Nursing Home Referral Within the Veterans Health Administration: Practice Variation by Payment Source and Facility Type

Abstract

Veterans enrolled within the Veterans Health Administration (VHA) of the U.S. Department of Veterans Affairs (VA) may receive nursing home (NH) care in VHA-operated Community Living Centers (CLCs), State Veterans Homes (SVHs), or community NHs, which may or may not be under contract with the VHA. This study examined VHA staff perceptions of how Veterans’ eligibility for VA and other payment impacts NH referrals within VA Medical Centers (VAMCs). Thirty-five semistructured interviews were performed with discharge planning and contracting staff from 12 VAMCs from around the country. VA staff highlights the preeminent role that VA priority status played in determining placement in VA-paid NH care. VHA staff reported that Veterans’ placement in a CLC, community NH, or SVH was contingent, in part, on potential payment source (VA, Medicare, Medicaid, and other) and anticipated length of stay. They also reported that variation in Veteran referral to VA-paid NH care across VAMCs derived, in part, from differences in local and regional policies and markets. Implications for NH referral within the VHA are drawn.

Eligible Veterans enrolled within the Veterans Health Administration (VHA) can receive U.S. Department of Veterans Affairs (VA)-paid nursing home (NH) care in VA-owned and VA-operated Community Living Centers (CLCs), community NHs that contract with the VA, and state government–owned and state government–operated State Veterans Homes (SVHs), paid for, in part, and overseen by the VA. Veterans may also pay for community NH care privately or receive care paid for by Medicare or Medicaid. In 2012, the VA devoted US$4.9 billion or 9% of its total health-care budget to NH care (U.S. General Accountability Office [GAO], 2013). That year, the average daily census for VA-paid NH care was 36,250 residents; more than half in SVHs (53%), followed by CLCs (28%) and contract NHs (CNHs; 19%). Long-stay residents accounted for 87.6% of the average daily census for VA-paid NH care in 2012 (U.S. GAO, 2013). Virtually all SVH residents were long stay (97%); a higher proportion of CLC residents (24%) and CNH residents (20%) were short stay.

Although insurance status is routinely assessed by hospital discharge team members in the broader health sector (Nazir, Little, & Arling, 2014), no study, to our knowledge, has examined the role of payment source in NH referral within the VHA. The few studies that have examined the role of financial considerations in NH search and selection asked residents, family members, and/or health personnel about the relative ranking or importance of costs (Reinardy & Kane, 1999; Shugarman & Brown, 2006), price (Castle, 2003; Collier & Harrington, 2005), or insurance (Konetzka & Perraillon, 2016; MCauley, Travis, & Safewright, 1997) in the decision-making process. Only a handful of these studies focused on investigating the views of professional intermediaries such as discharge planners and case managers with the greatest breadth of experience helping to identify and place patients in NHs (Collier & Harrington, 2005; Shugarman & Brown, 2006). Generally, evidence indicates that financial issues rank alongside geographic location, care needs, bed availability, and quality as among the attributes patients, families, and health-care personnel consider when searching for and selecting an NH.

Understanding the relationship between insurance status and NH use is important for ensuring consistent access to high-quality care among VHA-eligible Veterans nationally (Miller & Intrator, 2012). Few previous investigations, however, have examined extended care referrals within the VHA (Guihan, Hedrick, Miller, & Reder, 2011; Reder, Hedrick, Guihan, & Miller, 2009); none, specifically, on NH search and selection, let alone how payment eligibility influences those decisions. Consequently, the primary aim of this study was to examine VHA staff perceptions about how Veterans’ eligibility for multiple payment sources—VA, Medicare, Medicaid, and other—impacted the general type of NH—CLC, CNH, SVH, or other—to which VHA enrollees were admitted. We begin with a primer on VA NH care options, use, and payment, followed by a review of our methods, results, discussion, and conclusions.

Veterans, NHs, and Payment Sources

The VHA Office of Geriatrics and Extended Care Services (i.e., VA Central Office) establishes national policies regarding the provision of NH care. However, administration is decentralized to the VHA’s current 21 Veterans Integrated Service Delivery Networks (VISNs) and 152 VA Medical Centers (VAMCs). VHA policy permits any honorably discharged Veteran to sign up for VHA services and potentially receive care from the VHA. This includes Veterans who may be eligible for compensation by virtue of injuries received while serving in the military; in which case, they are required to file for disability and obtain a medical evaluation after which rating specialists determine the percentage of service-connected disability using standardized tables (VHA, 2013).

The Veterans Millennium Health Care and Benefits Act of 1999 (Public Law 106–117) established a hierarchy of priority status, with the VHA required to provide NH care to “mandatory Veterans” (Priority P1As), consisting of those seeking NH care for (1) a service-connected disability, (2) any reason but with a combined service-connected disability rating of 70% or more, and (3) any reason but with a combined service-connected disability rating of at least 60%, and is deemed unemployable or permanently and totally disabled (VHA, May 17, 2004). Other “nonmandatory” or “optional” Veterans may receive VHA-paid NH care at the discretion of individual VAMCs if resources are available and if all members of the aforementioned mandatory groups in need of such care are served (VHA, April 3, 2015). VA-paid NH is provided in CLCs, CNHs, and SVHs.

VA-Paid NH Options

In 2008, VA NHs were renamed CLCs as part of a transformation in culture away from the medical model toward more home-like environments providing person-centered care (Hojlo, 2010). By 2012, there were 135 VA CLCs (U.S. GAO, 2013). CLCs tend to be located on campus, as single units within VA hospitals or as freestanding buildings supported by an array of clinical specialties available at the VAMC. Local VAMCs cover the full costs of CLC care provided to mandatory Veterans, while optional Veterans may be required to pay a copayment depending on income and other factors. VA Central Office has issued several directives aimed at shifting the emphasis of CLC admissions from long-term to short-term care (VHA, 2005, 2012).

