How hospital clinicians select patients for skilled nursing facilities

Abstract

Background/Objective

Hospitalized older adults are frequently transferred to skilled nursing facilities (SNFs) for post-discharge care. Hospitals and SNFs are increasingly held jointly accountable for costs and outcomes of this care. However, it is unclear how clinicians select patients for SNF care; we sought to understand how hospital-based clinicians evaluate older adults in the hospital and decide who will be transferred to a SNF for post-acute care.

Design

Semi-structured interviews paired with a qualitative analytic approach informed by Social Constructivist theory.

Setting

Inpatient care units in three hospitals. We used purposive sampling to maximize variability in hospitals, units within hospitals, and staff on those units.

Participants

Twenty-five clinicians involved in evaluation and decision-making regarding post-acute care, including hospitalists, nurses, therapists, social workers, and case managers.

Measurements

Central themes related to clinician evaluation and discharge decision-making.

Results

Clinicians described pressure to expedite evaluation and discharge decisions, resulting in the use of SNF as a “safety net” for older adults being discharged from the hospital. The lack of hospital-based clinician knowledge of SNF care practices, quality, or patient outcomes resulted in the lack of a standardized evaluation process or a clear primary decision-maker.

Conclusion

Hospital clinician evaluation and decision-making about post-acute care in a SNF may be characterized as rushed, without a clear system or framework for making decisions, and uninformed by knowledge of SNF or patient outcomes in those discharged to SNF. This leads to SNF being used as a “safety net” for many older adults. As hospitals and SNFs are increasingly held jointly accountable for outcomes of patients transitioning between hospital and SNF, novel solutions for improving evaluation and decision-making are urgently needed.

INTRODUCTION

“Despite a general sense about where patients needing care should be treated, hospital discharge placement is far from an exact science…. Each decision to move an older person out of the hospital or along the long-term care continuum can affect the rest of that person’s life.”¹

Hospitalized older adults are increasingly discharged to skilled nursing facilities (SNFs) for post-acute care², leading to post-acute care becoming the most rapidly growing area of Medicare costs.^3–6 These older adults are commonly readmitted to the hospital^7–9 and often fail to rehabilitate and return to the community^10–13, leading to long-term institutional care placement.^10,14 The proportion of patients discharged to SNF who return home by 100 days post- hospital discharge is as low as 30%.¹² Recent legislation enforcing joint hospital-SNF responsibility for these outcomes and their associated costs may intensify the focus on patient selection for SNF^15–17, which has received relatively little evaluation.¹

There is substantial variability in patient selection for SNF across regions and providers; regional variability in the use of post-acute care explains 73% of all geographic variability in Medicare spending in the United States.^18,19 Further, hospitalists who care for older adults in the hospital²⁰ may have little familiarity with SNFs²¹ broadly or in their local area. Since there are no validated tools available in the public domain that reliably predict benefit from a SNF stay, recommendations vary between clinicians with different training.^22,23

In this context, we sought to understand how hospital-based clinicians select patients for SNF after hospital discharge. We focused on the in-hospital evaluation of older adults and the discharge decision-making process as key inputs to patient selection. In doing so, we hoped to gain insight into how patient selection for SNF could be optimized to achieve better outcomes for older adults.

METHODS

Study Design

We used purposive sampling to maximize diversity of hospitals, clinical units within these hospitals, and clinicians on these units. For example, the three participating hospitals included a VA hospital serving a predominantly male population often with significant comorbidity and poor social supports^24,25, a University hospital serving a large referral base as well as a predominantly urban population, and a safety-net public hospital serving a predominantly indigent and immigrant population. Clinical units included general medical wards, an Acute Care for the Elderly unit²⁶, and an orthopedic surgery unit. Clinicians with input on discharge decision-making (including hospitalist physicians, nurses, occupational and physical therapists, social workers, and case managers) were sampled based on their relative years of experience and experience working in the hospital or post-acute care setting. We interviewed at least one clinician of each type (hospitalist, nurse, occupational therapist, physical therapist, and either social worker or case manager) within each of the three hospitals. Participants provided verbal consent to participate. The study was approved by the Colorado Multiple Institutional Review Board.

