Pai Fe Z JSC ASPE_ || Office oF BEHAVIORAL HEALTH, a sea sane RetearTon DISABILITY, AND AGING POLICY Resident and Facility Factors Associated with High Risk of Discharge from Nursing Facilities, 2012-2017: Final Report Prepared for the Office of the Assistant Secretary for Planning and Evaluation (ASPE) at the U.S. Department of Health & Human Services by RTI International September 2022 Office of the Assistant Secretary for Planning and Evaluation The Assistant Secretary for Planning and Evaluation (ASPE) advises the Secretary of the U.S. Department of Health and Human Services (HHS) on policy development in health, disability, human services, data, and science; and provides advice and analysis on economic policy. ASPE leads special initiatives; coordinates the Department's evaluation, research, and demonstration activities; and manages cross- Department planning activities such as strategic planning, legislative planning, and review of regulations. Integral to this role, ASPE conducts research and evaluation studies; develops policy analyses; and estimates the cost and benefits of policy alternatives under consideration by the Department or Congress. Office of Behavioral Health, Disability, and Aging Policy The Office of Behavioral Health, Disability, and Aging Policy (BHDAP) focuses on policies and programs that support the independence, productivity, health and well-being, and long-term care needs of people with disabilities, older adults, and people with mental and substance use disorders. Visit BHDAP at https://aspe.hhs.gov/about/offices/bhdap for all their research activity. NOTE: BHDAP was previously known as the Office of Disability, Aging, and Long-Term Care Policy (DALTCP). Only our office name has changed, not our mission, portfolio, or policy focus. This research was funded by the U.S. Department of Health and Human Services Office of the Assistant Secretary for Planning and Evaluation under Contract Number #HSP2332016000211 and carried out by Research Triangle Institute. Please visit https://aspe.hhs.qov/topics/long-term-services-supports-long- term-care for more information about ASPE research on long-term services and supports (LTSS). RESIDENT AND FACILITY FACTORS ASSOCIATED WITH HIGH RISK OF DISCHARGE FROM NURSING FACILITIES, 2012-2017: FINAL REPORT Authors Kelly Hughes Zhanlian Feng Nan Tracy Zheng Sarita L. Karon RTI International lara Oliveira U.S. Department of Health and Human Services Office of the Assistant Secretary for Planning and Evaluation September 12, 2022 Prepared for Office of Behavioral Health, Disability, and Aging Policy Office of the Assistant Secretary for Planning and Evaluation U.S. Department of Health and Human Services Contract HHHSP2332016000211 The opinions and views expressed in this report are those of the authors. They do not reflect the views of the Department of Health and Human Services, the contractor or any other funding organization. This report was completed and submitted on September 2020. Contents Section Page Executive Summary ES-1 1. Background and Purpose 1-1 2. Methods 2-1 2.1 Data SOUPCES..... ccc nee nee eee 2-1 2.2 Study POPulation.........ccccccceceeceneeeeeeeeeeeeseeneeneeneeeaenaeeeeeeeteeteetnaenasnaseaneantnes 2-1 2.3 Statistical ANALYSES .........ccececcee ec eee eee eee cence een eens een e ee nee ESSER ESE EEE EES 2-1 3. Findings 3-1 3.1 Nationally, what are the characteristics of residents who are discharged live from nursing facilities, compared to residents who remain in the FACTILY? oo. ccc cee cece eee eee eee een een eee SESE ASSES EROS ROSAS SAOSR OSE OSE OST OS ESS SROSROS EOS EOSE OEE OSS 3-1 3.1.1 Characteristics of Nursing Facility Residents in 2017 ........ccceceeeeeeeeeeeees 3-1 3.1.2 Characteristics of Residents Discharged Live from Nursing Facilities OVER TING 00. cee eee cece cence eee eee eee n RRS E DEEDES ERS ERS ERE 3-6 3.2 What Are the Differences in Characteristics Among Residents Discharged Live Across All States and Types of Facility? .........:ccceccsceseeseeeeeneensenaeeaeeaetenes 3-9 3.2.1 Differences Across States .......cccccsccecesceneeneeeeeeseeeeeseteeteaenaenasnaneaeeenees 3-9 3.2.2 Differences by Facility TyP@S..........ccccsceeeeeeeeeeeeeeseeeeteeteaenaenasnaeeaeeags 3-11 3.2.3 Discharges from Terminated Facilities ............cccccesececseeeeeeeeeeeeeseeeaes 3-13 3.3 Do Post-discharge Outcomes Differ Among Residents Discharged Live With and Without Risk Factors for Live Discharge? ......cccccceccsceceseeeeeeeeeaeeeeaneneanees 3-14 3.3.1 Acute Care USC ......cceccccce eects seen eens eens nese eens eens ease tans nas ena S nant aE 3-14 3.3.2 Mortality ........ccccccceeceece eee e reese eset eee eeeneeneenaenaetaeteeeententsaenannannaneaeeggs 3-22 3.3.3 Analysis of Discharge Location .........ccccceseeeeeeeceeeeeeeetenteaeeasnaeeaneaneags 3-23 3.3.4 Outcome of Residents Discharged Live from Terminated Facilities ...... 3-24 4. Discussion 4-1 4.1 SuUMMary Of FINGINGS.........ccceeceeeeee seer erst eee e eens eneeneeeeeeeetaetaetsaenaenaenaneaenanees 4-1 4.1.1 Identification of RISK FactOrs...........cccccecseeeeeeeeeeeeeeeeeeeeeeaeaeaseeesnenenanas 4-1 4.1.2 Variation of Risk Factors Across State and Facility Types.........ccccccssee 4-2 4.1.3 Outcomes of NF Residents After Discharge .........c::scscsseeeeeeeeeaeeeneeeneass 4-2 4.1.4 Insights into FIDS.........cccceceeceece eee ee esse eeeneeneeneeeseeseegeeseenasnasnaseaseantags 4-3 4.2 Policy IMPHiCAtiONS ........ccccceceeeeeeeeeeee estes eeeeeeseeeeeeeeeeeeeeeeeereneeneaeagneaeaneaeeneags 4-5 4.3 LIMITATIONS oo ccc teeter reer errr ee eee eee Ree ROR R SESE C SECS SSS ESE 4-6 References R-1 Appen dic mOO@ > es Technical Details .........cccccececeee eee e eee e eee eee e eee eens esses eee eet ee eee ese S ESSERE ES EE ESTES A-1 M@thOdS........ccceceee eee eect eee eee ene ee A een Een S REEDS REEDS EE EES E DEES EDS ES ESS ES EERE SESE EEE TEES B-1 Additional Graphs and Data Tables for Risk Factors .......ccccssceseeeeneeneeeaeneeeenees C-1 Data Tables for State and Facility Characteristic...........cccsccseesseeeeneeneenaeeaeeees D-1 Graphs and Data Tables for Outcomes After Discharge ........::cccceeeeeeeeeeeeeeneees E-1 Figures Number Page 3-1. Change in Behavioral Symptoms over the Past Year by Discharge Status ........ Error! Bookmark not defined. 3-2. Incontinence at End of NF Stay and Change over the Past Year by Discharge SUAtUS.cccccce eee eee REESE SEC SECS ECS S ARSE RSE R SERS ER SE COSC SSC SS ESSE ROSE O EERE 3-4 3-3. Transitions to Medicaid Eligibility by Discharge Status............. Error! Bookmark not defined. 3-4. Differences in the Prevalence of Behavioral Symptoms between Residents Discharged Live and Not Discharged .........ccccccceeesceseeeeseeseeeeeeeeeaeaeeaeeeeanestanestanes 3-7 3-5. Differences in New Diagnoses of Psychiatric and Mood Disorder ...........::ecseeeeeeeees 3-8 3-6. Differences in Transitions to Medicaid Eligibility ........cccccecsscececeereeeeeneeseneeseneesneas 3-9 3-7. Percentage Point Differences in the Prevalence of Physical Behavior at End of Stay, by State ......cccccccccecesccceecceeeuecueveuccnvanvaneunuauuanuausaunvanvanvansaneuasuanuanvanvansa 3-10 3-8. Percentage Point Differences in the Prevalence of Severe Behavior, "Other", at End of Stay, by Facility SiZ@.........ccccccceceeeeeeeeeeneeneeneeneenaeeeetsetententnasnaenaseaneaetes 3-13 3-9. Hospitalization of Residents within 30 Days of NF Discharge, by Behavioral Symptoms Aggregated across Behavioral Types, at End of Stay ..........csccseeeeeeees 3-16 3-10. Outpatient Acute Care Use of Residents within 30 Days of NF Discharge, by Behavioral Symptoms Aggregated across Behavioral Type, at End of Stay.......... 3-17 3-11. Outpatient Acute Care Use of Residents with and without New Diagnoses of PSYCHOTIC DISOICer .......ccceceece eee e cence eee eens eee e seen eee O eRe SHOSROS COS TOS HOST OSE RaSROSROS HOSE OEE ES 3-19 3-12. Hospitalization of Residents with and without Severe Urinary Incontinence, at End Of Stay .....cccccccececceeceececcuvcuecuuenevancanvancausauuanuauueaseanvanvansaneansuauuasuansanvansas 3-20 3-13. Outpatient Acute Care of Residents with and without a Change in ADL to the most Severe Level Of DEPEndency ......cccceecscesceneeneeneeneeeaeteeteetententnaenasnaseaseantes 3-21 3-14. Mortality by Acute Care USage .......ccccccceceeeeeeeneeneeneeneenaeteeteetententnasnasnaneaneannes 3-23 Tables Number Page 3-1. Percent of Residents with New Diagnoses of Psychiatric and Mood Disorders in the Past Year by Discharge Status ........ccccsccseeseeneeneeeeeneeeeeeeeteetenteaenaenaeeaseaenenees 3-3 3-2. Percentage Point Differences in the Prevalence of Transitions to Medicaid Eligibility by Facility Profit Status..........cccccsccsceeceneeneeneeeeeeeeteetentestnaenaseaseaseaetes 3-12 3-3. Hospitalization of Residents within 30 Days of NF