Since 1965, the VA has contracted with CNHs approved by the VHA, including approximately 2,500 facilities in 2012 (U.S. GAO, 2013). VAMCs contract with community NHs when they determine that a need exists for additional NH options. Participating facilities must be state licensed and, with few exceptions, Medicare and/or Medicaid certified. Local VAMCs pay the full costs of care for mandatory Veterans residing in CNHs based on a negotiated rate (Intrator, 2013). Optional Veterans may be subject to copayments and a 6-month limit on the amount of VA-paid care provided.

SVHs were established after the Civil War; today, there are 162 facilities (Senior Veterans Service Alliance, 2017). SVHs must be certified by the VA, though states assume primary responsibility for regulating them. Unlike CLCs and CNHs, the VA’s contribution to SVHs is paid from a national budget source, rather than by local VAMCs. On average, the VA pays SVHs a per diem equivalent to one third the total costs of care (U.S. GAO, 2013). The VA pays the full costs for mandatory Veterans. The costs for optional Veterans are divided among the state, Veteran, and VA (U.S. GAO, 2013).

VHA Enrollees’ Payment Sources Beyond the VHA

Eligibility for multiple payment sources is high among Veterans enrolled in the VHA. In 2015, just one fifth of VHA enrollees only had VA coverage (Gasper, Liu, Kim, & May, 2015). A little more than half of VHA enrollees (51.3%) had Medicare, 28.4% private insurance, 18.3% TRICARE, and 6.8% Medicaid. The proportion of VHA enrollees qualifying for Medicare and Medicaid increased with age; younger enrollees were more likely to have private insurance and TRICARE coverage.

Outside the VHA short-stay NH residents typically pay for care privately or through Medicare, which includes a 100-day coverage maximum per episode of care after a 3 day qualifying hospital stay, full coverage for the first 20 days, and 80% coverage for days 21–100. Long-stay NH residents, however, typically enter under Medicaid or pay for care privately or through Medicare before qualifying for Medicaid after spending down their personal resources and/or exhausting their Medicare benefit. Long-stay residents with private supplemental insurance (Medigap) can forestall spend down through coverage of Medicare coinsurance and deductibles. Long-stay residents eligible for both Medicare and Medicaid receive Medicaid coverage of part or all of their Medicare cost sharing.

Method

Data for this study derived from semistructured interviews. Participants were selected through a combination of purposive and snowball sampling with selection first based on our own knowledge about the VHA but later on information provided by interviewees regarding specific actors who should be interviewed about the VHA’s extended care referral and contracting processes. Twelve VAMCs were chosen for study representing different geographic regions—northeast (2 sites), south (2), midwest (4), west (4), catchment areas—rural (4), urban (8), and available NH options—CLCs (0–2), CNHs (5–48), and SVHs (0–5; Department of VA, 2009). We selected most sites beforehand to ensure variation on these characteristics. However, we selected some sites based on information provided by our interviewees. Interviews were conducted until saturation or the point at which no new information was obtained and enough data had been collected to develop the categories required to describe the processes studied (Glaser & Strauss, 1967).

Interviewees were selected based on which VHA staff members were most knowledgeable about each VAMC’s extended care referral and contracting processes. Thus, at each VAMC, interviewees included those responsible for provider contracting and discharge planning. In a few instances, we interviewed knowledgeable participants in other roles (e.g., directors of geriatric services, extended care, and rehabilitation) when one of the main types of participants was unavailable.

In all, 35 in-depth interviews were conducted with 36 individuals between May 18, 2012, and December 6, 2012 (one interview included two participants). Two interviews took place in person, the remainder over the telephone. Three interviews were conducted at each site, except one site where two were conducted. Interviews lasted about 60 to 90 min each. Participants included 20’social workers, 12 nurses, and 4 geriatrics physician leaders. The first, second, and third authors—E. Miller, S. Gidmark, and E. Gadbois—conducted all interviews with at least two of the three present at each interview. Conversations were recorded (with each interviewee’s permission) and transcribed.

The interview protocol was drafted with an input from both VHA practitioners and researchers and personnel from VA Central Office. The protocol was then fielded at one VAMC before being refined for use in the study. Through the protocol, we sought to understand how NH referral works within the VHA, including the role that payment source, expected length of stay, and supply plays in the decision to enter a CLC, SVH, or CNH. We also sought to identify and explain similarities and differences that exist in this area across VAMCs. The role that Central Office regulation and regional and local rules, practices, resources, and capacities play in NH referral was examined as well. We asked interviewees “at your VAMCs, what options are available for you to refer patients to CLCs, SVHs, and/or CNHs?” and “do rules or other factors limit the specific facilities to which you can refer a Veteran.” We also asked interviewees to “describe the process for initial NH referrals” and to walk us through both “a typical referral” and “a difficult referral.” This question was followed by “what, in your opinion, are the most important patient characteristics that influence the NH referral process.” Use of follow-up questions enabled us to clarify and opportunistically explore responses. Probes were used when certain topics were not raised naturally, but it was felt going into the study might be important (e.g., “patient priority status,” “patient insurance status,” and “predicted length of stay”). The same protocol was used for respondents from different disciplines.

The first three authors used an open-coding process to identify recurring patterns and themes in responses in the transcripts (Miles & Huberman, 1994). This process was deductive to the extent that it was guided, in part, by a priori codes developed before the coding began. However, it was inductive to the extent that new codes were developed and initial codes modified or eliminated during the coding process. Initial codes derived from the topics covered in the interview protocol and the authors’ knowledge/recollection of the interviews conducted. They were refined through a process whereby each author read through 12 transcripts—one from each study site, and independently generated suggestions for potential coding categories. The authors then met to discuss their findings and to reach consensus on the updated coding scheme. The final coding tree consisted of progressively narrower branches under general topics such as CLCs, SVHs, and CNHs. At the broadest level, for example, codes under CNHs were divided into “facility characteristics” and “patient characteristics,” with the latter including somewhat narrower subcategories such as “payment scenarios” under which even more specific coding categories were assigned (e.g., “VA,” “Medicare,” and/or “Medicaid”). Two authors (S. Gidmark? and E. Gadbois) used the coding tree to separately code/recode all transcripts based on the common set of codes produced.

