Abstract
Individuals with serious mental illness, prisoners, and ex-offenders needing skilled nursing facility (SNF)–level care are difficult to place in traditional SNFs. SNFs accepting these historically marginalized individuals may offer them a more appropriate level of care. We compared health services use (emergency room (ER) visits, acute hospitalizations), total number of antipsychotic medications prescribed, and quality-of-life indicators (depressive symptoms, cognition, resident behaviors), before and after admission, among 86 individuals admitted to a Connecticut SNF for persons difficult to place. Residents were racially diverse, primarily male (89%), and 58.4 (±12.5) years of age; 56 percent were transferred from state psychiatric facilities. Twelve-month hospitalization rates decreased from 36.5 to 10.6 percent, 27 percent of those taking an antipsychotic medication at admission experienced a reduction in total number of antipsychotics prescribed by six months, and 13 residents transitioned into the community. Quality-of-life indicators did not change between admission and first quarterly assessment. A SNF for persons difficult to place may help prevent hospitalizations, optimize antipsychotic medication use, and serve as an intermediate step into the community. These findings may inform development of an evidence-based model for establishing SNFs in other states for persons who are difficult to place.
With the aging of the U.S. population, an unprecedented number of older adults will meet preadmission criteria for requiring skilled nursing facility (SNF)–level care in coming years.1 Included among the growing number of aging individuals requiring SNF care are historically marginalized persons who are regarded as difficult to place in Medicaid-paid SNF settings: those with severe mental illness (e.g., schizophrenia and bipolar disorder) and inmates and ex-offenders.
The U.S. Supreme Court decision in Olmstead v. L.C. (527 U.S. 581 (1999)) largely initiated a push toward deinstitutionalization of people with mental illness and admission to community and residential care settings.2 However, residents with severe mental illness (SMI) in state-run psychiatric facilities who have a history of behavioral symptoms or have frequent exacerbations of their illness thereby requiring hospitalization, may benefit from a higher level of care than complete deinstitutionalization.3,4 As these individuals age and acquire chronic conditions that limit function, they also may need SNF–level care. Yet, an unintended consequence of the Olmstead decision has been that individuals with SMI are more likely to be admitted to SNFs that have government-issued deficiency citations for both overall and clinical care quality.5 The growing need for SNF-level care may permit nursing home administrators at high-quality facilities to deny admission to selective individuals with SMI, given concerns about disruptive and potentially dangerous behaviors.6,7 As a result of this selectivity, if SNF placement for a person with SMI is found, the facility is likely to have more lenient admission standards, which translates to more quality deficiencies and a lack of an appropriate level of nursing and psychiatric care for these complex residents.8
In addition, over the past two decades, a large number of psychiatric hospitals in the United States have closed. Inadequate development of community-based health services to offset these closures has contributed to a disproportionate number of people with mental illnesses who are incarcerated.9 The criminalization of those with SMI who are now aging in prison, together with factors including minimum sentencing laws, more arrests at later ages, and population aging in general, have resulted in a 500 percent increase since 1990 in the population of U.S. inmates age 50 and older.10,–,12 Consequently, although exact estimates are not available, the number of inmates requiring SNF-level care is expected to increase substantially.13
Inmates and ex-offenders are also often denied SNF admission because of safety concerns, regardless of the severity of their prior offense or the amount of time that has passed since they committed the offense.14 Furthermore, many states have adopted legislation that has made the process of admitting an ex-offender to a SNF extremely cumbersome (e.g., background checks, resident notification, and Department of Correction–specified plan for supervision and monitoring), thereby increasing the likelihood of denied admission.14,15 These individuals often have no alternative but to remain in a state-run psychiatric facility or state correctional facility, neither of which may be appropriately equipped to provide SNF-level care.