Discharge, by Psychiatric rT aTelm i (oleleim B) ie) ge (-| anne 3-18 3-4. Hospitalization of Residents within 30 Days of NF Discharge, by Change to Severe IMpairmMent, PErCeNt....cccccecececeeceeeeeeeeeeeeeeeeeeneeeeeeeeeaeeaeeeeeaaenaenannaneanss 3-20 3-5. Hospitalization and Outpatient Acute Care of Residents within 30 Days of NF Discharge, by Medicaid Eligibility Status ..........cccccseseeseeseeeeteeteeteeenaenaenaeeaeeantes 3-22 iv Acknowledgements We thank Robert Baker and Douglas Kendrick for their programming support throughout this study, and Michael Lepore for his review of an early version of this report. Acronyms The following acronyms are mentioned in this report and/or appendices. ABS ADL CMS COVID-19 CY ED FFS FID ID/DD MBSF MDS NF OIG POS PPS PTSD SNF vi Aggressive Behavioral Scale Activity of Daily Living Centers for Medicare & Medicaid Services Novel Coronavirus Calendar Year Emergency Department Fee-For-Service Facility-Initiated Discharge Intellectual Disabilities/Developmental Disabilities Master Beneficiary Summary File Minimum Data Set Nursing Facility Office of the Inspector General Provider of Services Prospective Payment System Post-Traumatic Stress Disorder Skilled Nursing Facility Executive Summary Key Takeaways e This study identified the following risk factors for live discharges from nursing facilities: severe behavioral symptoms; impairments requiring more staff time; transitions to Medicaid eligibility; and psychiatric and mood disorders. e Risk factors were consistent across years but varied in presence and strength by state and facility type. Residents with risk factors were more likely to be discharged from for-profit, government and chain facilities than non-profit and non-chain facilities. e There were high rates of acute care use and mortality among residents discharged live from nursing facilities. Rates of post-discharge acute care use were higher among residents with risk factors than residents without risk factors. ES.1 Background People can be discharged from nursing homes for many reasons. Discharges may be a positive outcome and reflect an individual's choice. In other cases, discharges may be at the direction of the facility and against the will of the resident. There are strict rules about when involuntary facility-initiated discharges (FIDs) are allowed. Under federal law, involuntary FIDs are legal only if they are triggered by one of six reasons: the FID is necessary for the resident's welfare and the resident's needs cannot be met in the facility; the resident no longer needs nursing home care; the safety of other residents is endangered due to the clinical or behavioral status of the resident; the health of others in the facility would be endangered; the resident fails to pay; or, the facility closes (CMS, 2017a). FIDs are unlawful otherwise. The Office of the Inspector General (OIG) is currently investigating the extent to which state survey agencies investigated and took enforcement actions against nursing homes for inappropriate involuntary transfers and discharges (OIG, 2019). FIDs can be unsafe and traumatic for the residents involved and may result in higher costs of care, raising concerns to the public and to Medicare and Medicaid programs, which are the primary sources of payment. In the last few years, news media and Ombudsman programs report that FIDs may be increasing in frequency and as of 2018, were the leading cause of complaints for nursing home residents (Verdana & Pear, 2018). Recently, there has been heightened concern about the rise of FIDs amid the COVID-19 pandemic (Silver-Greenberg & Harris, 2020). This study builds on a recent review of available literature (Lepore, Zepeda & Yuen, 2018), which identified the following characteristics as placing residents at higher risk of potentially unlawful FIDs: transition to Medicaid payor status; Alzheimer's or other dementia diagnoses; behavioral symptoms; and impairments that require more staff time (e.g., ES-1 severe functional dependencies and bowel or urinary incontinence). This study was designed to increase understanding of FIDs and their impacts on nursing home residents. It identifies resident characteristics related to increased risk of live discharges and how these relationships vary across time, states, and facility types. It also presents findings on the relationship between live discharges of residents with risk factors and post-discharge outcomes. ES.2 Methods We could not observe FIDs directly with available data. Instead, this study focused on live discharges of nursing facility residents with characteristics that were found to be indicative of FIDs. This study defines both risk factors for FIDs, which were suggested in the literature and empirically supported in this study, and general risk factors, which were found in our study to be associated with higher rates of live discharges. We refer to both collectively as risk factors for live discharge hereafter. Our study population includes residents receiving either nursing facility or Skilled Nursing Facility (SNF) care in United States nursing homes, nationwide, from 2012 to 2017. We primarily studied nursing facility residents as there were no identifiable risk factors among SNF residents. First, we provided empirical evidence for risk factors by comparing the prevalence of resident characteristics, such as behavior, health and Medicaid eligibility, between residents discharged live? and not discharged, using the Minimum Data Set (MDS) resident assessments linked with Medicare enrollment data and claims from the Master Beneficiary Summary File (MBSF). Resident characteristics were considered risk factors if they were conditions customarily required to be treated in a nursing home and were found in our research to be more prevalent among residents discharged live than residents not discharged. Next, we investigated risk factors among nursing facility residents further by analyzing their prevalence across years, states and facility types. Lastly, we compared post- discharge outcomes, primarily hospitalizations and outpatient acute care visits, of nursing facility residents discharged live with and without risk factors. We did not find any resident characteristics among SNF residents that were more prevalent among residents discharged live than those not discharged. Thus, we did not empirically identify any risk factors for live discharge among SNF residents or conduct any further analyses with SNF residents. The lack of empirical findings among SNF residents is likely because of the nature of short-term SNF care, which has the primary goal of rehabilitating and returning resident to the community, resulting in an overwhelming large number of appropriate live discharges. 1 We considered residents to be discharged live if: (1) they were discharged live according to their discharge assessment; and (2) they did not return to the same facility within 30 days. ES-2 ES.3 Results Risk Factors for Live Discharges among Nursing Facility Residents. Behavioral symptoms, impairments requiring more staff time, and transitions to Medicaid eligibility were found to be more prevalent in residents discharged live than those not discharged. However, impairments were only more prevalent among residents discharged live when increasing in severity. These risk factors were also identified in the literature as risk factors for FIDs. New diagnoses of psychiatric and mood disorders (e.g., schizophrenia and manic depression) were not previously identified in the literature, but were also more prevalent in residents discharged live than those not discharged. The same risk factors for live discharge were identified across years. Risk Factors for Live Discharges by State and Facility Type. For each risk factor identified at the national-level, there was some variation across states - for example, in several states the prevalence of risk factors among residents discharged live was lower than among residents who were not discharged. We also found that the differences in the prevalence of risk factors between residents discharged live and those not discharged varied by facility type. Compared to non-profit facilities, for-profit and government facilities had a higher prevalence of risk factors among residents discharged live, relative to residents remaining in the facility. Likewise, rural facilities and facilities that were part of a chain had a higher relative prevalence of risk factors among residents discharged live than urban and non- chain facilities, respectively. Lastly, residents discharged live from terminated facilities had lower prevalence of risk factors than residents discharged live from active facilities. Post-discharge Outcomes for Nursing Facility Residents. We observed high rates of acute care use, defined as hospitalizations, emergency department