The final coding tree was used to identify, collect, and organize segments of text from across the coded transcripts into reports addressing specific topic areas (e.g., “CNHs/patient characteristics/payment scenarios/Medicaid” and “SVHs/patient characteristics/length of stay/long term”). One investigator (E. Miller) reviewed the coded segments within each report, grouped them into themes/subthemes and identified illustrative quotations (Table 1). Review of the report “CLCs/patient characteristics/payment scenarios/VA and Medicare,” for example, informed the content of “beginning VA payment immediately despite Medicare eligibility” and “beginning VA payment after using Medicare to pay for community NH care” subthemes. Themes/subthemes were reviewed and finalized by all three investigators. NVivo (Version 10), a widely used qualitative data software program, was used to manage the data. Institutional review board approval for this research was obtained.

Table 1.

Major Themes Arising From Key Informant Interviews With Illustrative Quotes.

Theme 1: Payment source as a consideration in the nursing home (NH) referral process

1.a. NH referral informed by eligibility for VA payment “The weighting of service connection—that is of course usually at the top.” (rural, West) “Now we only place what we call Millennium Bill eligible Veterans, so we only place service-connected Veterans.” (largely rural, West)

1.b. NH referral is informed by eligibility for non-VA payment sources “The first thing we do look at is their insurance. Do they have Medicare … Are eligible for Medicaid?” (urban, East)

Theme 2: Community Living Centers (CLCs)

2.a. Mandatory Veterans given priority for CLC beds but optional Veterans also served “The Mill-Bill/service-connected guys get priority over our beds.” (urban, South) “If they’re not high enough service-connected … we try to accommodate them … at our CLC if we have a bed.” (urban, South) “In better budget years sometimes they …provide …shortterm NH care for non-service-connected Veterans [in] need.” (largely rural, West) “Our CLC … serves as the safety net… for difficult to place Veterans …. They’re the folks that don’t have financial resources.” (largely rural, West)

2.b. Long-term placement in CLCs typically for mandatory Veterans, with exceptions “[The CLC has] four beds that are designated …for long-term care, but they have to be Millennium Billed.” (largely rural, West) “Level 3 sex offender[s] that by state law the community NHs can’t admit…. They wind up with us [at the CLC].” (urban, East) “We… get… a geriatric psych patient… that are very, very difficult… We don’t always have the right place to send [them].” (urban, Midwest) “We have a huge amount of homeless Veterans …. Because they have no place to go [we try] to get them [into the CLC].” (nurse, rural, West)

2.c. Short-term placements in CLCs represented a mix of mandatory and optional Veterans “[The] preference was always to bring [mandatory Veterans] in-house if they needed rehab. Part of it is because of cost.” (largely rural, West) “We have doctors on-site, we have [a] lab on-site… which makes it easier to take care of sub-acute patients.” (urban, East) “I would say probably the majority of those short-stay, skilled, and rehab patients in the CLC are probably not Mill Bill eligible.” (urban, East) “If he starts to get sick here [at the CLC] [it is easier] to ambulance him off to the [local VA hospital] rather than have him go 911 to a community hospital and then to try to retrieve him from there…. It saves some time and money.” (urban, East) “If [they] require [six weeks] we would use the CLC rather than Medicare knowing they would have a copayment under Medicare.” (urban, West)

2.d. Shifting CLCs away from long-term to short-term admissions “If they’re going to be a short stay, they’ll take a spin through our CLC for that short stay.” (highly rural, Midwest) “If they are long term care, they wouldn’t go into the CLC because they do not provide long term care there.” (urban, East) “They try to use the CLC for short term … care… [So] they can get the turnover and free up the beds again.” (largely rural, West) “I can either treat one patient for 365 days or I might be able to treat a half a dozen patients for 30 to 60 days each.” (urban, East) “Most of our patients have been here a long, long time, so … we can[‘t] just automatically switch to short stay.” (urban, Midwest)

Theme 3: Contract Nursing Homes (CNHs)

3.a. CNHs served mandatory Veterans with long-term care needs “Almost everybody who receives … paid VA care in a contract home are the Mill Bill eligible” (urban, East) “Most Veterans that require long-term care are routed through our Contracted NH Program rather than … the CLC.” (urban, Midwest) “[Optional Veterans] would go out and they would utilize the community NHs. Through Medicare for short term rehab or skilled care, or Medicaid. Or private pay, or private insurance, long-term care, pensioner’s plan.” (urban, South)

3.b. Veterans selected from among available payment sources for community NH care “If it’s a Veteran who is not going to be on VA contract we would just kind of give them the entire array of options.” (largely rural, West) “Essentially, [if you’re Mill Bill] you need to pick from the contracted NHs if you want the VA to pay.” (urban, East) “They have a choice, that they can initially start [with] their VA benefit or they can utilize their Medicare benefits for the first 20 days … and [then] switch to the VA.” (highly rural, Midwest) “If they have Medicare and they are Mill Bill eligible, generally those patients will… have the VA pay the contracted NH … for them.” (urban, East) “They have 100% [service-connection] and they have the Medicare and they don’t like the contract NH list, they go to another one.” (urban, East) “A lot… would say yeah lets have Medicare pay for it because I want to make sure other Veterans can access these services.” (largely rural, West)

3.c. Beginning VA payment immediately despite Medicare eligibility “If they go out to a contracted NH the VA will not use their Medicare…. The VA will just pick up the bill from day one.” (urban, Midwest) “We don’t want a service-connected Veteran who has a… benefit that he’s entitled to [paying a co-pay under Medicare].” (urban, Midwest)