We used the principles of Social Constructivist theory, which has been used previously in clinical decision-making research, to frame our study design and analysis.²⁷ This theory frames how clinicians, as a group, work together to create a system or environment that influences decision-making processes.^28–30

Data Collection and Analysis

Between February and September 2016, qualitative analysts (EL, AL, RA) conducted a 20–60 minute semi-structured, in-depth interview with each participant. We started with open-ended questions such as “Can you tell me about a time when you discharged an older adult to a SNF or rehabilitation facility?” We then explored their experience using a semi-structured interview guide that was informed by theories and frameworks related to transitions of care as well as clinical experience.^31–34 This interview guide was developed by the study team and iteratively refined based on pilot testing with feedback from the interviewee; pilot interviews were not included in the analysis. Topics included: evaluating the need for post-acute care; selecting post-acute care options; attitudes, knowledge, and beliefs about SNFs; influences on discharge decision-making, including team dynamics; and post-discharge follow-up. Interviews were audio-recorded, professionally transcribed, validated, and analyzed in Atlas.Ti (v7.5.11; Scientific Software Development, Berlin, Germany). Participant demographics were collected using a brief questionnaire before the interview. We employed a team-based approach to thematic analysis, beginning the analytic process while completing interviews to continue to refine the interview guide and identify new themes to pursue in interviews.^35,36 Codes were developed inductively through group discussion; additional codes that emerged were applied to all previously coded manuscripts until thematic saturation was reached.³⁷

RESULTS

We interviewed 25 clinicians across the 3 hospital sites, including 5 physicians, 7 occupational or physical therapists, 6 nurses, and 7 social workers or case managers (Table 1). We identified two perceived “drivers” that most strongly influenced patient evaluation and clinician decision-making, and two main perceived consequences of these drivers in terms of patient selection. The first driver was pressure to expedite patient evaluation and discharge decision-making, resulting in the use of SNF as a “safety net” for older adults being discharged from the hospital. The second driver was the lack of hospital-based clinician knowledge of SNF care practices, quality, or patient outcomes, resulting in the lack of a standardized evaluation process or a clear primary decision-maker for selecting patients for SNF care.

Table 1.

Participant characteristics

Characteristic	N (%)
Facility of Interview
Facility A - University	12 (48)
Facility B - VA	5 (20)
Facility C - public	8 (32)
Type of Provider
Hospitalist	5 (20)
Nurse	6 (24)
Occupational Therapist	3 (12)
Physical Therapist	4 (16)
Case Manager	2 (8)
Social Worker	5 (20)
Years of experience, mean (range)
Hospitalist	4 (1–10)
Nurse	4 (1–12)
Occupational therapist	25 (15–41)
Physical therapist	8 (4–14)
Case manager	2 (1–3)
Social worker	5 (1–10)
Race/Ethnicity, n, (%)
White	16 (64)
Latino/a	2 (8)
Other (Biracial/Mixed)	4 (16)
Missing	3 (12)
Veterans, n, (%)	1 (4)
Women, n, (%)	22 (88)

Driver: Pressure to expedite evaluation and discharge decision-making

“The reason this [discharge process to SNF] doesn’t work well isn’t because everyone doesn’t care, it doesn’t work well because doing the right thing is not the same as doing the easy thing, and it’s very hard with limited time to be able to always do the right thing for every patient, ‘cause it would take many more hours than there are in a day.” (Hospitalist)

Nearly all hospital-based clinicians identified pressure to create space for new admissions, pressure created by limited staffing resulting in incomplete evaluation of patients in the time available, and external pressure from payers to shorten hospital length of stay.

Pressure to “open up beds” was felt by nearly all hospital-based clinicians. As a result, planning and assessment for post-discharge care often occurred on the first day of hospitalization. (Supplemental Table S1, Quote 1a). Following this initial evaluation, hospital-based clinicians described frequent internal pressure from the hospital and external pressure from payers to discharge patients (Quotes 1b and 1c). These pressures become particularly acute when the medical team decides the patient is “medically ready” for discharge, but needs post-acute care. Social workers and case managers spoke of transmitting these pressures to patients and their family (Quote 1d). Hospital-based clinicians also described pressure to try to fully evaluate what seemed like too many patients in too little time, leading to expressions of strain (Quotes 1e and 1f).