visits and observation stays, and mortality among residents discharged live from nursing facilities. Overall, 53% of all residents discharged live experienced acute care, and 36% of all residents died within 30 days of nursing facility discharge. However, rates of acute care use were higher among residents discharged live with risk factors for live discharge than residents discharged live without these risk factors. Furthermore, we found the majority of residents discharged live went directly from the nursing facility to the hospital or emergency department, with a smaller percentage of them going briefly to the community or another non-acute care setting before using acute care. This pattern was most pronounced among residents discharged live with risk factors for live discharge. ES.4 Discussion and Conclusion This study identified several resident characteristics associated with a higher risk of live discharges from nursing facilities. Behavioral symptoms, impairments requiring more staff time, transitions to Medicaid eligibility and psychiatric and mood disorders all increased the risk of live discharge, especially when these conditions newly developed or worsened in the past year. Our study found that the prevalence of these risk factors for live discharge varied ES-3 across states and facility types. For-profit, government and chain facilities were more likely to discharge residents with risk factors than non-profit and non-chain facilities. Lastly, we also found that residents discharged live with these risk factors, particularly severe behavioral symptoms and impairments requiring more staff time, experienced higher rates of acute care than residents discharged live without these risk factors, partially because these risk factors are associated with worse health overall. We note that although the live discharges of residents with the risk factors identified in this research may be legal, our goal was to identify patterns in live discharges that indicate higher risk of FIDs in order to gain insight into the prevalence and appropriateness of FIDs. As such, these findings are intended to serve as a foundation for continued discussion and analyses and to inform strategies and policies for reducing potentially inappropriate FIDs. Further research would be needed to discern unlawful FIDs from legal discharges. Additional recommendations for future research include examining the discharge locations of residents and the impact of state policies, facility closures, and the COVID-19 pandemic on FIDS. ES-4 1. Background and Purpose People can be discharged from nursing homes for many reasons. Discharges may be a positive outcome and reflect an individual's choice. In some cases, however, discharges may be at the direction of the facility and potentially involuntary and inappropriate. When FIDs from nursing homes are involuntary and against the will of the resident, the discharge may be unsafe and traumatic for the residents involved and may result in higher costs of care, raising concerns to the public and to Medicare and Medicaid programs (CMS, 2017a). FIDs continue to be one of the most frequent complaints made to State Long-Term Care Ombudsman Programs, with over 14,000 complaints about improper eviction or inadequate discharge planning in 2017 (Roberts, 2019). Furthermore, FIDs may be increasing in frequency in recent years (Verdana & Pear, 2018). There is also concern that the COVID-19 pandemic has exacerbated FIDs, with over 6,400 involuntary discharges during the pandemic reported as of the end of June 2020 (Silver- Greenberg & Harris, 2020). One state reported a 30% increase in complaints of evictions from March 2020 to July 2020 of the pandemic (Serres, 2020). This has led the Task Force on Aging and Families to call for a prohibition on evictions from nursing homes during the pandemic (Task Force on Aging and Families, 2020). FIDs have also caught the attention of federal courts, with an appeals court recently ruling that federal law requires California to act against nursing homes who discharge patients to hospitals and then inappropriately refuse to readmit them (Egelko, 2019). The OIG is investigating the extent to which state survey agencies investigated and took enforcement actions against nursing homes for inappropriate involuntary transfers and discharges, using complaint data from the National Ombudsman Reporting System from 2011 through 2016 (OIG, 2019). The Centers for Medicare & Medicaid Services (CMS) has also established an initiative to prevent FIDs that violate federal regulations (CMS, 2017a). Under federal regulations, all nursing home residents have a right to remain in the facility and can only be legally discharged involuntary for a few reasons. Legal FIDs are defined as those triggered by one of six reasons: the FID is necessary for the resident's welfare and the resident's needs cannot be met in the facility; the resident no longer needs nursing home care; the safety of other residents is endangered due to the clinical or behavioral status of the resident; the health of others in the facility would be endangered; the resident fails to pay; or, the facility closes (CMS, 2017a). FIDs not triggered by these six reasons are deemed unlawful. In practice, however, it is often difficult to identify inappropriate or unlawful FIDs. The National Ombudsman Reporting System only collects data on improper evictions in cases of complaints, and no data on evictions is collected through nursing home assessments or CMS's Online Survey Certification and Reporting System. As a result, there is a lack of systematic studies to assess the nature, prevalence, trend, and contributing factors of FIDs at the national level. However, a recent review of prior research (Lepore, Zepeda & Yuen, 2018) identified several characteristics that place residents at higher risk of potentially unlawful FIDs, including: transition to Medicaid payor status; Alzheimer's or other dementia diagnoses; behavioral symptoms; and impairments that require more staff time (e.g., severe functional dependencies, or bowel or urinary incontinence). Building on and extending previous this work (Lepore, Zepeda, & Yuen 2018), we had two objectives in this study to help further understanding of FIDs and their impacts on nursing home residents. First, we sought to identify resident characteristics related to higher risk of live discharge and determine how these relationships vary across time, states, and facility types. Second, we sought to determine how live discharges of residents with risk factors relate to resident outcomes post-discharge. To accomplish these objectives, we addressed the following three primary research questions: 1. Nationally, what are the characteristics of residents who are discharged live from nursing facilities, compared to residents who remain in the facility? What percentage of residents who are discharged live demonstrate risk factors for live discharge, relative to the percentage among residents who remain in the facility? How have such risk factors for live discharge changed over time? 2. What are the differences in characteristics among residents discharged live across all states and types of facility (e.g., for-profit vs. non-profit; chain vs. non-chain)? 3. Do post-discharge outcomes (e.g., risk of death, worsening health condition leading to hospitalization or emergency department visits) differ among residents discharged live with and without risk factors for live discharge? The methods used to address these research questions are summarized in Section 2, and a full discussion is presented in Appendix B. The analyses focus on live discharges of residents with risk factors for live discharges, as the available national data do not support direct observation of FIDs. While such discharges are not perfect approximations of FIDs, they provide insight into the patterns of FIDs and outcomes of residents experiencing FIDs. 2. Methods 2.1 Data Sources This research used secondary data from multiple sources. The MDS Version 3.0 provided longitudinal information on nursing home residents' demographic and health characteristics. The MBSF provided information about Medicaid eligibility status for each resident who was a Medicare beneficiary, mortality data, and data on hospitalization and outpatient acute care use from fee-for-service (FFS) Medicare claims. Facility-level data came from the Provider of Services (POS) files, which contain information about facility characteristics. A single, resident-level analytic file including all variables of interest was created by merging stay- level MDS data with corresponding MBSF data for each resident and POS data for the facility at which the stay occurred. Data from calendar years (CYs) 2012-2017 were used in this analysis. 