“[We prefer they] start at a CNH [rather than elsewhere on Medicare] because while… we can move people that’s difficult to do.” (urban, West)

3.d. Beginning VA payment after using Medicare to pay for community NH care “Almost every VA in the country… use Medicare for the first days that are allowed and then switch to the contract.” (largely rural, West) “They get that first 20 days of Medicare paid for 100%; on day 21, if they don’t have a supplemental insurance to pick up that co-pay then they come off of Medicare and we put them on VA contract.” (urban, West) “Our contract rates are based on Medicaid plus a percentage. Medicare pays quite a bit more.” (urban, South) “We sent patients out for 20 days under Medicare and [then] converted them to VA contract. [CMS said] that was not allowable.” (urban, South)

3.e. VAMCs sometimes provided optional Veterans with short-term care contracts in community NHs “We have a large hospital frequently on bypass …. If we have someone that needs placement even though they don’t meet the criteria we will allow a shortterm 32-day contract when all else fails to do a timely discharge from the hospital setting.” (urban, Midwest) “If a Veteran needs six or eight weeks of an intravenous antibiotic in a skilled nursing facility… and then after [day 20] doesn’t really have the funds to pay the [Medicare] copay…we may… consider issuing a time-limited contract.” (urban, Midwest) “We occasionally have a non-service-connected Veteran with special needs, and the hospital administration will approve some NH care at VA pay…. This person would be out on the streets … if we didn’t help.” (largely rural, West) “Contract exceptions [are] for [when] the only way [individuals] can get discharged from an acute care setting would be to offer say a 30 day contract… for the NH to pull in … Medicaid or some other payment source to cover them.” (urban, West)

3.f. Fiscal considerations often primary when VAMCs determined whether to pay for optional Veterans in CNHs “There are larger facilities … that have the funding available to them that they can make those [non-mandatory] placements.” (rural, Midwest) “We have gotten more restrictive about who we do short-term contracts for… in order to save money.” (urban, South)

Theme 4: State Veterans Homes (SVHs)

4.a. VHA staff rarely referred to Veterans directly to SVHs due to lengthy waitlists “[At the]… largest [SVH]… the waitlist is at least a year long. The other ones are probably 6 to 8 months on average…. Most of the time when these families need NH care it’s become a crisis. So they [can’t] wait.” (largely rural, Midwest) “[Veterans]… sometimes [pay for a] community NH … and get on the waitlist for the Soldiers Home… and then transfer.” (urban, East)

4.b. SVHs served primarily nonmandatory Veterans “For the [SVHs] they don’t have to [be] service-connected they just need to be a Veteran. It’s probably the best deal if they don’t have eligibility to stay [in the CLC].” (largely rural, Midwest) “Out of… let’s say a 140–150 patients at each facility, you might have about 10 patients that we’re paying for 100%.” (urban, South) “I’m urging [non-service connected Veterans], put in the applications for the SVH because the financial burden is so much less onerous on the families - It’s huge!” (urban, Midwest)

4.c. SVHs were a financially attractive referral option for VAMCs serving mandatory Veterans “If I have a Veteran … say… who is 100% service-connected … I’ll say to them … that… either way you’re going… to be paid for by the VA as long as you need NH level of care.” (urban, West) “A mandatory Veteran, if he comes [to the CLC], he fills up a bed. If I send him out on contract, he’s costing me…. But if I send him to an [SVH], Central Office pays for it and so it’s not my money. It’s in my best interest to get as many as possible into the [SVH]…. I’m being completely crass and mercenary here - but that is the best way to do that.” (urban, Midwest)

4.d. SVHs served long-term residents with comparatively lower care needs than CLCs and CNHs “Not really medically complex [or] short-term, skilled nursing, or rehab and discharge. Generally, it’s long-term placement.” (urban, East) “I have not known of a Veteran going there from the hospital. It’s usually permanent placement… Usually they’re on the wait list for about six months before they get in.” (urban, South) “Predominantly you see patients who are more high functioning… choose the [SVH] because they can enter into that… lower level of [assisted living type] care and have it not be as costly for them [as it would be in the community].” (highly rural, Midwest)

4.b. SVHs served primarily nonmandatory Veterans “For the [SVHs] they don’t have to [be] service-connected they just need to be a Veteran. It’s probably the best deal if they don’t have eligibility to stay [in the CLC].” (largely rural, Midwest) “Out of… let’s say a 140–150 patients at each facility, you might have about 10 patients that we’re paying for 100%.” (urban, South) “I’m urging [non-service connected Veterans], put in the applications for the SVH because the financial burden is so much less onerous on the families - It’s huge!” (urban, Midwest)

4.c. SVHs were a financially attractive referral option for VAMCs serving mandatory Veterans “If I have a Veteran … say… who is 100% service-connected … I’ll say to them … that… either way you’re going… to be paid for by the VA as long as you need NH level of care.” (urban, West) “A mandatory Veteran, if he comes [to the CLC], he fills up a bed. If I send him out on contract, he’s costing me…. But if I send him to an [SVH], Central Office pays for it and so it’s not my money. It’s in my best interest to get as many as possible into the [SVH]…. I’m being completely crass and mercenary here - but that is the best way to do that.” (urban, Midwest)

4.d. SVHs served long-term residents with comparatively lower care needs than CLCs and CNHs “Not really medically complex [or] short-term, skilled nursing, or rehab and discharge. Generally, it’s long-term placement.” (urban, East) “I have not known of a Veteran going there from the hospital. It’s usually permanent placement… Usually they’re on the wait list for about six months before they get in.” (urban, South) “Predominantly you see patients who are more high functioning… choose the [SVH] because they can enter into that… lower level of [assisted living type] care and have it not be as costly for them [as it would be in the community].” (highly rural, Midwest)

