A Focus on the Program of All-Inclusive Care for the Elderly

Originally published by the Journal of Aging and Social Policy in 2017

By Lori Gonzalez


For over four decades, the Program of All-Inclusive Care for the Elderly (PACE) has been operated by nonprofit organizations. Research has demonstrated that nonprofit PACE provides quality, cost-effective community-based care to older adults who would otherwise require a nursing home level of care. Recently, the U.S. Secretary of the Department of Health and Human Services has authorized for-profit entities to operate PACE, contingent on their ability to demonstrate that they can provide care that is similar to nonprofit PACE with regard to access to care, quality of care, and cost-effectiveness. In 2013, a study was conducted to evaluate how PACE operates under for-profit versus nonprofit status. The results were presented to Congress which, in turn, authorized for-profit PACE providers. This article critiques the 2013 study, offers a comparison to for-profit hospice, and argues that at best there is not enough evidence to conclude that for-profit PACE provides the same quality of care as existing nonprofit operators.

Providing quality, comprehensive, and cost-effective services for elders has been a challenge in long-term care. The Program of All-Inclusive Care for the Elderly (PACE) provides a continuum of medical and social services to older adults who qualify for nursing home care but wish to remain in the community. Beginning in the 1970s, PACE has been operated solely by not-for-profit organizations that receive a flat fee (capitated reimbursement rate) from Medicare, Medicaid, and/or private payers to cover the cost of care. In 2009, the Centers for Medicare and Medicaid Services (CMS) issued a call for proposals to develop a for-profit PACE program that could provide care on par with existing nonprofit providers. After a for-profit demonstration in Pennsylvania and a corresponding 2013 study submitted by Mathematica to CMS and Congress reported no significant differences in access, quality, or cost, CMS authorized the states to allow for-profit PACE operators. Colorado has already passed laws that allow for-profit PACE providers, and California and New Mexico are expected to follow suit (Wallace, 2015).

This paper examines recent movement from nonprofit to for-profit PACE and compares this shift to an analogous one experienced in the hospice industry. First, the PACE program is described. Next, the for-profit PACE demonstration is discussed. Then a comparison is made between the advent of for-profit PACE and similar changes that took place in the hospice industry. Finally, the paper concludes with policy recommendations.

A primer on PACE

In 1973 the Chinatown community of San Francisco, California, opened On Lok—the nation’s first PACE adult day care center. The program was designed to provide social day care in the center and in-home care and meals to the frail elderly in the community (National PACE Association, 2015). On Lok primarily served the immigrant population, consisting largely of Chinese, Italian, and Filipino Americans and their families, a group that valued community care, believing that nursing homes were unacceptable places for their loved ones to age (Greenwood, 2001). On Lok, therefore, provided a natural fit between the care that elders and their families required and their cultural views on where that care should take place.

Today, PACE operates 116 programs in 32 states, guided by the continuing belief that “it is better for the well-being of seniors with chronic care needs and their families to be served in the community whenever possible” (National PACE Association, 2015). Each PACE program serves between 100 and 2,500 elders for a total of over 30,000 enrollees in the United States (National PACE Association, 2015). Participants’ needs and preferences inform the written care plan for each individual, including the frequency of visits to the PACE center (CMS, 2011). Eligibility for PACE requires that individuals be 55 years old or older, certified by the state to need a nursing home level of care, reside near a PACE program, and be able to live safely in the community. The typical PACE enrollee is 80 years old, has 7.9 medical conditions, and is limited in about 3 activities of daily living (ADLs); close to half have a diagnosis of dementia. Yet, with the assistance of the PACE program, 90% of participants who might otherwise enter a nursing home are able to live in the community—a clear preference common among elders (National PACE Association, 2015).

In PACE, interdisciplinary teams consisting of primary care physicians, nutritionists, nurses, social workers, and representatives of other disciplines meet with enrollees and their families to assess the services needed to provide comprehensive care and allow participants to remain in the community. PACE services include primary care, hospital care, medical specialty services, nursing home care, emergency services, home care, physical therapy, occupational therapy, adult day care, dentistry, social services, and transportation. Families of enrollees also receive services, such as counseling or instruction on how to care for their loved ones. Care is provided at PACE centers or at home or in the community via contracts with other providers. When an enrollee comes to a PACE center, they might, for example, receive medication, counseling, dialysis, a hot meal, or a flu shot. There are also opportunities to socialize with other enrollees and PACE staff during the visit.