2.2 Study Population The study population included all nursing home residents receiving either nursing facility or SNF care in United States nursing homes from 2012 to 2017. Each year was examined separately and included residents with at least one stay during the year. Details about the definition of nursing facility and SNF care and stay construction can be found in Appendix A. 2.3 Statistical Analyses Identifying FIDs. The available data sources do not contain information that identifies whether a discharge is facility-initiated, or whether it is voluntary or involuntary. Such information would only be attained through medical record review. Because of the lack of observed data on actual FIDs, our analyses relied on identifying discharges with resident characteristics that are conditions which a nursing facility is required to treat under federal regulations and which are indicative of FIDs. Prior to beginning the statistical analysis, the project team conducted a thorough literature review to understand what factors were likely to be associated with FIDs. The following are a list of factors suggested in the literature as associated with FIDs: Transition to Medicaid payor status. Alzheimer's or other dementia diagnoses. Behavioral symptoms. Impairments requiring more staff time (e.g., severe functional dependencies, bowel or urinary incontinence, or cognitive impairment). The team also included additional resident characteristics as general risk factors (i.e., psychiatric and mood disorders) not identified in the literature but which were found in our study to be associated with higher rates of live discharges. We refer to both collectively as risk factors for live discharge hereafter. Identifying Risk Factors. We compared the characteristics of residents discharged live and residents not discharged, nationwide, in 2017. First, we compared the characteristics at end of stay, defined as the most recent status according to either the discharge assessment or the last available assessment (in the case of residents not discharged). We also compared changes, either onset or worsening, in resident characteristics over the past year. These changes may indicate an improper discharge if the nursing home should be able to meet the changing care needs (e.g., if a health condition appears which is customarily required to be treated in a nursing home). We defined risk factors for live discharges as characteristics which were customarily required to be treated in nursing homes and were more prevalent among residents discharged live than residents not discharged. Below is a list of risk factors for live discharge and a list of additional characteristics examined in analyses which were not identified as risk factors because they were less prevalent among residents discharged live than residents not discharged. Asterisks indicate whether the risk factor was identified in literature, analyses, or both, with a single asterisk (*) indicating "identified in literature and analyses" and double asterisks (**) indicating "not identified in literature but identified in analyses." Risk Factors for Live Discharge: e Transition to Medicaid eligibility (over last 3 months, 6 months, or year). * e Severe behavioral symptoms at end of stay and changes in behavioral symptoms to severe over past year (physical, verbal, other disruptive, or when aggregated across behavioral types). * e Psychiatric and mood disorders at end of stay: manic depression (bipolar disorder) and post-traumatic stress disorder (PTSD). ** e New diagnoses of psychiatric and mood disorders over the past year: anxiety, depression, bipolar disorder, psychotic disorder (other than schizophrenia), schizophrenia, PTSD. ** e Total dependency in activities of daily living (ADLs) at end of stay and changes to total dependency over the past year. * Changes in urinary and bowel incontinence to severe over past year. * Increases in the severity of cognitive function over the past year. * Characteristics Examined but not Risk Factors for Live Discharge: e Full Medicaid eligibility at end of stay. * e Alzheimer's disease and non-Alzheimer's dementia, at end of stay and new diagnoses over the past year. * e Psychiatric and mood disorder at end of stay: anxiety, depression, psychotic disorder (other than schizophrenia) and schizophrenia. e Severe urinary and bowel incontinence at end of stay. * Mood distress, at end of stay and increases in the severity over the past year. Severe suicidal thoughts at end of stay and changes in suicidal thoughts to severe over the past year. e Dependencies for ADL (other than total dependency), at end of stay and increases in the severity over the past year. * e Moderate and severe cognitive impairment at end of stay. * e Significant change assessment in last 2 months. Because we did not find any characteristics among SNF residents that were more prevalent among residents discharged live than residents not discharged, we did not empirically identify any risk factors for live discharges among SNF residents. All subsequent analyses were limited to nursing facility residents. Analyses of risk factors and outcomes. We determined variation in risk factors for live discharge across time by comparing the differences in the prevalence of risk factors among residents discharged live and among residents not discharged, from 2012 to 2017. We also compared the differences in the prevalence of risk factors across states and facility types in 2017. We focused on 2017 as the latest year of data as risk factors were consistent across years. Lastly, we analyzed differences in the post-discharge outcomes of residents with and without risk factors for live discharge by comparing their hospitalization and outpatient acute care rates. This analysis focuses on 2016 as the latest year of data with sufficient subsequent data for identifying outcomes. Appendix B provides a more comprehensive description of our study methods. 3. Findings 3.1 Nationally, what are the characteristics of residents who are discharged live from nursing facilities, compared to residents who remain in the facility? What percentage of residents who are discharged live demonstrate risk factors for live discharge relative to the percentage among residents who remain in the facility? How have such risk factors for live discharges changed over time? 3.1.1 Characteristics of Nursing Facility Residents in 2017 Of the nursing facility population (1,345,105 residents) in 2017, 16.62% of residents? (223,522) were discharged live at some point during the year, and 65.85% of residents (885,788) were not discharged by the end of 20173 (Appendix C, Table C-1). We found several characteristics to be more prevalent among residents discharged live from nursing facilities than among residents remaining in facility. These characteristics include: (a) severe behavioral symptoms; (b) psychiatric and mood disorders; (c) impairments requiring more staff time; and (d) transitions to Medicare eligibility. Each characteristic was examined two ways: (1) at end of stay, defined as the most recent status recorded on either the discharge assessment or the latest assessment; and (2) as changes over the past year. Severe Behavioral Symptoms Key Takeaway e Severe behavioral symptoms at end of stay were more prevalent among residents discharged live than not discharged, for each behavioral symptom type individually (verbal, physical or other disruptive) and when aggregated across behavioral types. Changes to severe behavioral symptoms over the past year were also more prevalent. Severe behavioral symptoms at the end of stay, were more prevalent among residents discharged live than among residents not discharged. This pattern was consistent across different types of behavioral symptoms, including physical behavior, verbal behavior, and other disruptive behaviors (e.g., disrobing in public or throwing food). The difference in prevalence was largest for severe verbal behavior, with a prevalence of 8.61% among residents discharged live, compared to 7.26% among residents not discharged (Appendix C, Figure C-1). Severe and very severe behavioral symptoms aggregated across behavior types were similarly more prevalent among residents discharged live. In 2In 2017, the vast majority (95.55%) of the residents had only one stay. To ease the presentation of results below, we simply refer to stays as residents. 