A SNF specifically for individuals transitioning from state-operated psychiatric or correctional facilities may enhance the quality of care and quality of life of persons who are difficult to place by providing appropriate care that is otherwise unavailable. In February 2012, the Connecticut Department of Mental Health and Addiction Services (DMHAS)16 issued a Request for Proposal (RFP) to identify and enter into a contract with a vendor to provide “nursing home level of care for individuals in state care who are difficult-to-place.” These individuals are defined as “transitioning from a correctional facility, have criminal justice involvement and/or [are] individuals transitioning from a higher level of care provided by the Department of Mental Health and Addiction Services.” The May 2013 opening of one such facility in Connecticut (Public Act 11-44, §117; Public Act 12-1, §104) provided a new option for long-term care of these patients. To optimize care for this unique population, this SNF provides staff with specialized sensitivity training that emphasizes maintaining the dignity of residents. Staff are taught awareness of the environments from which the residents are admitted and how their prior surroundings may affect adjustment to a nursing home setting. Staff training also includes mindful approaches to dealing with aggressive behaviors, understanding how mental illness may be complicated by dementia, and working with residents with a history of traumatic brain injury. The staff to resident ratio of this 95-bed facility is considerably higher than what is specified in the Public Health Code of the State of Connecticut (Sec. 19-13-D8t). Between the hours of 7 a.m. and 11 p.m., the ratio is 1 to 6 for Certified Nursing Assistants and 1 to 25 for Registered Nurses. These ratios are 1 to 15 and 1 to 45, respectively, between 11 p.m. and 7 a.m. The staff retention rate is tracked and used as a guide to gauge how well leadership ensures staff support. In addition, there is a social worker for each unit and a psychiatric consultation group that includes a psychiatrist, social worker, psychologist, and advanced practice registered nurse (APRN) that meets regularly with residents and staff to support individualized care planning. Furthermore, recreational programming often focuses on specific interests of the residents (e.g., music, sports, and food) and “giving back” (e.g., coordinating a bake sale for staff and visitors and donating proceeds to veterans' organizations). We sought to use routinely collected data to describe this unique SNF population and to evaluate the SNF's potential impact on key resident outcomes.
Methods
Sample
The Institutional Review Boards of the University of Connecticut Health Center and the Connecticut Department of Mental Health and Addiction Services exempted the study from review. A total of 97 individuals were referred to the specialized SNF from either a Connecticut state psychiatric facility or correctional facility between May 1, 2013, and May 31, 2015. In addition to meeting nursing home level of care/preadmission screening and resident review (PASARR) processes, admission referrals had to include evidence that all community options had been explored and that other nursing homes would not admit the individual, and had to provide an explanation detailing why the specialized SNF would be the most appropriate, least restrictive setting for the individual. In addition, per Public Act 12-1, § 104, individuals cannot be referred from a correctional facility if they have been “convicted of a capital felony under the provisions of section 53a-54b of the general statutes in effect prior to April 25, 2012, or murder with special circumstances under the provisions of section 53a-54b of the general statutes in effect on or after April 25, 2012.” (The State of Connecticut repealed the death penalty on April 25, 2012, per Substitute Senate Bill No. 280 Public Act No. 12-5; An Act Revising the Penalty for Capital Felonies.) A specific team of nurses assessed each referral for admission, focusing on individuals needing skilled nursing care whose challenging behaviors were likely to be mitigated by behavioral care plans established before admission. A total of 86 (90.6%) of the persons referred were admitted.
Measures
Data from the nursing home Minimum Data Set (MDS) version 3.0 were collected as part of routine practice at admission and quarterly thereafter, or until death or discharge. The (MDS) is:
… part of the federally mandated process for clinical assessment of all residents in Medicare and Medicaid certified nursing homes. This process provides a comprehensive assessment of each resident's functional capabilities and helps nursing home staff identify health problems. MDS assessments are completed for all residents in certified nursing homes, regardless of source of payment for the individual resident. MDS assessments are required for residents on admission to the nursing facility, periodically, and on discharge. In most cases, participants in the assessment process are licensed health care professionals employed by the nursing home. MDS information is transmitted electronically by nursing homes to the national MDS database at CMS.17
The MDS includes some previously validated scales, as well as measures designed and tested for the MDS specifically. For the purposes of this study, data from MDS Sections C (Cognition),18 D (Depressive symptoms),19 and E (Behavior) were used and refer to the past 7 days. Cognition is assessed via the Brief Interview for Mental Status (BIMS), which includes several questions measuring repetition (repeating 3 words), temporal orientation (correctly indicating the year, month, and day of the week), and recall (restating each of the three words provided in the repetition question, either with or without a cue).18 BIMS scores range from 0 to 15, and higher scores indicate better cognition, with scores of <7 indicating severe impairment. Depression was assessed with the nine-item Physician Health Questionnaire (PHQ-9).19,20 The PHQ-9 asks respondents to indicate how often they have been bothered by problems that map directly to the nine DSM-IV criteria used for establishing major depressive disorder (e.g., loss of interest or pleasure; thoughts of hurting oneself or being better off dead).21 Scores range from 0 to 27, with higher scores indicating greater severity. Behaviors were assessed as the overall presence of physical (e.g., hitting, kicking, grabbing), verbal (e.g., screaming, cursing), and other (scratching self and throwing food or bodily waste) behavioral symptoms. For the purposes of this study, MDS admission data were compared with residents' 3- and 6-month MDS quarterly assessments. Utilization review data identified ER visits and acute inpatient hospitalizations (i.e., unplanned medical or psychiatric admissions) occurring up to 12 months before and after SNF admission and also provided data on discharge location or death. Finally, medical record review conducted by a medical student researcher identified antipsychotic medication use via prescriptions for antipsychotic medications at admission and 6 months later. Data were reviewed independently by two study investigators to ensure consistency in coding, and were deidentified before transfer to the principal investigator (L.C.B.).