Consequence: Use of SNF as a “safety net” in older adults being discharged from the hospital

“I do think [SNF] is utilized…along the lines of trying to find a place to put somebody who can’t quite go home, and I think from that standpoint…we defeat the purpose. It continues to be just a place that, well, they’re done medically, they don’t have any needs here in the hospital, but they can’t go home and we don’t really have any place to put them….we utilize these skilled nursing facilities as a kind of repository.” (Hospitalist)

Faced with these pressures, clinicians described trying to place patients into one of two categories as early as possible in the hospitalization: those that could return home, and everyone else. SNF discharge was a common default option for patients in the latter category, because it could be arranged more quickly than other care options (Quote 2a).

Patients who were “on the border” between these categories were also encouraged to go to SNF, because this was felt to be “safer” from the perspective of hospital-based clinicians (Quotes 2b and 2c) and would allow more time for recuperation and rehabilitation in the hope of the patient returning home. However, even patients whom hospital-based clinicians thought would never return to the community were discharged to SNF for rehabilitation (Quote 2d), because this was seen as the most expeditious way to connect the patients to what was perceived as their real need – long-term nursing home care. In these cases, SNF was justified by clinicians as treating hospital-based deconditioning (Quote 2e). The prevailing attitude of clinicians was one of resignation that the “system” was constructed this way. Some even expressed feelings of frustration (Quote 2f).

While SNF was most commonly identified as a bridge to long-term nursing home care in these populations, it was also seen as a solution to difficult discharges of homeless or socially isolated patients (Quotes 2g and 2h).

Driver: Lack of knowledge of SNF care delivery, quality, or patient outcomes

“I would love if everybody involved, and I include physicians in that, had a better sense of what this process is like, what a SNF is like, what happens there, what doesn’t happen there…” (Hospitalist)

Most clinicians confessed they did not know what level of care could be provided in a SNF. Surprisingly, this did not vary between clinician roles (Quotes 3a, 3b, 3c). Hospital clinicians did seem to believe there was significant variability in SNF quality and care practices (Quote 3d). Hospital-based clinicians not only lacked exposure to SNFs, they also lacked feedback about the quality of their evaluation (Quotes 3e and 3f) and subsequent outcomes of patients they discharged to SNF (Quotes 3g and 3h).

Consequence: Lack of standardized process or clear decision-maker for selecting patients for SNF care

“It’s just looking at the whole, the big picture I guess, it’s kind of everything together…I mean, there’s not like a checklist that you go through. Just kind of everything together.” (Physical therapist)

Since clinicians did not know much about the care provided in a SNF nor learned of patient outcomes related to their discharge decision-making, they did not have formalized “criteria” for selecting patients. As a result, neither the patient evaluation criteria nor the clinician decision-making process was standardized, and clinicians did not identify evidence-based tools they used as a part of their decision-making. Instead, clinicians described trying to quickly gather enough information about a variety of factors (Table 2) in order to reach a “gestalt” sense about whether the patient could safely return home (Supplemental Table S1, Quotes 4a and 4b).

Table 2.

Information hospital-based clinicians commonly evaluate to help select patients for SNF