4.b. SVHs served primarily nonmandatory Veterans “For the [SVHs] they don’t have to [be] service-connected they just need to be a Veteran. It’s probably the best deal if they don’t have eligibility to stay [in the CLC].” (largely rural, Midwest) “Out of… let’s say a 140–150 patients at each facility, you might have about 10 patients that we’re paying for 100%.” (urban, South) “I’m urging [non-service connected Veterans], put in the applications for the SVH because the financial burden is so much less onerous on the families - It’s huge!” (urban, Midwest)

4.c. SVHs were a financially attractive referral option for VAMCs serving mandatory Veterans “If I have a Veteran … say… who is 100% service-connected … I’ll say to them … that… either way you’re going… to be paid for by the VA as long as you need NH level of care.” (urban, West) “A mandatory Veteran, if he comes [to the CLC], he fills up a bed. If I send him out on contract, he’s costing me…. But if I send him to an [SVH], Central Office pays for it and so it’s not my money. It’s in my best interest to get as many as possible into the [SVH]…. I’m being completely crass and mercenary here - but that is the best way to do that.” (urban, Midwest)

4.d. SVHs served long-term residents with comparatively lower care needs than CLCs and CNHs “Not really medically complex [or] short-term, skilled nursing, or rehab and discharge. Generally, it’s long-term placement.” (urban, East) “I have not known of a Veteran going there from the hospital. It’s usually permanent placement… Usually they’re on the wait list for about six months before they get in.” (urban, South) “Predominantly you see patients who are more high functioning… choose the [SVH] because they can enter into that… lower level of [assisted living type] care and have it not be as costly for them [as it would be in the community].” (highly rural, Midwest)

Note. VA = Veterans Affairs; VAMCs = VA Medical Centers; VHA = Veterans Health Administration.

Results

This section begins by placing eligibility for VA and other payment sources in the context of NH referral, followed by a review of factors informing referral to CLCs, CNHs, and SVHs, specifically.

Theme 1: Payment Source as a Consideration in the NH Referral Process

NH referral informed by eligibility for VA payment (1.a).

Veterans need to be clinically and financially eligible in order to receive NH care paid for by the VA. VHA staff reported that the most important financial characteristic discharge planners looked at in informing the decision to refer to a CLC, CNH, or SVH was Veterans’ eligibility for VA-paid care. Respondents variously referred to this eligibility as Veterans’ “service-connected” status, “mandatory status,” “Millennium Bill (Mill Bill) eligibility,” or “contract eligibility.”

NH referral informed by eligibility for non-VA payment sources (1.b).

VHA staff agreed that assessing Veterans’ eligibility for payment sources beyond the VA was critical to determining who the payer could be, and therefore which types of facilities—CLC, CNH, SVH, Medicare/Medicaid NH—were fiscally realistic possibilities for Veterans when choosing an NH from among the options available.

Theme 2: Community Living Centers

Mandatory Veterans given priority for CLC beds but optional Veterans also served (2.a).

VHA staff reported that VHA personnel gave mandatory Veterans preferred placement in CLCs over other Veterans in need of NH care. According to interviewees, the extent to which optional Veterans were served in CLCs depended, in part, on bed availability and local fiscal circumstances. It was reported that optional Veterans served in CLCs tended to lack the financial resources to pay for care in other settings because they did not have private insurance or were not eligible for Medicare or Medicaid.

Long-term placement in CLCs typically for mandatory Veterans, with exceptions (2.b).

VHA staff reported that although largely reserved for mandatory Veterans, certain optional Veterans received long-term care in CLCs. The latter included optional Veterans with no other payment options, histories of sexual abuse or violence, or homelessness, or required specialized care needs that could not be met elsewhere (e.g., behavioral problems).

Short-term placements in CLCs represented a mix of mandatory and optional Veterans (2.c).

VHA staff reported that if beds were available, mandatory Veterans in need of short-term NH placements were typically admitted to CLCs, reflecting, in part, medical staff belief that care coordination, care continuity, and outcomes would be better than if CNH placement took place. VHA staff also reported that optional Veterans in need of short-term care were often served in CLCs as well. Reasons include fewer administrative burdens arranging care and superior coordination, continuity, and outcomes than when community NH placement occurred. They also include lower out-of-pocket costs incurred by Veterans in CLCs (limited to no co-pays) than in community NHs under Medicare’s skilled nursing facility benefit (20% coinsurance, days 21–100).

Shifting CLCs to short-term admissions (2.d).

Many interviewees reported that CLCs were increasingly being used for short-term postacute and rehabilitative care with long-term placements being reduced or eliminated and referred to community NHs, even for mandatory Veterans. This shift, in part, reflected a desire to make CLC beds more readily available for patients discharged from inpatient hospital care and to increase the bed turnover rate, so that greater numbers of Veterans could be served. VHA staff reported that most VAMCs had made progress rebalancing CLC admissions toward short stays and that the few long-stay patients who remained at these facilities either consisted of cases grandfathered in under prior rules or new cases representing the difficult to place Veterans noted earlier. Interviewees from a few VAMCs, however, reported that their CLCs had either struggled or chosen not to conform with the increased short-stay emphasis, and, thus, continued to serve predominately long-stay residents.

Theme 3: CNHs

CNHs served mandatory Veterans with long-term care needs (3.a).

VHA staff reported that most CNH placements were for mandatory Veterans, typically referred for long-term care. By contrast, most VHA personnel reported that optional Veterans were typically admitted to a community NH under a non-VA payment source.

Veterans selected from among available payment sources for community NH care (3.b).