Data Analyses
Descriptive statistics were calculated for participants' characteristics. Chi-square or Fisher's exact tests for categorical variables, and t tests and Wilcoxon rank-sum tests for continuous variables were used to compare participants' demographics, preadmission health services use (ER visits, acute inpatient hospitalizations, and use of antipsychotic medication), and the MDS elements (cognition, depression, and behavior) according to preadmission facility (state psychiatric facility or state correctional facility). Chi-square tests compared the difference between the proportion of ER visits and hospitalizations in the 12 months before and after SNF admission for the sample overall. Linear mixed models were used to examine the change in MDS cognition and depression scores over time (admission, three months, six months). We then reran these models while controlling for pretransfer facility. We also compared the posttransfer hospitalization and ER rates between those admitted from a psychiatric facility and those from a correctional facility by using the chi-square test. Finally, chi-square tests and Wilcoxon rank-sum tests were used to evaluate differences in the MDS elements between those who were prescribed versus those not prescribed antipsychotic medications at admission. All statistical tests used p < .05 to indicate statistical significance. Data were analyzed by using SAS version 9.4.
Results
The 86 SNF residents were 58.4 ± 12.5 (SD) years of age (range, 27–92 years), ethnically and racially diverse (51.2% white, 31.4% black, and 17.4% Hispanic), and most were men (89.5%). As compared with the 48 (55.8%) residents transferred from a state psychiatric facility, the 38 persons referred from a correctional facility were significantly younger and had significantly higher rates of preadmission ER visits and acute inpatient hospitalizations, but lower rates of antipsychotic medication use (Table 1). Those transferring from a correctional facility also had higher and better average MDS cognition scores. Furthermore, the 10 residents with cognitive impairment (score <7 on the cognition screen) had all transferred from a psychiatric facility. Similarly, seven of the nine residents who could not complete the cognitive assessment at admission had transferred from a state psychiatric facility. Depression symptom severity scores and behavioral symptoms did not differ significantly between the two groups at admission.
Characteristics of Residents of a Specialized Skilled Nursing Facility According to Pretransfer Facility
Table 2a and Table 2b present data regarding change after admission in health services use and quality-of-life indicators. Among the 86 SNF residents, 12-month hospitalization rates decreased significantly from 36.5 percent pretransfer to 10.6 percent posttransfer (p value for difference in proportions <.001). The 12-month pre- and posttransfer ER visit rates did not differ (14% and 13%). Mean cognition (F = 0.08; p = .93) and mean depression symptom severity scores (F = 1.18; p = .31) did not differ significantly over time for the sample overall. Nor did mean cognition (F = 0.40; p = .67) or depressive symptom scores (F = 0.82; p = .44) change over time when controlling for pretransfer facility.
Change in Health Services Use Among Residents of a Specialized SNF
Quality-of-Life Indicators Among Residents of a Specialized SNF
As indicated previously, the pretransfer hospitalization rates and pretransfer ER rates were higher among those who transferred from a correctional facility. In contrast, in comparing those who had transferred from a psychiatric facility versus a correctional facility, respectively, the posttransfer hospitalization rates (10.6% versus 10.5%; χ2 = .0; p = .99) and the posttransfer ER visit rates (8.5% versus 18.5%; χ2 = 1.4; p = .24) did not differ significantly. Finally, as indicated by MDS-Section E data, eight (9.3%) residents exhibited behavioral symptoms in the week preceding SNF admission: one experienced physical behaviors only, three experienced verbal behaviors only, and four experienced a combination of physical, verbal, and other behaviors. Six of these 8 residents experienced a repeated behavioral incident in the 12 months after admission. An additional nine residents experienced at least one behavioral incident over this time frame, for a total of 15 residents experiencing behavioral symptoms in the year after admission. The type of behavioral symptoms (e.g., physical, verbal, other) after admission was not available.