Data point	Quote	Clinician (clinical setting, years of experience)
Physical function	We also consider strongly prior level of function…insurance is…more likely to approve if they’re functioning below their baseline status.	Physical therapist (University, 4)
Social support	I just know that sometimes the alternative to go home is worse because there’s nobody at home…you’ll get family members like, ‘I can’t take care of him. I don’t know what you’re going to do with him but I can’t do anything with him.’ And that’s tough.	Nurse (University, 1)
Cognition	Patients that have maybe dementia…that’s a hard one I really struggle with, like what’s their ability to learn, what’s their ability to make progress…what are their abilities to actually learn some new skills and improve their balance so that they can go home.	Physical therapist (Safety-net, 8)
Pain	If pain isn’t controlled, then to me, they’re not ready to go. To me that’s criteria number one, I mean you’ve gotta have pain controlled, well controlled so that they can work with therapy, and obviously, if they’re going to rehab because they’re mostly there for therapy…	Physical therapist (VA, 5)
Home Environment	He’s [the patient] definitely kind of not meeting the criteria as far as for discharge home. He lives independently, he has stairs, again, has to be pretty independent obviously with him living alone, but in this case we’re recommending subacute rehab to kind of help…	Physical therapist (VA, 5)
Insurance status	I think the biggest factor I see is insurance. Does this person have resources for that beyond because…we get a lot of people that are self-pay or no benefits, and then it becomes a situation, well what can we do for them at home, and that’s kind of tough because you know that they would benefit from a skilled nursing facility.	Nurse (University, 1)
Medical readiness	I guess the decision…has more to do with, do I expect there to be a decline in status in the next 24 to 48 hours and can the place where we’re sending the patient do the things that need to be done for this patient, and if not, then they stay.	Hospitalist (University, 10)
Patient goals and motivation	A big part is gonna be, does the patient want to be there [SNF], and is the patient willing to actively participate. If the patient is in the hospital and says I’m not going to participate and then we have a couple days in the hospital where they are refusing therapy in the hospital…it’s gonna be a waste of everyone’s resources, right, to send them.	Social worker (University, 2)
Value of SNF	So if we focused on…explicitly thinking about what we can offer the patient and what a short-term stay would do for them, if we say yes, they would benefit then I think that it’s something that can really be improved upon.	Hospitalist (University, 1)

This heterogeneous approach also meant it was unclear who ultimately made the decision for any particular patient to go to a SNF. While most clinicians cited the patient as the ultimate decision-maker if the patient had decision-making capacity, many noted the patient would often go along with the physician’s recommendation (Quote 4c). However, physicians often did not feel like the ultimate decision-maker (Quotes 4d and 4e) and suggested therapists ultimately decided the disposition of the patient, an idea therapists did not concur with (Quote 4f). Nurses were not involved in the process (Quote 4g). Several case managers spoke of legal restrictions to recommending one SNF over another, and as a result uniformly described providing patients with a list of SNFs closest to their homes, and allowing the patient to make the decision regarding which they preferred.

The lack of familiarity with SNF practices or patient outcomes following a SNF stay also resulted in variability in when a patient was considered “ready” to discharge to SNF (Quotes 4h and 4i). Without a clear sense of criteria for SNF placement and variable methods of evaluation, patients could have been discharged to a variety of post-discharge support options at different points in their inpatient stay.

Perhaps as a result, hospital-based clinicians’ estimates of how many patients discharged to a SNF from the hospital successfully rehabilitate and return home ranged from 45–90%, with significant variability within and across provider groups. A contrasting example was provided by hospital clinicians working on the orthopedic surgery wards, where the criteria for discharge to home versus a SNF following elective joint replacement were made clear to the patients and clinicians up-front according to a relatively strict protocol even prior to the introduction of a bundled joint replacement program at their hospital (Quote 4j).

Interactions between themes and clinician-proposed solutions

While important in isolation, these themes were often all present in a single case, and there was dynamic interplay between them (Table 3). Clinicians also identified solutions to their perceived barriers to optimal selection of patients for SNF. These included more provider education about SNF capabilities and patient outcomes in SNF, as well as a standardized checklist for evaluating patients that might lead to a scoring system or tool to predict benefit from a SNF stay (Table 4).

Table 3.