Research on PACE’s effectiveness

Extant evidence suggests that PACE programs are effective in their primary goal of allowing frail elders to remain safely in the community. Studies comparing PACE participants with similar individuals have found that PACE participants are less likely to be admitted to a nursing home or hospital, more likely to experience delays before entering a nursing home, and more likely to die at home than in a hospital (Beauchamp, Cheh, Schmitz, Kemper, & Hall, 2008; Chatterji, Burstein, Kidder, & White, 1998; Kane, Homyak, Bershadsky, & Flood 2006a2006b; Segelman, Cai, Van Reenen, & Temkin-Greener, 2015; Temkin-Greener & Mukamel, 2002). A recent study, for example, found that PACE participants were less likely to be institutionalized than their 1915(c) home- and community-based services waiver counterparts (Segelman et al., 2015). Further, when they were admitted to the nursing home, they were frailer, suggesting that PACE was able to keep them out of the nursing home until it was absolutely necessary.

PACE programs are viewed favorably by participants and their families. PACE enrollees report higher satisfaction with the care they receive compared to those who are eligible but not enrolled (Chatterji et al., 1998). Family members and PACE participants rank quality of care high, including with respect to feelings of safety and security, being treated with respect, and control over decision making (Mitchell, Polivka, & Wang, 2008). PACE participants also report improved quality of life as reflected in increased social interaction, less depression, and fewer worries after enrollment (Beauchamp et al., 2008). Some enrollees, however, report being less satisfied with certain aspects of the PACE experience, including the amount of information provided on their conditions, having to give up their primary care physician, and a lack of concern among providers for participants’ input into their own care (Beauchamp et al., 2008). On balance, though, most research indicates that PACE provides quality care that both participants and their families value.

Research on PACE’s cost-effectiveness

Research on PACE’s cost-effectiveness has yielded mixed results. For instance, a Washington state study found that the per-member, per-month Medicaid acute medical and long-term care expenditures of PACE participants were $1,442 higher than those receiving 1915(c) home- and community-based waiver program services (Mancuso, Yamashiro, & Felver, 2005) but that the gap narrowed to $1,018 after 4 years due to rising nursing home costs among those initially receiving home- and community-based services. Moreover, the same study found comparable costs between PACE and nursing home residents. Another study that compared Medicare and Medicaid expenditures among PACE enrollees and non-PACE enrollees in nine states found that while Medicare expenditures were similar for both groups, Medicaid expenditures were higher for PACE enrollees, resulting in overall higher costs for PACE (Foster, Schmitz, & Kemper, 2007). Once again, however, researchers found that this difference in expenditures declined over time from an average of $962 per person per month in the first 6 months to an average of $536 in months 19 to 24.

In contrast, three state assessments found that PACE provides significant cost savings relative to alternative forms of care (Damons, 2001; Rylander, 2000; Weiland et al., 2013). In Tennessee, Medicaid expenditures for PACE enrollees were estimated to be about 17% lower than Medicaid expenditures for those enrolled in managed care and receiving nursing facility services. In Texas, savings to Medicare and Medicaid were estimated to be about 14% lower when comparing PACE enrollees to comparable populations residing in nursing homes and the community. In South Carolina, researchers estimated that state payments for comparable fee-for-service Medicaid beneficiaries were 28% higher than payments made under PACE’s Medicaid capitated rate (Weiland et al., 2013).

Taken together, extant evidence indicates that although PACE might be costlier than other programs initially, over time its expenditures are similar to (or even less than) alternatives.

Findings regarding PACE and its quality of care and cost-effectiveness have been based on studies of the program under nonprofit ownership. The next three sections describe the shift to for-profit PACE programs and the results of the only study to date that compared outcomes between the two models.