Residents with a prescription for an antipsychotic medication at admission (n = 37) had significantly more behavioral symptoms at admission (28.2% versus 4.4%; χ2 = 5.13; p = .003) and had worse cognition (mean (SD)): 10.1 (4.0) versus 12.83 (2.9); Z = −3.52; p = .0004) or were unable to complete the cognitive screen (18.6% versus 2.3%; χ2 = 6.08; p = .02) as compared with those not taking an antipsychotic at admission. Table 3 presents antipsychotic medication use data for the 77 residents with medication data available at baseline and six months later. Of those, 33 had a prescription for at least one antipsychotic medication at admission. Twelve (36.4%) experienced a change: 9 (27.3%) experienced a reduction, and 3 experienced an increase in the total number of antipsychotic medication prescriptions. Of the five residents who experienced a decrease from two or more to one antipsychotic medication prescriptions, three had an as-needed (PRN) prescription removed, and two had a twice-per-day (BID) prescription removed. The three residents who experienced an increase from one to two or more antipsychotic medication prescriptions did not have a different medication added. Rather, a prescription for the same medication as either a PRN (n = 1) or a once-per-day (QD), short-acting medication (n = 2) was added. Of the 40 residents not taking an antipsychotic medication at admission, 4 (10.0%) were taking at least one antipsychotic medication at six months.
Changes in the Number of Antipsychotic Medication Prescriptions for Specialized SNF Residents Between Admission and Six Months
By November 30, 2015, 11 (12.8%) residents had died (with 5 dying within 6 months of admission), and 29 (33.7%) had been discharged. Of these 29 people, 9 were discharged to a nonspecialized SNF (31.0%), 6 returned to a correctional facility (20.7%), 1 returned to a state psychiatric facility (3.4%), and 13 (44.8%) were discharged into community-based settings. Of those discharged into the community, nine had been admitted from a state psychiatric facility.
Discussion
Individuals with SMI and aging prisoners and ex-offenders needing SNF–level care are difficult to place in typical Medicaid-funded SNFs. The recent opening of a SNF for persons who are difficult to place provides an alternate source of long-term care for residents who are aging in other, perhaps less appropriate, institutional settings. Using routinely collected data, we sought to describe this population and better understand the potential impact of the specialized SNF on health-related outcomes. Notably, we found that the rate of acute hospitalizations decreased in the year after SNF transfer. Some residents transferred to the SNF for end-of-life care. For others, the facility served as an intermediate step before a subsequent move to a nonspecialized SNF or return to the community.
Although each of the residents of the specialized SNF needed to meet a minimum level of care eligibility criteria for SNF admission, our findings suggest that the type of pretransfer facility has implications for determining care planning and service needs. In addition to being older, residents admitted from a state psychiatric facility were more likely to exhibit other characteristics consistent with a dementia diagnosis, including antipsychotic medication use, more severe cognitive impairment, and more behavioral symptoms at admission as compared with those admitted from a state correctional institution. As previously noted, however, individuals with a history of significant behavioral problems were not referred for placement. In contrast, although few residents admitted from a correctional facility had probable dementia, these individuals had a high proportion of ER visits or hospitalizations in the past year, likely indicating significant physical illness. Future SNFs dedicated to caring for individuals who are difficult to place may consider the primary referral source as a factor when structuring how to best meet incoming residents' needs (e.g., memory-impairment units and rehabilitative services).
More than a third of the SNF residents experienced hospitalization in an acute-care inpatient facility in the year before transfer. In contrast, only 10 percent were admitted to an acute-care hospital in the year following transfer. This lower rate may indicate a more appropriate care environment; chronic, debilitating conditions may be more optimally managed to prevent unnecessary hospitalizations. Furthermore, as most of those hospitalized before SNF admission transferred from a state correctional facility, this finding may be of particular interest to state departments of correction. A cost–benefit analysis was beyond the scope of this study, but reports have documented the considerable costs of caring for prisoners with substantial physical and medical needs.22,23 Admitting individuals with corrections involvement to the SNF may translate to cost savings in areas including prisoner transportation to or from a medical facility and 24-hour correctional officer supervision of the hospitalized patient. There was no difference in the rate of ER visits in the years before and after SNF transition. However, this finding, together with the reduced hospitalization rate, may indicate that the SNF is using the ER judiciously for unavoidable acute occurrences rather than as a possible prelude to acute inpatient hospitalization.