Interviewees reported that it was, ultimately, up to Veterans themselves to choose which available payment source—VA, Medicare, and other—that they were initially placed under. Interviewees explained that Veterans who received NH care paid for by non-VA sources of payment were not limited by the list of CNHs within their VAMC’s catchment area but could generally choose an NH from among the broader pool of community NHs willing to serve them. On the other hand, mandatory Veterans with no other payment options had a strong incentive to select an NH from among the list of CNHs because otherwise the VA would not have paid for their care. Veterans dually eligible for Medicare and VA payment (or triply eligible, in the case of U.S. military’s program, TRICARE) could choose whether to utilize their VA or Medicare (or TRICARE) benefit. Veterans eligible for multiple programs often chose to go out on a VA contract, even when non-VA sources (Medicare, TRICARE) would cover them fully for the first 20 or 100 days. These decisions stemmed, in part, from a desire to avoid potential co-pays once their non-VA benefits ran out. They also stemmed, in part, from a desire by some Veterans to use a benefit that they felt that they had earned. Other times, VHA staff indicated that dually eligible Veterans opted to go out under Medicare because their preferred NH or the one that best met their needs was not on a VA contract, or altruistically, because using Medicare freed up VA funds for other Veterans without alternative sources of payment.

Beginning VA payment immediately despite Medicare eligibility (3.c).

VHA staff reported that some VAMCs tended to prefer that mandatory Veterans begin VA payment immediately even though Medicare might otherwise have paid for the first 20 days of care in a CNH. Several reasons were indicated. First, sites did so out of concern that Veterans might be liable for co-payments if’ stays lasted greater than 20 days. Second, sites did so to avoid the administrative burden of switching Veterans to VA payment once the Medicare co-pays started. Third, sites did so due to risk that NHs could go without payment if the Centers for Medicare and Medicaid Services (CMS) declined to pay because, in the agency’s view, the primary payer should be the VA.

Beginning VA payment after using Medicare to pay for community NH care (3.d).

VHA staff reported that some VAMCs tended to prefer that mandatory Veterans eligible for both VA and Medicare NH payment first use Medicare, beginning VA payment when the Medicare co-pay would start, after 20 days. Interviewees believed that initial reliance on Medicare payment increased the number of facilities available for placement, thereby maximizing Veteran choice. They also believed that the initial reliance on Medicare resulted in financial savings for referring VAMCs by providing them with an alternative to VA payment. They felt that it benefited receiving facilities as well which typically receive higher reimbursement under Medicare than VA. Some VHA staff reported that CMS had made clear that mandatory Veterans should enter CNHs under a VA contract, not using their Medicare benefit; that CMS might otherwise refuse to pay. Some VAMCs, however, asserted that Veterans had the right to choose which benefit would pay for their care under VA regulations. VA staff communication with CMS on behalf of NHs following denial of Medicare payment for mandatory Veterans seemed to resolve matters in favor of NHs’ Medicare payment claims, according to one interviewee.

VAMCs sometimes provided optional Veterans with short-term contracts in community NHs (3.e).

VHA staff indicated that contract exceptions were made when optional Veterans were referred to VA-paid NH care in a CNH. It was reported that VHA staff made contract exceptions to free up inpatient hospital beds. This was especially true at VAMCs that lacked CLCs to serve, in part, as a release valve to turnover scarce inpatient hospital capacity. VA staff also reported that contract exceptions were made if Veterans’ skilled nursing needs were expected to result in NH stays that extended beyond the period in which Medicare provided full coverage or if they were otherwise uninsured but their care needs might have better been met in a community NH than in a CLC. Mentioned by interviewees were 30-day contracts, Medicaid pending. Here, the objective was to get a Veteran’s foot in the NH’s door while other payment sources were sorted out, often with the aim of facilitating discharge from the VA hospital.

Fiscal considerations often primary when VAMCs determined whether to pay for optional Veterans in CNHs (3.f).

VHA staff reported that the requisite resources and motivation for issuing short-term contracts to optional Veterans varied across VAMCs, with better resourced sites often better able to fund nonmandatory Veterans than less well-resourced sites. Moreover, it was reported that as fiscal pressures rose, VAMCs often became more restrictive about who received contracts for short-term care in community NHs. One Southern VA, for example, used to provide short-term contracts to certain optional Veterans until pressure came down from the VISN to eliminate the practice due to an increasingly adverse budgetary environment.

Theme 4: SVHs

VHA Staff rarely referred Veterans directly to SVHs due to lengthy waitlists (4.a).

Interviewees reported that discharge planners seldom referred Veterans straight to SVHs because of waitlists of 6 months to a year or more for admission. Because of waitlists, Veterans who were eventually admitted to an SVH typically experienced an intervening stay in another setting before SVH placement took place.

SVHs served primarily nonmandatory Veterans (4.b).

Typically, the cost of caring for optional Veterans in SVHs was divided evenly among the state, VA, and Veteran. VHA staff reported that this division in financing made SVHs a particularly attractive option for Veterans with no other payment options or those who would have had to pay the full cost out-of-pocket before spending down to Medicaid eligibility in a community facility. Thus, when interviewees reported referring Veterans to an SVH, one of the primary reasons was a lack of other payers available for those Veterans’ NH stays.

SVHs were a financially attractive referral option for VAMCs serving mandatory Veterans (4.c).

VHA staff reported that relatively few mandatory Veterans used SVHs. Those that did so, however, qualified for the same level of financial support no matter whether they received care in an SVH, CNH, or CLC. The only difference was the source of that support varied, with two thirds typically deriving from the VA and one third from the state in the case of SVHs as compared to 100% VA in the case of CNH and CLC care. Some interviewees suggested that it might have been financially beneficial for VAMCs to refer mandatory Veterans to SVHs rather than CNHs or CLCs because VA Central Office paid the VA’s contribution to SVHs, rather than the local VAMCs, which paid for CNH and CLC care.

SVHs served long-term residents with comparatively lower care needs than CNHs and CLCs (4.d).

VHA staff reported that SVHs served long-stay residents with lower care needs compared to CNHs and CLCs. According to interviewees, one reason that care needs tended to be less intensive in SVHs was that the lengthy waitlists for admission tended to preclude short-term postacute and rehabilitative placements. Another reason noted by interviewees was the availability of a domiciliary or assisted living option in some SVHs, which was otherwise unavailable to these Veterans and certainly much less expensive than what could be found in the community.