3 The rest of residents died during the year while in a nursing facility and were not included in the analysis. addition, changes to more severe behavioral symptoms over the past year were more frequent among residents discharged live. As shown in Figure 3-1, this pattern was seen for all behavioral types and for behavioral symptoms aggregated across behavioral types. Overall, the differences in prevalence between residents discharged live and those not discharged were more pronounced when observing increases in the severity of behavioral symptoms over the past year than when looking only at the presence of these behavioral symptoms at the end of stay, without regard to when these symptoms first developed. These findings were statistically significant (p<0.001). FIGURE 3-1. Change in Behavioral Symptoms over the Past Year by Discharge Status T% 6% 5% 4% L 5 3% So 14 0 2 2% 2 1% Cc 5 ia 5 0% oO Physical Verbal Other Behavior, Aggregate, Aggregate, Behavior, Behavior, to Severe*** to Severe*** to Very to Severe*** to Severe*** Severe*** Not discharged Discharged live */**/*** = Difference is statistically significant based on a p-value cutoff of 0.05/0.01/0.001. Psychiatric and Mood Disorders Key Takeaway e Only manic depression and PTSD at end of stay were more prevalent among residents discharged live than those not discharged. New diagnoses of all psychiatric and mood disorders over the past year were more prevalent among residents discharged live. Of all psychiatric and mood disorders examined at end of stay, only manic depression (bipolar disorder) and PTSD were more prevalent among residents discharged live than those not discharged. Of residents discharged live, 7.28% had a bipolar disorder and 1.16% had PTSD at end of stay, compared to 6.37% and 0.84% of residents not discharged, respectively (Appendix C, Table C-1). However, for all psychiatric and mood disorders, new diagnoses in the past year were more prevalent among residents discharged live than those not discharged (Table 3-1). These findings were statistically significant (p<0.001). Table 3-1. Percent of Residents with New Diagnoses of Psychiatric and Mood Disorders in the Past Year by Discharge Status Discharge New Diagnosis in Past Year, % Status Bipolar*** | PTSD*** | Anxiety*** | Depression* ** | Psychotic* ** | Schizophrenia* * * Not 1.10 0.16 6.52 7.78 2.74 1.32 discharged ischarged 1.77 0.28 8.04 9.15 2.88 1.62 ive */**/*** = Difference is statistically significant based on a p-value cutoff of 0.05/0.01/0.001. Impairments Requiring More Staff Time Key Takeaway e Impairments requiring more staff time were more prevalent among residents discharged live than those not discharged, but only when these impairments increased in severity over the past year. For ADLs, only the most severe level of dependency was more prevalent among residents discharged live. Impairments requiring more staff time such as urinary or bowel incontinence, cognitive impairments and functional dependencies were more prevalent among residents discharged live than those not discharged, but only when these impairments increased in severity over the past year. As shown in Figure 3-2, severe urinary or bowel incontinence at end of stay was less prevalent among residents discharged live than those not discharged, while increases in the severity of urinary or bowel incontinence from none or mild to severe were slightly more prevalent among residents discharged live (p<0.001). Likewise, moderate and severe cognitive impairments at the end of stay were less prevalent among residents discharged live than those not discharged, but there were greater increases in the severity of cognitive impairments over the past year among residents discharged live than those not discharged (p<0.001) (Appendix C, Table C-1). 3-3 Figure 3-1. Incontinence at End of NF Stay and Change over the Past Year by Discharge Status 70% 60% 50% 2 40% Cc S 30% wn YY 20% 5 LC : | Oo 0% L Severe Urinary Severe Bowel Urinary Bowel Incontinence, Incontinence*** Incontinence*** Incontinence, Change to Change to Severe*** Severe*** GNot discharged wmDischarged live */**/*** = Difference is statistically significant based on a p-value cutoff of 0.05/0.01/0.001. On average, residents who were discharged live were less dependent in ADLs than those who were not discharged. However, the most severe level of ADL dependency (total dependency on or rare or no occurrence of each ADL) was more prevalent among residents discharged live than those not discharged (5.26% compared to 4.82%). Changes to the most severe level of dependency were also more frequent among live discharges: 2.48% of residents discharged live had declined to the most severe level of ADL dependency over the past year, compared to 1.81% of residents not discharged. These findings were statistically significant (p<0.001). (Appendix C, Figure C-2). Transitions to Medicare Eligibility Key Takeaway e Transition to Medicaid eligibility over the 3 months and 6 months prior to discharge was more prevalent among residents discharged live than those not discharged. Lastly, transitions to Medicaid eligibility, but not Medicaid eligibility at end of stay, were more prevalent among residents discharged live than residents not discharged. As Medicaid is the primary payor for nursing facility care, 73.51% of residents who were discharged live had Medicaid eligibility compared to 84.90% of residents not discharged (Appendix C, Table C-1). However, a higher percent of residents discharged live than residents not discharged had transitioned to Medicaid eligibility over both the 3 months and 6 months prior to discharge, as shown in Figure 3-3. These findings were statistically significant (p<0.001). 3-4 Figure 3-2. Transitions to Medicaid Eligibility by Discharge Status 9% 8% 7% 6% 5% Percent of Residents SS&s x x 1% 0% Change in last 3 months Change in last 6 months HNot discharged Discharged live */**/*** = Difference is statistically significant based on a p-value cutoff of 0.05/0.01/0.001. Summary Key Takeaways e Several characteristics found in the literature to be risk factors for FIDs were not identified as risk factors for live discharge in this study. e The characteristics found in this study to be risk factors for live discharge include behavioral symptoms, psychiatric and mood disorders, impairments requiring more staff time, and transitions to Medicaid eligibility. We also evaluated the following risk factors based on findings from the literature: Alzheimer's disease and other dementias, suicidal thoughts, mood distress, and having a significant change assessment in the last 2 months. However, each of these characteristics had a lower prevalence among residents discharged live than among residents not discharged and thus were not considered further. Based on these analyses, the following characteristics meet the definition of risk factors for live discharge set forth in Section 2: behavioral symptoms, psychiatric and mood disorders, impairments requiring more staff time, and transitions to Medicaid eligibility. These risk factors were the focus of subsequent analyses. The analysis results for the full list of characteristics examined and additional years are included in Appendix C. 3-5 3.1.2 Characteristics of Residents Discharged Live from Nursing Facilities Over Time Key Takeaways e The number of residents discharged live was stable over time. The same risk factors were observed across years. e The difference in prevalence between residents discharged live and not discharged widened for severe behavioral symptoms and most psychiatric and mood disorders and was stable for other risk factors The number of residents discharged live from nursing facility care remained largely stable over time, from 16.27% (223,182) in 2012 to 16.62% (223,522) in 2017. The number of residents who were not discharged and remained in nursing facility care at the end of the year decreased from 67.18% (921,487) in 2012 to 65.85% (885,788) in 2017 (Appendix C, Table C-1).4 The majority of risk factors for live discharge decreased in prevalence or remained largely stable across years for both residents discharged live and those not discharged. There were some exceptions which increased in prevalence for both groups, including transitions to Medicaid eligibility, select psychiatric and mood disorders, and bowel incontinence. The differences in the prevalence of risk factors between residents discharged live and those not discharged were compared across years. As shown in Figure 3-4, severe behavioral symptoms and changes to more severe behavioral symptoms were more prevalent among residents discharged live across years (as indicated by positive differences in prevalence). The exception was severe behavioral symptoms, "other", at end of stay, which was only more prevalent among residents discharged live after 2015, as indicated by negative differences prior to then. The differences in the prevalence of most behavior symptoms between residents discharged live and those not discharged widened. However, these trends were only significant for risk factors related to verbal behavior (p<0.05) and other disruptive behavior (p<0.01). Only very severe behavior symptoms aggregated across behavioral types had a narrowing difference in prevalence (p>0.05). 