We also found that average depressive symptoms scores and cognition scores did not change over time. Although we would not expect to see rapid changes in cognition, symptoms of depression could change over a three- or six-month period after a major residential transition. Average depression scores were very low at admission and remained low, perhaps reflecting optimism about the new residential environment. Earlier assessment of depression before individuals are referred to transfer and a longer-term posttransfer depression assessment would provide a fuller picture of depression symptom patterns for this group.
Individuals with SMI and prisoners are often difficult to place in traditional nursing homes because of concerns regarding the potential for problematic behaviors. Yet, our findings and those of others do not support this assumption regarding behavior. A 12-year retrospective analysis of one California-based SNF found that only 10 (22%) of 46 residents with correctional histories exhibited behaviors that potentially posed safety risks to staff and other residents,24 and none of them had been directly admitted from a correctional facility. Another study found that SNF residents with SMI were no more likely to display physically aggressive or socially inappropriate behavior than residents with dementia but no SMI.7 We found that only 15 (17%) of the specialized SNF residents exhibited behavioral symptoms during the year after admission, 6 of whom had behavioral symptoms in the week before admission. Although this low rate likely reflects the selective nature of the referral and admission process, it is still considerably lower than that reported for typical SNFs, where about 75 percent of residents with dementia (or 50% of all residents) display behavioral symptoms.25 Future research evaluating outcomes of this specialized SNF model should include more detailed tracking of adverse behavior, as resident-to-resident abuse is common in other SNFs.26
Nearly 43 percent of the residents from the specialized SNF were taking an antipsychotic medication on admission. Although this rate is considerably higher than has been reported,27 it likely reflects, at least in part, the unique characteristics of this population. However, by 6 months, 27 percent of those taking an antipsychotic at admission experienced a reduction in the number of prescriptions for antipsychotic medications. These reductions may reflect national efforts to reduce antipsychotic use in nursing homes. Furthermore, our findings may suggest more prudent use of PRN antipsychotic medications. However, we did not have information regarding dosage changes, nor did we know the reasons that prompted a prescription change. More in-depth investigation is needed to characterize antipsychotic medication use in this unique population and to understand whether changes in antipsychotic medication prescribing affect outcomes, including mood and cognition.
Finally, the specialized SNF was an intermediate step to a community-based setting for 15 percent of the 87 residents admitted during its first two years of operation. This finding indicates an important and unique role for this facility in the state of Connecticut's larger long-term services and supports system rebalancing goals. Furthermore, an additional 10 percent of the residents were transferred to another SNF. These transfers suggest that initial placement at the specialized SNF may have increased other traditional SNFs' willingness to admit individuals who are difficult to place. In future research, understanding the reason for these transfers (e.g., resident moves to a SNF that is closer to family) will help to determine how the specialized SNF may affect the quality of care and quality of life of some of society's most vulnerable and historically marginalized persons.
There are several additional limitations that should be acknowledged. Despite its widespread use as a data collection tool, the MDS has been criticized for concerns regarding reliability and data accuracy.28 However, two of the three MDS measures used (BIMS and PHQ-9) are well-validated scales developed independently and subsequently included in the MDS because of their measurement strength.18,–,20 Because we focused on the results from one facility, nurse assessors were likely to have received similar training, thereby minimizing the likelihood of inter-rater differences. Although the minimal number of residents with problem behaviors at admission reflects the judicious selection process, these results may not generalize to all individuals who are difficult to place and therefore may not reflect the experience of future SNFs dedicated to this population. Finally, as previously indicated, we did not have data regarding reasons that residents were transferred from the specialized SNF to another location.
This evaluation of the first-of-its-kind specialized SNF for persons who are difficult to place indicates that this type of facility may have a positive impact on the quality of care and quality of life of these vulnerable individuals. The specialized SNF resulted in significantly fewer hospitalizations and, for many of those who were discharged, served as an interim step before final transition into the community. These findings may help to inform the development of an evidence-based model that other states can use to establish specialized SNFs for individuals who are difficult to place, subsequently providing long-term services and support in a more appropriate and cost-effective setting for this population.
Acknowledgments
The investigators thank the Connecticut Department of Mental Health and Addiction Services for participating in the research project.
Footnotes
This work was supported by a 2013 “Another Look: Better Health for Elders in Care Facilities” grant (Grant Number 20130498) from The Patrick and Catherine Weldon Donaghue Medical Research Foundation, West Hartford, CT.
Disclosures of financial or other potential conflicts of interest: None.
- © 2018 American Academy of Psychiatry and the Law