For-profit PACE and the Balanced Budget Act

PACE has operated since the 1970s as a nonprofit program. Initially, PACE was authorized as a research and demonstration program under federal Medicare and Medicaid waivers. It became a formal Medicare program benefit and state option under Medicaid with passage of the Balanced Budget Act (BBA) of 1997 (Public Law 105–33, 4801–4804). A 1999 interim final regulation implementing this provision of the BBA officially established nonprofits as permanent PACE program providers under Medicare and Medicaid as per sections 1894(a)(3)(A)(i) and 1934(a)(3)(A)(i) of the act. Although permanent PACE sites are required to be operated by nonprofit entities, sections 1894(h) and 1934(h) of the act allow the secretary of the U.S. Department of Health and Human Services (DHHS) to establish for-profit PACE sites, but on a research and demonstration basis only.

Under the BBA, the secretary of the DHHS could establish for-profit PACE sites as permanent providers, but only after an evaluation study of their cost-effectiveness, quality, and access to care had taken place (Federal Register Notices, 2001). Several parameters for the evaluation were established in the legislation. In particular, the final rule implementing this provision required that the total number of enrollees in the demonstration evaluation be at least 800 (or a number that was still statistically sufficient for an evaluation as determined by the DHHS secretary) and that participating plans be in operation for 3 years before they could be included in the evaluation. It also required that the evaluation assess whether or not enrollees in for-profit PACE were frailer than enrollees in nonprofit PACE, whether access to and quality of care was lower in for-profit than nonprofit PACE, and whether the cost to Medicare and/or Medicaid was higher in for-profit PACE compared to nonprofit PACE.

The first for-profit PACE demonstration program began in 2007 in Pennsylvania. Three others, also in Pennsylvania and operated by the same company, followed in 2011. Mathematica contracted with CMS to conduct an evaluation of the for-profit demonstration between 2012 and 2013, all sites of which were located in Pennsylvania. The focus of the evaluation was on access and quality. CMS made the October 2013 report of the evaluation results available to Congress in May 2015 and to the public in October 2015 (CMS, 2015a; Jones et al., 2013). The next two sections describe the design and results of Mathematica’s evaluation.

PACE study design and sampling issues

The Mathematica evaluation took place entirely in Pennsylvania because it was the only state where for-profit organizations operated PACE programs (see Jones et al., 2013 for a full description of the study design). The evaluation had two design flaws. The first is that it did not use a randomized controlled design—a limitation, but not an uncommon one in PACE evaluations. The second is that it did not adhere to all of the requirements for the evaluation set out by the BBA.

The study’s sample was assembled in several stages using a matching process. First, researchers selected all four of the for-profit programs operating in Pennsylvania at the time—Senior LIFE Washington, Senior LIFE York, Senior LIFE Altoona, and Senior LIFE Johnstown—and compared them with the state’s 11 nonprofit PACE programs in regard to length of time in operation, geographic location (urban/rural), and the population characteristics of the regions where the for-profit PACE organizations were operating (socioeconomic and demographic information). The goal was to match the nonprofit programs chosen for the evaluation as closely as possible to the four for-profit programs participating in the study. One of the for-profit programs had been in operation for 59 months, while the other three had been operating for 17 months. Therefore, the researchers decided to limit the nonprofit programs considered to those operating between 17 and 59 months. Three of the six nonprofits—LIFE Lutheran, LIFE Geisinger, and LIFE Beaver—identified on the basis of length of operation also matched the for-profit sites on the basis of the geographic location and population characteristics criteria employed for purposes of site selection. These three sites were thus included in the evaluation.

Once Jones and colleagues (2013) identified the participating PACE providers, the next step involved selecting a sample of enrollees. First, the sample was limited to those who were aged 65 years or older, despite that those aged 55 years and older are eligible for PACE. The rationale for restricting study enrollment in this way was that enrollees aged 65 years and older made up 90% of the PACE enrollees in Pennsylvania. Excluding younger enrollees, however, is a potential study limitation. It is not clear, for example, whether the evaluators tested for statistically significant differences between 55- to 64-year-old enrollees and enrollees aged 65 years and older. To the extent that important differences exist, findings may not be generalizable to the full population of PACE enrollees or to other sites where the proportion of younger enrollees is higher than in Pennsylvania.

The second step in the sample selection process was to select those who were enrolled in for-profit PACE for at least 6 months and then stratify them into three groups: (a) short-term enrollment (7 to 12 months), (b) medium-term enrollment (13 to 36 months), and (c) long-term enrollment (37 to 59 months). These three groups were, in turn, used to create numerical distributions of the nonprofit programs’ enrollees for purposes of comparing them to the for-profits programs’ enrollees.