4 The rest of the residents died during the year while in a nursing facility. 3-6 Figure 3-3. Differences in the Prevalence of Behavioral Symptoms between Residents Discharged Live and Not Discharged 2% Severe Verbal Behavior* Jo _ 1% Physical =~ Behavior -- Very Severe Behavior, Aggregate 0% =e Severe Behavior, Aggregate Percent difference between residents discharged live and not dischaged === Severe Other -1% Behavior** 2012 2013 2014 2015 2016 2017 2012 2013 2014 2015 2016 2017 Symptom Present at End of Stay Symptom Developed in the Past Year * /**/*** = Difference is statistically significant based on a p-value cutoff of 0.05/0.01/0.001. Consistent with findings in Section 3.1.1, only bipolar disorder and PTSD at the end of stay were more prevalent among residents discharged live than those not discharged across years. The differences in the prevalence of bipolar disorder and PTSD at end of stay widened over time (p>0.05 and p<0.01, respectively) (Appendix C, Figure C-3). Additionally, as shown in Figure 3-5, new diagnoses of the majority of psychiatric and mood disorders examined were more prevalent among residents discharged live than those not discharged across years. The exception was new diagnoses of psychotic disorders, which were less prevalent among residents discharged live for some years. Overall, the differences in the prevalence of new diagnoses of psychiatric and mood disorders widened from 2012 to 2017, with only the trend in PTSD being significant (p<0.01). The difference in the prevalence of depression narrowed (p<0.05) Also consistent with findings in Section 3.1.1, urinary or bowel incontinence at the end of nursing facility stay was less prevalent among residents discharged live than those not discharged across years. Changes to severe urinary or bowel incontinence were more prevalent across years, and the differences in the prevalence of changes to severe urinary and bowel incontinence were relatively stable across time (p>0.05) (Appendix C, Figure C-4), 3-7 Figure 3-4. Differences in New Diagnoses of Psychiatric and Mood Disorder 2.50% 2.00% 1.50% === Depression* em Anxiety 1.00% Bipolar Disorder 0.50% e=== Schizophrenia === PTSD*** 0.00% === Psychotic Disorder Percent difference between residents discharged live and not discharged -0.50% -1.00% 2012 2013 2014 2015 2016 2017 * /**/*** = Difference is statistically significant based on a p-value cutoff of 0.05/0.01/0.001. The most severe level of dependency for ADL and a change to the most severe level of dependency was more prevalent among residents discharged live than residents not discharged across years. The differences in the prevalence of the most severe level of dependency at the end of stay and changes to most severe level of dependency narrowed across years (p<0.05 and p<0.01, respectively). (Appendix C, Figure C-5),. The difference in the prevalence of increases in the severity of cognitive impairment was largely stable across years (p>0.05) (Appendix C, Table C-3). Lastly, transitions to Medicaid eligibility in either the last 3 months or 6 months was more prevalent among residents discharged live than those not discharged for all years (Figure 3-6). The differences in the prevalence of transitions to Medicaid eligibility over the last 3 months and 6 months were largely stable over time (p>0.05). Full results can be found in Appendix C. 3-8 Figure 3-5. Differences in Transitions to Medicaid Eligibility 4.00% 2 To Cc s © 3.50% SS H © 2S 3.00% c 2 ov g 6 2.50% SC o 2 E 2.00% oO ® D2 1.50% Lo = Oo 2 5 1.00% =_---_ ns 95 5 2 0.50% o. To 0.00% 2012 2013 2014 2015 2016 2017 --Transition to Medicaid Eligibility in Last 6 Months =-- Transition to Medicaid Eligibility in Last 3 Months Trend across years not statistically significant (p>0.05). 3.2 What Are the Differences in Characteristics Among Residents Discharged Live Across All States and Types of Facility? 3.2.1 Differences Across States Key Takeaways e Risk factors for live discharge varied across states. e For example, severe physical behavior at the end of stay places residents at a higher risk of discharge nationally and in 30 states, with a concentration in the southern states; however, in 20 states and the District of Columbia (D.C.), severe physical behavior at end of stay places residents at lower risk of live discharge. We found large variation in risk factors for live discharge across states. As an example, Figure 3-7 below shows the difference in the prevalence of one risk factor between residents discharged live and those not discharged across states. Severe physical behavior at end of stay is more prevalent among residents discharged live than those not discharged at the national level and also in 30 states (with clusters in the Southwest and the lower Southeast), as indicated by positive differences in prevalence, suggesting physical behavior is a risk factor for live discharge in these states. It is less prevalent among residents discharged live than those not discharged in 20 states and D.C. (with a slight cluster in the upper Midwest), as indicated by the negative differences in prevalence. For example, as shown in Figure 3 7, in Oklahoma a resident with severe physical behavioral issues would be 3-9 more likely to be discharged live than a resident without severe physical behavioral issues, whereas in Vermont a resident with severe physical behavioral issues would be less likely to be discharged live. Figure 3-6. Percentage Point Differences in the Prevalence of Physical Behavior at End of Stay, by State 32 peasas 5 4 -S700> Zz o ®NS_50 ® > $62223 J 29 z z ze S5_zes=85853° oO oO OL afd ox oP 3 2 of 3 xe2 -nG25 805 o< eo 5 SOUS TES SSaS5RB® S20 © N =O. - 8 = ODaSC AO GSO SxS TO 5 5 6 COMfS3QLP SOAS BS ® aw 4 CBIOoI=DVeOW =O GIONS OO 5 eee HI oe poder' e Hdl gs ro Qeenes - ID Caaa =. 58 ogsseeeee eee 233 5 <S32989533 SI=D+DO we Oo => aIVFO2® BO BOSH © oO PSSS§ox<S505 S gu Ota oD Dare sOSOAY < OF 02g CD 3 IBeps-gs859 5 3 OS ES S8gs "SS S35 5 £8 oOo 9 2 a5 o » » 3 co Wo 53 5° oa oS 3 5 2. oO 5 S For all the risk factors, some states showed negative differences in prevalence, indicating that in those states, residents with those risk factors would be less likely to be discharged live than residents without those risk factors. Some risk factors were more consistent than others. For example, the differences in the prevalence of new diagnoses of anxiety were positive in all but four states. Additionally, characteristics with negative differences in prevalence at the national level had positive differences in prevalence in many states. For example, anxiety at end of stay had positive difference in prevalence in 13 states. Geographic clusters in presence or absence of risk factors varied by the risk factor examined. Full results can be found in Tables D-1 through D-3. 3-10 3.2.2 Differences by Facility Types Key Takeaways e For-profit facilities and government facilities were more likely than non-profit facilities to discharge live residents with risk factors, for the majority of risk factors. Non-profit facilities were only most likely to discharge residents with very severe behavioral symptoms aggregated across behavioral types. e Rural facilities were more likely to discharge live residents with risk factors than urban facilities, for the majority of risk factors. Urban facilities were only more likely to discharge residents with increases in the severity of cognitive impairments. e Facilities that were part of a chain where more likely to discharge live residents with risk factors than facilities that were not part of a chain, for the majority of risk factors. Non-chain facilities were only most likely to discharge residents with impairments requiring more staff time. e Mid-size and small facilities were more likely to discharge live residents with risk factors than large facilities for the majority of risk factors. Large facilities were only most likely to discharge residents with select psychiatric and mood disorders. e Facilities with two or more ownership changes or one ownership change were more likely than facilities with no ownership changes to discharge residents with risk factors, for most risk factors. Facilities with no ownership changes were only most likely to discharge residents with select psychiatric and mood disorders. Most risk factors were present across all facility types, as evidenced by positive differences in prevalence between residents discharged live and those not discharged. However, some facility types had larger differences in the prevalence of certain risk factors between resident discharged live and those not discharged, indicating these facility types were more likely to discharge residents with those risk factors. Profit status. For-profit and government facilities were more likely than non-profit facilities to discharge residents with risk factors, as