Discussion

Findings indicate that the most important financial characteristic informing placement in a CLC, CNH, or SVH was Veterans’ eligibility for VA-paid NH care. This result suggests that VAMCs understand and support the VA’s mission to provide NH care to eligible Veterans. Consistent with other reports (Office of Inspector General [OIG], 2006a), VA personnel report conforming to Millennium Act requirements that they prioritize NH access for mandatory Veterans while serving other Veterans to the extent that capacity allows. The impressions of VA personnel are also largely consistent with the cross-setting distribution in Veterans served nationally (U.S. GAO, 2013), including the preponderance of mandatory and optional placements to CNHs and SVHs, respectively, and mix of mandatory and optional placements to CLCs, particularly for postacute and rehabilitative care.

Findings suggest that VAMC NH referral practices reflected, in part, a desire to achieve operating efficiencies in managing patient workflow. This search for operational efficiencies is illustrated by the adoption of strategies to free up scarce inpatient hospital capacity, a common theme throughout the health sector (Collier & Harrington, 2005; McAuley et al., 1997; Nazir et al., 2014; Shugarman & Brown, 2006). These strategies were, in turn, consistent with current practice, which suggests that once medically stable, patients should be transferred out of acute therapy to the least restrictive environment possible. Findings suggest that the CLC program is often used as a mechanism to transition both mandatory and optional Veterans to independent living while restoring function to prehospitalization levels. They also reveal a preference for CLC over community NH placement for this purpose because providing postacute and rehabilitative care to Veterans previously hospitalized at the CLC eliminated administrative burdens associated with arranging community NH care while improving the coordination and continuity of the care provided.

VA staff reported that eligibility for VA-paid NH care increased the array of NH placement possibilities from which Veterans may have chosen, including CLCs, SVHs, and CNHs, and potentially, community NHs paid for privately or by Medicare or Medicaid. Thus, it is important that all potential payment sources be identified and placement options associated with each assessed. Indeed, VHA staff highlighted the prominence of Veterans’ choice in selecting the payment source under which NH placement occurred, consistent with VA expectations that Veterans serve as active participants in the care planning process (Department of VA, n.d.). It was reported, for example, that some mandatory Veterans preferred to enter a community NH under VA payment even though they could have been placed under Medicare. By contrast, other mandatory Veterans preferred to go under Medicare first because doing so gave them more facilities to choose from or because they wished to prioritize VA funding for other Veterans with no alternative payment sources.

VA staff reported that mandatory Veterans opting to use their VA benefit to receive NH care within the community had to choose an NH from among the list of NHs that contracted with their particular VAMC; otherwise, the VA would not have paid for their NH care. On average, however, NHs participating in the CNH program tend to provide lower quality care than community NHs that do not contract with the VA (Geriatrics & Extended Care Data Analysis Center, 2015). Impediments to expanding the list of high-quality CNHs include burdensome contracting requirements and too few referrals in light of the level of reimbursement received (Miller, Gadbois, Gidmark, & Intrator, 2015). This suggests that alternative contracting practices should be explored to enable mandatory Veterans to choose from a larger set of NHs and still receive VA payment. Recently, the VA has considered allowing VAMCs to use Medicare procedures to enter into provider agreements with CNHs (Halvorson, 2013). These agreements would eliminate contracting and reporting requirements associated with VA contracts (Gatty, 2014). The VHA also recently adopted the Community NH Dashboard that uses CMS five-star performance data to compare the quality of VA CNH and non-CNHs within each VAMC market. This information should better enable contracting officials to evaluate the quality of CNHs and to identify other high-quality NHs for potential contracting.

Findings suggest that fiscal issues played a role in variation in regional and local practices regarding the placement of Veterans eligible for multiple payment sources. This result is consistent with reports that highlight the discretion VISNs and VAMCs possess in allocating financial resources across NH and other service offerings (OIG, 2006a, 2006b, 2013; U.S. GAO, 2013). The influence of existing resource constraints is reflected in the tendency of some of the VAMCs studied to prefer initial placement under Medicare for Veterans otherwise eligible for VA-paid community NH care. It is also reflected in the perceived benefit of SVH placement to VAMCs, both because it is cheaper and because Central Office rather than the VAMC picks up the tab. It is further reflected in the observed role that the local fiscal environment plays in influencing the extent to which optional Veterans were placed in CLCs or in CNHs on short-term contracts. This variation in optional Veterans served is consistent with the national data. In 2012, for example, close to two thirds (62%) of the VHA’s average daily NH census consisted of optional Veterans, but the proportion of optional Veterans served ranged from 40% to 69% across VISNs (U.S. GAO, 2013).

Fiscal issues are not the only local factor that may influence the prodvision of NH benefits to optional Veterans. Cross-site differences in service availability and demand may contribute as well (OIG, 2006a; U.S. GAO, 2013; U.S. Department of VA, 2009). Most VAMCs studied had one affiliated CLC, though some had two or more while a few lacked this option. Moreover, CLCs ranged in size from 15 to 20 beds to over 100. Nationally, regional and market differences in noninstitutional alternatives to NH placement are also evident (OIG, 2006b; U.S. GAO, 2003). Variation in the number, size, vicinity, and scope of VA extended care options relative to local demand for those services likely impacted the extent to which VHA staff allocated VA-paid NH benefits to optional enrollees with those sites with greater (lesser) NH and lesser (greater) home and community-based services capacity being more (less) likely to do so.