It is at this point that the researchers determined that among the other nonprofit programs, Life St. Mary would also be selected for the evaluation. The justification for also including Life St. Mary was that it was the only plan among the six nonprofit plans with a comparable length of service to the for-profit sites that also matched the three enrollment strata in the for-profit programs. The decision to include Life St. Mary is potentially problematic because this site did not conform to two of the three initial program-selection criteria: geographic location and population characteristics.

In sum, all 4 for-profit programs and 4 of the 11 nonprofits programs were included in the evaluation. This process yielded 407 for-profit enrollees and 406 not-for-profit enrollees aged 65 years and older. Of these, 333 of the for-profit enrollees responded to the telephone interview used to collect data from study participants compared with 326 of the not-for-profit enrollees.

The design and sampling approach introduces several limitations. The authors acknowledge that “… because all of the [for-profit] demonstrations were owned by the same company and operated in the same state (Pennsylvania) their common ownership or idiosyncratic characteristics of the areas in which they are located could limit the degree to which these programs are representative of future for-profit PACE programs” (Jones et al., 2013, p. xi). In other words, the findings should be interpreted with caution because they might not be generalizable to other contexts in which PACE operates. Moreover, as noted, one of the four nonprofit programs was included in the study even though it did not conform to the geographic location and population characteristic criteria employed for purposes of site selection.

Another limitation is that matching may exaggerate the effect of the treatment (i.e., being in a for-profit PACE) on outcomes. Matching in general can be problematic because unlike an experimental design, it is difficult to gauge hidden bias and there is a possibility of obtaining estimates that are extremely overestimated or underestimated (see Arceneaux, 2010).

Still another limitation is that the BBA required a minimum of 800 enrollees (or enough to be statistically significant as determined by the secretary) before an evaluation of the demonstration could be conducted. Mathematica determined that 585 people were enrolled in the four for-profit PACE sites and that 2,787 people were enrolled in the 11 nonprofit PACE sites. Although more than 800 lives were covered, only 326 enrollees from the four participating nonprofit organizations and 333 enrollees from the four participating for-profit organizations were included in the study.

Finally, only one of the participating for-profit programs met the federal requirement to operate for 3 years before a study could be conducted, yet all four for-profit programs were included in the analyses. One of the participating nonprofit programs also did not meet this requirement.

Findings of the Mathematica for-profit demonstration

The for-profit PACE evaluation compared the performance of the for-profit and nonprofit sites with respect to heath and demographic characteristics, participants’ frailty, access to and quality of care, satisfaction with care, and cost-effectiveness.

Health and demographic characteristics prior to enrollment

Jones et al. (2013) used administrative data to compare health characteristics and demographics of for-profit and nonprofit PACE participants both prior to and after their enrollment in a PACE program. The results indicate that before enrollment, those in for-profit PACE were slightly less likely to be dually eligible for Medicare and Medicaid and more likely to be enrolled in Medicare Advantage—optional plans managed by Medicare-certified private companies—than nonprofit enrollees. They also indicate that a higher percentage of non-Whites (excluding Hispanics) were enrolled in for-profit than nonprofit PACE. No statistically significant differences were identified with regard to chronic conditions.

Characteristics of for-profit and nonprofit enrollees were also compared after enrollment. Results indicate that for-profit enrollees were slightly less likely to report high rates of chronic conditions but slightly more likely to report mental health problems than nonprofit enrollees. Compared to the nonprofit sites, the for-profit PACE sites also had a higher percentage of those who were dually eligible for Medicare and Medicaid and a lower percentage of those who were eligible for Medicare due to disability or end-stage renal disease.


Findings regarding profit status and frailty were mixed. Participants were asked whether they needed assistance in each of six ADLs. A lower percentage of for-profit enrollees than nonprofit enrollees required help with getting around indoors, bathing, toileting, and transferring in and out of bed, indicating a less frail population. Help with eating or getting dressed did not significantly differ by profit status. When the researchers examined the percentage of those who both needed and actually received help in each setting, they found no significant differences except that a higher percentage of those in for-profits than nonprofits needed and received help eating. Of interest, when the evaluators looked at who received help and who still had unmet needs, they found a higher percentage of for-profit than nonprofit enrollees who received help with getting around and toileting to also report having unmet needs relating to these two ADLs. This finding suggests that the help for-profit enrollees receive for these two ADLs is insufficient to meet needs relative to the help provided to nonprofit enrollees.

