analysis may provide a useful framework for these discussions, as we developed the model using service categories that have been identified as important to patients and their families (Casarett et al., 2005).

This study has several limitations. We used a model of information needs developed from interviews with patients referred to a single urban hospice (Casarett et al., 2005). This may limit generalizability, although the patients interviewed in the original study represented referrals from a large number of physicians and care sites. We were limited to service and organizational characteristics that were recorded as part of the NHHCS study (Haupt, 1994). Although the NHHCS is comprehensive, it is likely that it omitted relevant service characteristics, particularly for the categories of practical support and emotional, social, and spiritual support. For example, we had no detailed data on information services (e.g., information about resources, referrals, appointment making) or caregiver instruction (Kane, 1999). Additional omitted organizational characteristics included the date of establishment of the agency, as well as its relationships with hospitals, nursing homes, and health agencies beyond ownership (Casey, Moscovice, Virnig, & Durham, 2005). Future studies should focus on identifying additional service and organizational characteristics that are important to patients and their families. Lastly, as is the case for all categorization schemes that attempt to summarize a large number of varied entities into a small number of simplified groups, our approach to categorizing services using the GOM model has limitations. It is possible that a few agencies were not scored or classified precisely, particularly those agencies that offered a small total number of services, because by definition such agencies offer little by which to distinguish them. Our findings should be validated with additional research and alternative approaches.

Casarett and colleagues (2005) suggested that by anticipating common information needs, physicians might better support patients and families in their decision making about hospice care. Based on information that patients and their families found valuable during an initial discussion of hospice care, we characterized agencies that provide formal and informal hospice care in the United States using data from a nationally representative survey. Because our approach utilized the patient perspective, we provide an initial assessment of potential gaps in services according to geographic and organizational characteristics. Additional research is critical to determine whether these gaps represent a mismatch between available services and services desired by patients, or whether variation in the mix of services provided by agencies is representative of variation in patient preferences for those services.

	Footnotes

 
All work should be attributed to the University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin.

¹ Department of Population Health Sciences, University of Wisconsin School of Medicine and Public Health, Madison.

² Center on Aging and Health and Department of Population, Family and Reproductive Health, Johns Hopkins University, Baltimore, MD.

³ University of Wisconsin School of Medicine and Public Health, Health Sciences Learning Center, Madison.

⁴ Department of Medicine and Palliative Care Program, Comprehensive Cancer Center, University of Wisconsin Hospitals and Clinics, Madison.

⁵ Department of Medicine, University of Wisconsin School of Medicine and Public Health, Madison.

Decision Editor: Linda S. Noelker, PhD

Received for publication November 17, 2006. Accepted for publication April 2, 2007.

	References

TOP Abstract Methods Results Discussion References

 

• Ahmed, N., Bestall, J. C., Ahmedzai, S. H., Payne, S. A., Clark, D., & Noble, B., (2004). Systematic review of the problems and issues of accessing specialist palliative care by patients, careers and health and social care professionals. Palliative Medicine, 18, 525-542.[Abstract/Free Full Text]
• Berkman, L. F., Singer, B. H., & Manton, K. G., (1989). Black/white differences in health status and mortality among the elderly. Demography, 26, 661-678.[Medline]
• Carlson, M. D., Gallo, W. T., & Bradley, E. H., (2004). Ownership status and patterns of care in hospice: Results from the National Home and Hospice Care Survey. Medical Care, 42, 432-438.[Medline]
• Casarett, D., Crowley, R., & Hirschman, K. B., (2004). How should clinicians describe hospice to patients and families? Journal of the American Geriatrics Society, 52, 1923-1928.[Medline]
• Casarett, D., Crowley, R., Stevenson, C., Xie, S., & Teno, J., (2005). Making difficult decisions about hospice enrollment: What do patients and families want to know? Journal of the American Geriatrics Society, 53, 249-254.[Medline]
• Casey, M. M., Moscovice, I. S., Virnig, B. A., & Durham, S. B., (2005). Providing hospice care in rural areas: Challenges and strategies. American Journal of Hospice & Palliative Care, 22, 363-368.[Medline]
• Charpentier, P., (1996). GOM3, Version 3.4. Retrieved December 15, 2002, from http://lib.stat.cmu.edu/DOS/general/.index.html.
• Emanuel, E. J., Ash, A., Yu, W., Gazelle, G., Levinsky, N. G., & Saynina, O., et al (2002). Managed care, hospice use, site of death, and medical expenditures in the last year of life. Archives of Internal Medicine, 162, 1722-1728.[Abstract/Free Full Text]
• Friedman, B. T., Harwood, M. K., & Shields, M., (2002). Barriers and enablers to hospice referrals: An expert overview. Journal of Palliative Medicine, 5, 73-84.[Medline]
• Hamilton, V., (1993). The Medicare hospice benefit: The effectiveness of price incentives in health care policy. Rand Journal of Economics, 24, 605-624.[Medline]
• Haupt, B. J., (1994, September). Development of the National Home and Hospice Care Survey (Vital and Health Statistics Series 1, No. 33). Hyattsville, MD: National Center for Health Statistics.
• Haupt, B. J., (2003, August). Characteristics of hospice care discharges and their length of service: United States, 2000 (Vital and Health Statistics Series 13, No. 154). Hyattsville, MD: National Center for Health Statistics.
• Iwashyna, T. J., Chang, V. W., Zhang, J. X., & Christakis, N. A., (2002). The lack of effect of market structure on hospice use. Health Services Research, 37, 1531-1551.[Medline]
• Kane, R. A., (1999). Goals of home care: Therapeutic, compensatory, either, or both? Journal of Aging and Health, 11, 299-321.[Abstract/Free Full Text]
• Manton, K. G., Woodbury, M. A., & Tolley, H. D., (1994). Statistical applications using fuzzy sets. New York: Wiley.
• Ogle, K. S., Mavis, B., & Wyatt, G. K., (2002). Physicians and hospice care: Attitudes, knowledge, and referrals. Journal of Palliative Medicine, 5, 85-92.[Medline]
• Rickerson, E., Harrold, J., Kapo, J., Carroll, J. T., & Casarett, D., (2005). Timing of hospice referral and families' perceptions of services: Are earlier hospice referrals better? Journal of the American Geriatrics Society, 53, 819-823.[Medline]
• Singer, B., (1989). Grade of membership representations: Concepts and problems. In T. W. Andersen, K. B. Athreya, & D. L. Iglehart (Eds.), Probability, statistics, and mathematics: Papers in honor of Samuel Karlin (pp. 317–334). New York: Academic Press.
• StataCorp. (2003). Stata statistical software: Release 8.2 (Version 8.2). College Station, TX: Author.
• Teno, J., Clarridge, B. R., Casey, V., Welch, L. C., Wetle, T., & Shield, R., et al (2004). Family perspectives on end-of-life care at the last place of care. Journal of the American Medical Association, 291, 88-93.[Abstract/Free Full Text]

This Article Abstract Full Text (PDF) Services Download to citation manager PubMed PubMed Citation