Future research should explore the relationship between service capacity, demand, and the number of optional Veterans served. It should also examine the role of VA budgeting and resource allocation practices in influencing these decisions. VA Central Office allocates resources to the VHA’s 18 regional networks (VISNs) using a risk-adjusted per person allocation formula, adjusted for labor/contracting costs and other factors (U.S. GAO, 2011). VISNs, in turn, allocate resources to the VAMCs within their jurisdictions. It may be that the resource allocation methodology employed does not sufficiently account for case-mix differences precipitating NH use. It may be that the methodology does not sufficiently account for the costs of’serving lower priority groups, that is, optional Veterans. It may be that discretion VISNs and VAMCs possess in distributing the funds allocated to specific service modalities contributes to varying levels of resources available for NH care relative to other priorities.

Findings suggest that the VHA’s (2005, 2012) emphasis on shifting CLCs admissions from long-term to short-term care has filtered through to most of the VAMCs studied with CLCs increasingly being used for postacute care and rehabilitation and Veterans in need of long-term care being referred to CNHs. At the same time, findings suggest limitations in transforming the nature of CLC care. Staff from a few VAMCs reported struggling or choosing not to conform with the increased short-stay emphasis. Some Veterans, especially those who had lived in CLCs for many years, were “grandfathered” into permanent residency under the Millennium Act. Other Veterans were difficult to place elsewhere. These Veterans included those with histories of sexual abuse/violence, homelessness, no other payment options, and specialized care needs that could not be met in available community facilities. Further research should be devoted to identifying social, geographic, and/or case-mix characteristics hindering VAMC compliance with VA policy in this area. This includes determining the extent to which the VHA admits CLC residents for long periods who may have been potential candidates for alternative HCBS placement (Mor et al., 2007), a VHA priority since 1999 passage of the Millennium Act (U.S. GAO, 2003).

Limitations

There are several potential study limitations. First, we investigated how Veterans’ eligibility for multiple payment sources impacted NH referral at 12 VAMCs only. Although we believe that our findings may nevertheless be generalizable given the range of VAMCs studied and the common policy environment within which they operate (i.e., VA Central Office regulations), external validity may be limited due to variation in local norms and market conditions that influence program availability, interpretation of policy, types of personnel, and reporting structure. Moreover, the VHA has undergone changes since the collection of our interviews, which may lessen the applicability of study findings to the current environment. In 2014, President Obama signed the Veterans Access, Choice, and Accountability Act of 2014 (Public Law 113–146) into law to improve Veterans’ access to care in response to reports that larger numbers of Veterans never received requested doctor appointments or had to undergo unreasonably long waits before doing so (U.S. GAO, 2017). The target of this legislation, however, was primary and specialty care, rather than the focus of our research, geriatrics, and extended care services. The regulations and policies that govern the VA’s three NH programs have remained unchanged since we conducted our interviews.

Second, we used a combined purposive-snowball sampling approach in the absence of a known sampling frame. Even though we are confident that we interviewed most key staff, potentially knowledgeable personnel may have been overlooked. Third, this study took a hospital discharge planning perspective; findings do not reflect community-based clinics, which may have different practices and norms around referrals. Fourth, we did not interview Veterans and family members about their perceptions of the role of payment source in NH referral so do not know the extent to which their impressions complement, reinforce, or oppose those of VHA staff. Last, the study focused entirely on extended care referral within the VHA; as such, care should be taken when applying our findings to outside the VA.

Conclusion

This study examined VHA staff perceptions of how Veterans’ eligibility for VA and other payment sources impacted NH referral across 12 VAMCs. Findings highlight the preeminent role that VA priority status played in determining placement in VA-paid NH care. They indicate that CLC, CNH, and SVH admission was driven by Veteran preferences and choice and contingent, in part, on potential payment source (VA, Medicare, Medicaid, and other) and anticipated length of stay. They also indicate that variation in optional Veteran referrals to VA-paid NH care across VAMCs derived, in part, from differences in local and regional policies and markets, most notably with respect to availability and payment. VA Central Office should concentrate its efforts on monitoring the care provided and enforcing prevailing quality standards in order to ensure that Veterans receive the best care wherever they are placed no matter region or setting. VA Central Office should also determine the extent to which VAMCs’ discretion in the provision of NH care to nonmandatory Veterans results in differences in utilization by social, geographic, and/or case-mix characteristics.

Author Biographies

Edward Alan Miller, PhD, MPA, is a Professor of Gerontology & Public Policy and Fellow, Gerontology Institute, at the University of Massachusetts Boston, and Adjunct Professor of Health Services, Policy & Practice at the Brown University School of Public Health. Dr. Miller’s research focuses on understanding the determinants and effects of federal and state policies affecting frail and disabled elders. He edits the Journal of Aging & Social Policy and is a Fellow of the Gerontological Society of America.

Stefanie Gidmark, MPH, received her master’s degree at Brown University investigating the relationship between body image, obesity and HIV in South Africa. She is currently a project coordinator at the Providence VA Medical Center where she participates in data acquisition and analysis and prepares research reports, grant proposals and presentations.

Emily Gadbois, PhD, is a project director at the Center for Gerontology and Health-care Research at Brown University. Last year, she completed her doctorate in Gerontology at the University of Massachusetts Boston, having successfully defended her dissertation, Understanding How State Policy Influences Driving Status in Older Adults.

James L. Rudolph, MD, SM, is Director of the Center of Innovation in Long-Term Services and Supports at the Providence VA Medical Center, and Associate Professor of Medicine at Brown University’s Warren Alpert Medical School. His innovative clinical research, program development and leadership involve identifying at risk older patients during hospitalization and intervening at that critical juncture. An overriding goal is to measure, manage, and improve the services VA provides to vulnerable elder Veterans across the nation.

Orna Intrator, PhD, is a Research Health Scientist at the Canandaigua VA Medical Center, Director of the VA Geriatrics & Extended Care Data & Analysis Center, and Professor of Public Health Sciences at the University of Rochester School of Medicine and Dentistry. Dr. Intator’s research focuses on the needs and services provided to older adults who require long-term services and supports, paying particular attention to public policy and organizational management. She is a Fellow of the Gerontological Society of America.