Access and quality

The study examined several factors relating to access to and quality of care. For-profit enrollees were more likely than nonprofit enrollees to report having fallen in the past 6 months, having been injured by a fall in the past 6 months, and that it takes a great deal of energy to get services. Compared to nonprofit PACE enrollment, being in a for-profit PACE was also associated with a decrease in the probability of having had a flu shot and pneumonia vaccination and having eyesight tested regularly. Taken together, these findings suggest that for-profit PACE participants experience greater barriers to access to care and face a lower quality of care than nonprofit PACE participants.

Satisfaction with care

The Mathematica study also compared satisfaction with care by profit status. For-profit enrollees were less likely to report being satisfied with overall care, with information from doctors, with information on medications, and with coordination of care than nonprofit enrollees. For-profit enrollees were also less likely to report being treated with respect and less likely to report being able to obtain appointments with specialists than their nonprofit counterparts. With regard to therapy, for-profit enrollees were more likely than nonprofit enrollees to report being satisfied with therapy. Taken together, the findings suggest that for-profit enrollees are less likely to be satisfied with care when compared to beneficiaries enrolled in nonprofit sites.


Jones et al. (2013) did not explicitly compare the cost-effectiveness (or even the cost to Medicare and Medicaid) of for-profit and nonprofit PACE. Instead, they simply state that, “Prospective payment for for-profit PACE organizations is calculated using the same methodology as not-for-profit PACE organizations. Therefore, expenditures were equal … after controlling for beneficiary risk score, organization frailty score, and county rates …” While it is true that CMS and the administering state agency make prospective payments using a formula, doing so does not guarantee that the money will be spent as efficiently in for-profit and nonprofit PACE programs or that costs to Medicare and Medicaid will be the same. Until these possibilities have been explored, it is premature to conclude that for-profit PACE and nonprofit PACE are equally cost-effective.

Summary of findings

The evaluation of the for-profit PACE demonstration revealed mixed findings regarding health and demographics prior to and after enrollment and mixed findings regarding frailty. The evaluation, however, showed that nonprofit PACE performed better than for-profit PACE on all measures of access and quality and on all but one measure of satisfaction with care. Moreover, the study did not include a formal analysis of cost-effectiveness or cost equivalency by profit status. This is an important point not only because it was a requirement under the BBA demonstration rules to examine Medicare and Medicaid expenditures but also because the for-profit industry often claims cost-effectiveness as a justification for the takeover of nonprofit enterprises. Together, the study’s findings and limitations (particularly with respect to setting, sample size, and duration of operation) suggest that it is premature to conclude that for-profit PACE and nonprofit PACE are comparable. The study did not demonstrate equivalent care and cost by profit status and, in fact, nonprofit PACE often outperformed for-profit PACE where differences were observed.

Lessons from the rise of for-profit hospice

The shift in profit status that has occurred over the last two decades in hospice is instructive in understanding recent changes in PACE. Hospice and PACE both serve vulnerable populations who are often at or near the end of life. Moreover, both PACE and hospice providers rely heavily on fixed payments—hospice providers via the fixed per diem Medicare benefit and PACE via Medicare, Medicaid, private pay (or a mix) capitated payment—which provides financial incentives to control costs. Like PACE, hospice care is associated with reductions in hospitalizations and the likelihood of dying in a hospital and is rated highly by patients and their families (Kelley, Deb, Du, Carlson, & Morrison, 2013). Given the aforementioned similarities, understanding the shift to for-profit hospice could shed light on the implications of a similar such shift in PACE.

The number of for-profit hospice providers has tripled in the last 15 years. Until recently, most hospices were run by nonprofit or community groups with extensive volunteer participation. Today, more than 60% of hospices are for-profit (National Hospice and Palliative Care Organization, 2013). Medicare expenditures for hospice have increased from $10 billion in 2007 to almost $15 billion in 2012, primarily due to a growing number of enrollees and an increase in the number of Medicare-participating hospices (CMS, 2014; Iglehart, 2009). Medicare beneficiaries qualify for hospice if the attending physician declares, along with the hospice physician, that they have 6 months or less to live. An individual or their representative files an election statement that waives Medicare Part A for treatment of their terminal illness and identifies the hospice that will provide care in accordance with the patient’s treatment plan.