indicated by having larger differences in the prevalence of risk factors between residents discharged live and those not discharged. Out of the three profit statuses, for-profit facilities were most likely to discharge live residents with most behavior-related risk factors. Government facilities were most likely out of the three profit statuses to discharge live residents with most psychiatric and mood disorders- related risk factors and with changes in impairments requiring more staff time (incontinence, functional dependency and cognitive function). However, government facilities had negative differences in the prevalence of several types of behavioral symptoms, at end of stay, indicating they were less likely to discharge live residents with these severe behavioral symptoms than those without these severe behavioral symptoms. Non-profit facilities were only most likely of the three profit statuses to discharge live residents for a few risk factors, primarily those related to very severe behavioral symptoms 3-11 aggregated across behavioral types. Lastly, as shown in Table 3-2, government facilities were most likely to discharge live residents with transitions to Medicaid eligibility in the last 3 months, while for-profit facilities were most likely to discharge live residents with transitions to Medicaid eligibility in the last 6 months (as indicated by the largest differences in prevalence). (Appendix D, Table D-4). Table 3-2. Percentage Point Differences in the Prevalence of Transitions to Medicaid Eligibility by Facility Profit Status For-profit | Non-profit | Government Transitions to Medicaid eligibility in last 3 months 1.12 0.87 1.35 Transitions to Medicaid eligibility in last 6 months 3.84 3.02 2.60 Urban/rural. Rural facilities were more likely than urban facilities to discharge residents with risk factors, displaying larger differences in prevalence of most risk factors. For example, the prevalence of severe verbal behaviors was 1.96 percentage points higher among residents discharged live than residents not discharged in rural facilities, compared to 1.27 percentage points higher in urban facilities. Urban facilities were only more likely than rural facilities to discharge live residents with transitions to Medicaid eligibility over the last 3 months and 6 months and new diagnoses of bipolar disorder and schizophrenia. Urban facilities had a negative difference in the prevalence of increases in the severity of cognitive impairments, indicating they were less likely to discharge live residents with cognitive impairments than without cognitive impairments. (Appendix D, Table D-5). Chain status. Facilities that were part of a chain where more likely to discharge live residents with risk factors than facilities that were not part of a chain. The exception was residents with impairments requiring more staff time, who were more likely to be discharged live in non-chain facilities. (Appendix D, Table D-6),. Facility size. Small facilities (less than 100 beds) and mid-size facilities (101-200 beds) were more likely than large facilities (over 200 beds) to discharge live residents with risk factors. Small facilities were most likely to discharge residents with behavior-related risk factors and new diagnoses of anxiety, psychotic disorders and PTSD, compared to larger facilities. Mid- size facilities were most likely to discharge live residents with transitions to Medicaid eligibility in the past 3 months and 6 months and impairments requiring more staff time, compared to smaller and larger facilities. Lastly, large facilities were more likely to discharge live residents with new diagnoses of depression, bipolar disorder and schizophrenia than smaller facilities. Large facilities also had some negative differences in the prevalence of risk factors, primarily among behavioral symptoms at end of stay. For example, as shown Figure 3-8, large facilities had a difference in prevalence of severe behavior, "other",° at the end of stay of -0.49 percentage points compared to 0.67 ° Severe behavior, "other", are disruptive behaviors outside of physical/verbal behavior (e.g., disrobing in public or throwing food) 3-12 percentage points among small facilities and 0.33 percentage points among mid-size facilities. This means that large facilities were less likely to discharge live residents with behavioral symptoms than residents without behavioral symptoms while mid-size and small facilities were more likely (Appendix D, Table D-7), Figure 3-7. Percentage Point Differences in the Prevalence of Severe Behavior, "Other", at End of Stay, by Facility Size 0.8 2 TO Cc ®o SP 06 non © 25 c 2 0.4 ov 2s 6 = 0.2 a0 o & oO 52 ° oO =D -0.2 oD <2 g 5 -0.4 OD o -0.6 Small Mid-size Large Ownership changes. Facilities with two or more ownership changes and 1-2 ownership changes were more likely to discharge live residents with risk factors than facilities with no ownership changes. Facilities with two or more ownership changes were most likely to discharge live residents with behavior-related risk factors and new diagnoses of anxiety, bipolar disorder and PTSD, compared to facilities with fewer ownership changes. Facilities with one to two ownership changes were most likely to discharge live residents with the remaining behavioral-related risk factors, and facilities with no ownership change were most likely to discharge live residents with new diagnoses of depression and schizophrenia, out of all ownership change counts. The association between number of ownership changes and likelihood of discharging live residents with impairments requiring more staff time depended on the impairment. (Appendix D, Table D-8). 3.2.3 Discharges from Terminated Facilities Key Takeaway e Only a small percent of residents were discharged live from terminated facilities. Residents discharged live from terminated facilities had lower prevalence of risk factors than residents discharged live from active facilities. 3-13 The nature of discharges from terminated facilities differs from that of active facilities. Nursing facility termination indicates a nursing facility has lost its Medicare and/or Medicaid certification, requiring it to discharge all Medicare and Medicare-Medicaid dually eligible residents. Nursing facilities may continue to care for a portion of residents after the scheduled termination date (e.g., if they are private pay or Medicaid only, depending on the certification lost). For the purposes of this study, live discharges from terminated facilities were considered unrelated to an individual's condition since they were discharged due to facility reasons and not for patient reasons. From 2012 to 2017, between 0.68% and 1.21% of residents were discharged live from a terminated facility each year. The majority of these residents, between 75% and 90%, were discharged live on or before the termination date. The remaining residents were discharged live after the termination date (Table D-9). Because terminated facilities discharge all Medicare residents, discharge of a resident with a risk factor does not indicate potential inappropriateness. Compared to residents discharged live from active facilities, residents discharged live from terminated facilities, either before or after the termination date, had lower prevalence for the majority of risk factors. For example, 9.79% of residents discharged live from terminated facilities had schizophrenia at the end of stay, compared to 12.81% of residents discharged live from active facilities (p<0.001). There were some exceptions, such as some types of behavioral symptoms at end of stay, which were more prevalent among residents discharged live from terminated facilities than residents discharged live from active facilities. Not all these results were statistically significant. The higher prevalence of most risk factors among residents discharged live may indicate potentially inappropriate live discharges among active facilities. However, there could also be other explanations such as differences in the populations served by terminated and active facilities, which are beyond the scope of this study (Appendix D, Table D-10). 3.3. Do Post-discharge Outcomes Differ Among Residents Discharged Live With and Without Risk Factors for Live Discharge? 3.3.1 Acute Care Use We define "acute care" as hospitalizations, emergency department visits and observations stays, with emergency department and observation stays collectively referred to as "outpatient acute care".® In 2016, 52.79% of residents experienced acute care within 30 days of nursing facility discharge. Specifically, 44.33% of residents were hospitalized, and 15.40% had an outpatient acute care visit within 30 days of discharge, with a small percentage having both a hospitalization and an unrelated emergency department visit within 30 days. The majority of acute care episodes happened immediately after nursing 6 Emergency department visits and observation stays are grouped as only a very small proportion of residents had observation stays and these stays are typically billed with emergency department visits. 