Case examples of interplay of drivers and results in patients selected for SNF

Exemplar case	Clinician (clinical setting, years experience)
We’re taking care of an elderly woman who was living in an assisted living facility and has had repeated falls, so was admitted to the hospital. She has about 30 different reasons for having falls. She had an anoxic brain injury so is not cognitively totally normal, however, she is conversant, however that is not in the records…so some confusion about who is the decision maker (Lack of standardized process or clear decision-maker for selecting patients). It’s [the facility] closing next month so the goal is for her not to return there…sometimes that becomes very complicated…unless it’s a different level of care, which is the case here due to her fall (SNF as safety net). Then today her potassium…is a little bit elevated, so we’re trying to make decisions of is she medically still safe for her to go or not. Will she be able to get blood work within two or three days and have someone actually look at it when she arrives, when she goes to her facility (Lack of knowledge of SNF care delivery, quality, or patient outcomes)… it’s challenging for me to know whether that could or would happen and whether it will happen correctly. At the same time..there’s always kind of forces at play of getting people out of the hospital as soon as able (Pressures to expedite evaluation and decision-making).	Hospitalist (University, 1)
61-year-old male and I encountered him I believe initially on the Obs unit and he couldn’t remember why he was here. He needed about mod assist with bed mobility, just some poor planning, couldn’t figure out how to get out of bed on his own, once he got up and started walking with a walker, he did pretty well and I would probably call him more supervision (Lack of knowledge of SNF care delivery, quality, or patient outcomes) …I said, ‘do you ever fall at home,’ he said ‘oh 2 to 3 times a day, I fall down the stairs in front of my trailer’, I was like ‘you told me you didn’t have stairs’, and he said ‘oh yeah, I fall down those 2 to 3 times a day and there’s no handrail.’ Turns out he has a roommate but she works and she’s not really there full-time. I don’t think that there was anything …for his reason for admission that would’ve been like an acute change in cognition where you would think that that would necessarily get better, so it kind of leads me to believe that it’s a little bit more chronic (Lack of standardized process or clear decision-maker for selecting patients). He’s somebody that, … it’s a really hard call because he’s getting around OK, but it’s really more that he’s not safe from like a cognition point of view and falling all the time. (SNF as safety net)	Physical Therapist (Safety-net, 8)

Table 4.

Clinician suggestions to improve patient selection process

Suggestion	Quote	Clinician (clinical setting, years experience)
Educate hospital clinicians about SNF capabilities	I would love if everybody involved, and I include physicians in that, had a better sense of what this process is like, so what a SNF is like, what happens there, what doesn’t happen there, what the timeframe is for that, if people had a better sense of the actual transition…	Hospitalist (University, 1)
Identify variability in SNF care quality or outcomes	If there were certain centers that did better or had better physical therapy…I think that’d really change my way or my view of which centers or sites that I would like to send my patients to, and then I can counsel the patients a little bit better about those sites…… I think would be the most helpful if there was some way to really assess level of care, level of physician involvement, level of physical therapy at all of these potential centers that these patients go to next.	Hospitalist (VA, 1)
Checklist for standardizing evaluation of patients	Taking all these things into consideration…maybe even like a spreadsheet of oh, did all of these different disciplines see this patient… maybe even a tool that you check off boxes of… all different family members and disciplines have you talked to and the process of looking for facilities, too, cause I think the process of looking for facilities, especially when you’re new can be very confusing. I think I would’ve especially found it handy to have a tool.	Case manager (University, 3)
Scoring system or tool to predict likelihood of patient benefit from SNF stay	I think if there was some sort of you know, marker or score system that could show effectiveness of therapy or likelihood of improvement and ability to have an outcome which is consistent with going home at the end of the short-term subacute stay, it would be really helpful clinically.	Hospitalist (VA, 1)
	Boy, if there’s a tool that we can employ or if I had a better sense of how to make sure that what I’m saying is the best thing for this patient from a functional perspective, that would be fantastic…	Physical therapist (University, 14)
Feedback on patient experience and outcomes from SNF	What I almost would like….just feedback from patients after they’ve gone. You know, like a short survey just so that we know, cause I know what liaisons tell me, but I don’t really ever see patients again unless for some reason, unfortunately, they’re readmitted.	Case manager (University, 3)

DISCUSSION

Hospital clinicians’ selection of patients for post-acute care in a SNF was heavily influenced by internal and external pressure to expedite their evaluation and decision-making. However, the process for gathering relevant information and making discharge decisions was largely based on clinical “experience.” Unfortunately, this experience was neither informed by personal knowledge of SNFs nor fine-tuned through feedback about patient outcomes in SNFs. The pressure to quickly identify which patients could not “safely” discharge home, coupled with uncertainty about when patients were “appropriate” for SNF, led to SNF being used as a “safety net.”

In sum, these findings may suggest hospitals (and by proxy SNFs) will face significant challenges in a new era of shared accountability for patient outcomes and costs. This accountability is enforced by programs such as Bundled Payments for Care Improvement, Medicare Spending Per Beneficiary (MSPB) metrics^15,16, and new SNF readmission and community discharge measures that will be used for value-based purchasing by Medicare.¹⁷

These new programs emphasize there may be no more vulnerable time in the life of an older adult than the period following acute hospitalization. Failure to rehabilitate often results in long-term nursing home placement,^13,14 and those who are readmitted to the hospital have high short-term mortality rates.¹⁰ Most importantly, preliminary evidence indicates patient outcomes are directly affected by the type of post-acute care received.^38,39 Improved evaluation of the match between patient needs and post-discharge care, coupled with informed decision making regarding SNF placement, is crucial to optimize outcomes.