Much of the research on profit status and hospice has focused on examining enrollees’ length of stay, the number and intensity of services provided, and staffing. Per diem reimbursement provides hospice providers with incentives to keep costs below the amount paid by Medicare in order to maximize “profits” and to maximize the number of low-care days while minimizing the number of high-cost days. Hospice care is especially costly during initial enrollment and near the end of life. However, care is much less expensive in between these two times, with the relationship between cost and length of enrollment taking on a U-shaped curve (Carlson, Barry, Cherlin, McCorkle, & Bradley, 2012). This middle period (low-care days) can provide hospices with incentives to increase the length of stay and to select patients with diagnoses that provide a longer stay (e.g., those without an aggressive cancer diagnosis). Other strategies hospices can use to minimize costs in light of fixed reimbursement include restricting services, reducing the number of visits per patient from staff, and using less qualified staff when more qualified staff might be better suited to provide care.

The Office of the Inspector General (OIG) of the U.S. DHHS has commissioned several reports on hospice care and has found some differences by profit status. In 2009 the OIG found that more than 80% of hospices in nursing facilities did not comply with Medicare’s coverage requirements—certification that a patient is terminally ill, compliance with election statements, and plans of care—which cost Medicare about $1.8 billion (OIG, 2009). However, nonprofit hospices were less likely to comply with Medicare’s coverage requirements in nursing homes than for-profit hospices (89% of claims vs. 74% of claims).

Another OIG report (2015) found that hospices serving patients in assisted living facilities (ALFs) compared to hospices serving patients in other long-term care settings provided care for a longer period of time and received much higher Medicare payments. The report also showed that over a period of 5 years Medicare payments to hospices serving ALF residents more than doubled to a total of $2.1 billion in 2012. Again, significant differences by profit status were evident. For-profit hospices were more likely than nonprofit hospices to serve beneficiaries in ALFs and to generate more of their revenue from ALFs. The median amount paid by Medicare to for-profit hospices serving ALFs was more than $18,000 per beneficiary compared to a median of about $14,000 per beneficiary for nonprofit hospices. The length of stay for beneficiaries being served by for-profit hospices in ALFs was 4 weeks longer than similar beneficiaries served by nonprofit hospices.

The most recent OIG report (2016) examined Medicare expenditures for general inpatient care. General inpatient care is intended for hospice beneficiaries who need short-term intensive pain or chronic symptom management that cannot be provided in nonintensive settings. General inpatient care is the second most expensive level of hospice care, with average costs of about $670 per day as compared to $880 and $155, respectively, for continuous home care and inpatient respite care (OIG, 2016). According to the report, hospices cost Medicare $268 billion in 2012 for inappropriately billed general inpatient care. At 40% and 27%, respectively, inappropriate billing for general inpatient care was much more likely among for-profit than nonprofit or government-owned hospices.

Other studies have found that there are some differences in length of stay, patient enrollment, Medicare billing, staffing, and services provided with nonprofit hospices, for the most part outperforming for-profit hospices, and for-profit hospices are more likely than nonprofit hospices to adopt cost-reducing and profit-maximizing strategies. Carlson and colleagues (2004) examined 422 hospice agencies from a nationally representative survey of hospices and found that controlling for factors such as gender, disability, and location, for-profit hospices provided a narrower range of services than nonprofit hospices, suggesting lower quality of care. Canavan et al.’s (2013) study of hospice use in nursing homes found that for-profit hospices have a lower staff to patient ratio, also possibly indicating lower quality. Lorenz et al.’s (2002) study of 176 hospices in California found that there was no difference in patients’ length of stay by profit status. Compared to nonprofit hospices, however, for-profit hospices reported a higher percentage of non-cancer discharges (those discharged for all causes including death) (46% vs. 28%), more referrals from long-term care facilities (19% vs. 4%), more patients who use government payments (88% vs. 81%), and fewer skilled nursing visits as a percentage of all nursing visits (49% vs. 60%), possibly indicating that for-profit hospices are intentionally attempting to control their patient mix and costs related to staffing.