This study provides novel insights into the attitudes, knowledge, and beliefs of hospital-based clinicians regarding post-acute care in a SNF. Others have established that different types of clinicians have different attitudes about each type of long-term care,²³ and that medicine residents have limited knowledge about SNFs.^1,21 The pressure to discharge patients quickly is well-described,³² and may result in patients being discharged from the hospital “quicker and sicker.”^33,40,41

Our results suggest several potential avenues for improving selection of patients for post-acute care in SNF. First, it is important to establish standardized, evidence-based criteria for evaluation. The intent of the Continuity Assessment Record and Evaluation (CARE) tool is to standardize evaluation; it has not yet been validated to predict the optimal post-discharge care for an individual patient.⁴² The timing of evaluation could be delayed until the patient has met specific medical criteria (e.g., normal vital signs, recovery of baseline mental status) to best assess functional ability.⁴¹ Further, the relative roles of clinicians in evaluation and decision-making must be made explicit, including the processes each uses to evaluate patients for post-acute care. These processes and findings could be shared through dedicated time in daily interdisciplinary rounds. The involvement of caregivers and their ability to support patients at home could be more robustly tested in the hospital prior to discharge and caregiver support could be enhanced.^44,45 Fine-tuning patient selection requires improved knowledge of SNF capabilities and context; novel models for exposing housestaff and hospitalists to SNF are needed. Ultimately, the current “push” model from hospitals to SNFs may require a paradigm shift to a “pull” model, where SNF clinicians who know the post-discharge context are more involved in the decision-making process. Our results also suggest the need for systems for obtaining feedback about post-discharge outcomes in patients discharged to SNF.⁸

These interviews also suggest the need for tools that better predict prognosis in SNF. Preliminary evidence suggests loss and subsequent recovery of function may follow a predictable pathway in a majority of patients.⁴⁵ This, coupled with better prognostic scores for predicting end-of-life, may improve decision-making that is congruent with a patient’s long-term goals.^46,47

There are also policy solutions to consider. Currently, there are no widely-available mechanisms to facilitate rapid placement of hospitalized older adults into home-based primary care, assisted living, or long-term care. Clinicians felt they could not recommend one SNF over another for fear of violating Stark laws or Anti-Kickback statutes.⁴⁸ Applying waivers more broadly for hospitals and SNFs trying to enhance their collaborative care for transitioning patients may improve care.

The results of this study should be interpreted in the context from which they were derived. Strengths include purposive sampling of hospitals, wards, and clinicians with a robust number of interviews to reach thematic saturation. Analysis was conducted as a team with triangulation to clinicians to verify results, including interim presentations of results to external stakeholders to inform further analysis. However, all hospitals were located in Denver, and attitudes, clinician behaviors, and hospital relationships with SNFs may vary by region or population density. All three hospitals have housestaff teaching programs, and results may be different in community hospitals. Hospitalists were the only physician group sampled as they provide the majority of inpatient care to older adults, but other clinicians may have had alternative responses. We acknowledge there are “missing voices” in this analysis that are relevant to decisions about post-acute care in a SNF (such as those of patients, caregivers, primary care physicians, and SNF clinicians); future work will focus on additional stakeholders to understand their perspectives. To confirm the generalizability of our findings, they should be studied in a larger, more representative sample of respondents in future work.

Hospital clinician evaluation of older adults and decision-making about post-acute care in a SNF may be characterized as rushed, without a clear system or framework for making decisions or decision-maker, and uninformed by knowledge of SNF or patient outcomes in those discharged to SNF, leading to SNF being used as a “safety net.” As hospitals and SNFs are increasingly held jointly accountable for outcomes of patients transitioning between hospital and SNF, novel solutions for improving evaluation and decision-making are urgently needed.

Supplementary Material

Supp Table S1

Supplemental Table S1: includes all supporting quotes cited in Results above.