O’Neill et al.’s (2008) study of 185 hospices in California revealed that length of stay was longer in for-profit hospices compared to nonprofit hospices and that for-profit patients had fewer visits from registered nurses. A study of California’s Medicaid-certified hospices revealed that for-profit hospices were more likely than nonprofit hospices to enroll those with a cancer diagnosis (Gandhi, 2012). This finding runs contrary to profit maximization strategies that include avoiding enrolling costly cancer patients. Other study findings, however, suggest that for-profits use profit maximization strategies. For example, the study found that for-profit hospices were more likely to enroll individuals who were referred by long-term care facilities (i.e., enrolling those who are predicted to have longer stays with more low-care days). The researchers also found that for-profit hospices in general have enrollees with longer lengths of stay, employ less skilled staff, and offer fewer social services.

Aldridge and colleagues (2015) analyzed Medicare hospice claims from 2000 and 2010 and found that for-profit hospices were more likely than nonprofit hospices to enroll patients for more than 6 months and less likely to enroll patients for shorter lengths of stay (e.g., 1 week or less). A study of 93 home care agencies and 29 hospices in North Carolina found that direct care staff in for-profit agencies and hospices had higher turnover rates (34% points higher) than nonprofits (Dill & Cagle, 2010). Finally, using data from the National Home and Hospice Care Survey, Wachterman and colleagues (2011) found that for-profit hospices compared to nonprofit hospices enrolled fewer individuals with a cancer diagnosis and a higher proportion of those living with dementia and that enrollees in for-profit hospice had longer lengths of stay.

In sum, the research literature on hospice and profit status shows how for-profit hospices are more likely than nonprofit hospices to adopt strategies to earn and maximize a profit under Medicare’s fixed per diem payment structure. Several studies showed that for-profit hospice, compared to nonprofit, was more likely to provide services in settings where Medicare reimbursements are higher, to maximize low-care days by selecting enrollees with a longer length of stay (e.g., those with dementia and those with a non-cancer diagnosis), to cut back on the number and quality of staff, and to provide fewer services. Together these findings suggest the need for concern and vigilance as a similar shift from nonprofit to for-profit operators takes place with PACE.


At present, PACE is undergoing significant changes. In June 2015, CMS announced that it would no longer distinguish between for-profit and nonprofit PACE operators. This decision was based, in part, on the Jones et al. study described above (CMS, 2015b; National PACE Association, 2016a). Already, Colorado’s PACE provider, InnovAge, has converted from nonprofit to for-profit status (Wallace, 2015) and several others are expected to follow suit. In late 2015, President Obama signed into law the PACE Innovation Act, signaling another significant development. The act allows CMS to develop pilot projects that serve different populations than the PACE program’s original target. Those who are younger than 55, those with multiple and chronic conditions, and elders who do not yet qualify for a nursing home level of care will be served by the demonstration programs established under this act (National PACE Association, 2016b). Similar to profit-maximizing strategies in for-profit hospice, for-profit PACE providers could soon enroll those who are less costly—younger, healthier individuals who are not in need of a nursing home level of care—in addition to or even instead of older, frailer individuals. Recently, CMS proposed a rule change that would permit PACE providers to use nonphysician medical staff in place of primary care physicians (CMS, 2016). Like for-profit hospice, for-profit PACE will likely take advantage of rules such as this to substitute less costly staff (nurses and/or direct care workers) for more costly staff (physicians). Together, the PACE Innovation Act, pilot programs, and CMS rule changes could fundamentally change PACE and its original focus on care for the frail elderly.

PACE has been operated by nonprofit organizations for more than four decades. During that time, they have provided cost-effective, high-quality care that has allowed elders to remain in the community. Results from the only study that examined the relationships among access, frailty, quality, and profit status do not provide a sufficient rationale for turning a well-functioning nonprofit operation into a for-profit one—especially in light of the findings that for-profit PACE performed worse on many measures of satisfaction and access to and quality of care and in light of lessons learned from a similar shift that has occurred in hospice. Still, the future of PACE looks to be heading in the same direction as hospice—increasingly for-profit and incentivized to constrain costs at the expense of access and quality. At a minimum, more research is needed, and policy makers should proceed cautiously before making the decision to allow a for-profit conversion of PACE.