3-14 facility discharge: 86.9% of hospitalizations and 56.6% of outpatient acute care visits occurred either on the day of nursing facility discharge or the following day. Overall, 43.13% of all residents discharged live went directly from the nursing facility to an acute care setting (hospital or emergency department), while 9.66% of residents went first to a community or other non-acute setting (such as a different nursing facility) and then experienced acute care within 2-30 days of discharge. The remaining residents experienced no acute care or acute care after 30 days. (Appendix E, Table E-1). Based on this study's definition of live discharge, residents discharged live did not return to their original nursing facility within 30 days of discharge. However, they may have been transferred to a second nursing facility, either directly or after their acute care ended. For more discussion on discharge location, see Section 3.3.3. In the remainder of this section, we describe the acute care use of residents with risk factors. One caveat is that we cannot determine the cause of acute care use, such as whether it was the result of an inappropriate discharge. Some of these risk factors, particularly severe behavioral symptoms and impairments requiring more staff time, may indicate declining overall health of a resident and increased the risk of acute care. Thus, higher acute care use in these populations partially reflects appropriate discharging of residents to hospitals when the nursing facilities can no longer meet their needs. Residents with Severe Behavioral Symptoms Key Takeaways e Residents discharged live with behavior-related risk factors had higher rates of acute care than residents discharged live without behavior-related risk factors, and differences in acute care rates were relatively large. e Most residents discharged live with behavior-related risk factors who experienced acute care went to acute care directly after discharge. Residents with behavioral risk factors had much higher rates of hospitalization within 30 days of nursing facility discharge, compared to residents without such factors. Between 54.40% and 61.38% of residents with severe behavioral symptoms at end of stay were hospitalized, with residents with severe verbal behavior having the lowest rates and residents with very severe behavior symptoms, aggregated across behavior types, having the highest. In contrast, residents without severe behavioral symptoms consistently had hospitalization rates around 43%. Similar differences were observed between the hospitalization rates of residents with and without changes to more severe behavioral symptoms. As an example, we use a representative graph to illustrate post-discharge hospitalization patterns by one behavior risk factor for live discharge. As shown in Figure 3-9, hospitalization rates increased with increasing severity of behavioral symptoms aggregated across behavioral types, at end of stay, and the majority of hospitalizations 3-15 happened directly after discharge. Specifically, as shown by the green line in Figure 3-9, 55.5% out of 61.38% of residents with very severe behavior symptoms aggregated across behavioral types, at end of stay, were hospitalized on the day of discharge or the following day (Days 0 and 1), compared to 5.88% out of 61.38% hospitalized between days 2 and 30 post-discharge (Days 2-30). We see a similar trend for residents with severe aggregated behavior symptoms (orange line in Figure 3-9) and none or mild behavioral aggregated symptoms (blue line in Figure 3-9), but the pattern is less pronounced as severity decreases. In general, among residents with behavioral symptoms at end of stay and with changes to severe behavior symptoms who were hospitalized within 30 days of nursing facility discharge, roughly 90% of them were hospitalized within one day of discharge, regardless of the type of behavioral symptoms. (Appendix E, Table E-2). These findings were statistically significant (p<0.001). Figure 3-8. Hospitalization of Residents within 30 Days of NF Discharge, by Behavioral Symptoms Aggregated across Behavioral Types, at End of Stay a HD ou o oa 1) yO Ww fs a oa oO Cumulative Percent of Residents 01234567 8 9 1011121314151617 18192021 222324 25 26 27 28 29 30 Days since Discharge =--- Mild or no behavior symptoms -- = Severe behavioral symptoms*** KKK -- Very severe behavioral symptoms */**/*** = Difference is statistically significant based on a p-value cutoff of 0.05/0.01/0.001. Residents with behavioral symptoms also had higher outpatient acute care use. For example, as shown in Figure 3-10, rates of outpatient acute care visits were highest for residents with severe and very severe behavior symptoms aggregated across behavior type, at end of stay, compared to residents with mild or no aggregated behavioral symptoms. Outpatient acute care use was highest immediately after discharge but was more evenly distributed across the 30 days post-discharge than hospitalizations. For example, 62% of all outpatient acute care use for residents with very severe behavioral symptoms aggregated across behavioral types, at end of stay, occurred on the same day as discharge or the following day, compared to almost 90% of hospitalizations. Similar differences in outpatient rates were observed between residents with and without severe behavioral symptoms by 3-16 individual behavioral type and with and without changes to more severe behavioral symptoms (Appendix E, Table E-2). These findings were statistically significant (p<0.001). Figure 3-9. Outpatient Acute Care Use of Residents within 30 Days of NF Discharge, by Behavioral Symptoms Aggregated across Behavioral Type, at End of Stay 25 n e 20 oO 3S & 15 -_- a ue _- oO = 10 Cc oO _" 2 e 5 0 01234567 8 9 101112131415161718192021 2223 24 25 26 27 282930 Days since Discharge -- Mild or no behavior symptoms -- Severe behavioral symptoms*** -- Very severe behavioral symptoms*** * /**/*** = Difference is statistically significant based on a p-value cutoff of 0.05/0.01/0.001. Residents with Psychotic and Mood Disorders Key Takeaway e Residents discharged live with most psychiatric and mood disorders had higher rates of acute care than residents discharged live without these disorders, and the majority of acute care happened directly after hospitalizations. However, the differences in acute care rates, particularly outpatient acute care, were relatively small. Residents with most psychotic and mood disorders had higher hospitalization rates than residents without these disorders. The exception was residents with PTSD at end of stay and new diagnoses of PTSD, who had lower rates of hospitalization than residents without PTSD. Table 3-3 shows the hospitalization rates among residents with and without psychotic and mood disorders. The largest positive difference (6.95%) was observed between residents with and without psychotic disorders at end of stay and the smallest positive difference (0.82%) was observed between residents with and without new diagnoses of depression. As in the case of residents with behavioral symptoms, the majority of hospitalizations happened on the day of discharge or the day following discharge. These findings were statistically significant (p<0.001). 3-17 Table 3-3. Hospitalization of Residents within 30 Days of NF Discharge, by Psychiatric and Mood Disorder Schizo- Bipolar PTSD Anxiety Depression Psychotic phrenia*** Diagnoses at 47.86 29.17 46.45 45.58 50.58 48.65 end of stay No diagnosis 44.07 44.46 43.34 43.04 43.63 43.92 at end of stay New 49.69 28.10 46.89 44.71 49.29 49.30 diagnosis No new 44,18 42.73 43.96 44.06 43.93 44,23 diagnosis */**/**X = Difference is statistically significant based on a p-value cutoff of 0.05/0.01/0.001. Residents with psychotic and mood disorders also had higher outpatient acute care use than residents without these diagnoses. For new diagnoses, outpatient acute care visits ranged from 13.69% among residents with PTSD to 18.11% among residents with bipolar disorders (Appendix E, Table E-2). In contrast, residents without a given new diagnosis mostly had an outpatient acute care use rate around 15%. This finding was only statistically significant for new diagnoses of anxiety (p<0.01), bipolar disorder (p<0.01) and psychotic disorders (p<0.05). As an example, we use a representative graph to illustrate post-discharge outpatient patterns by one disorder. As shown in Figure 3-11, the differences in outpatient acute care use rates between residents with and without psychiatric and mood disorders were small (less than 2% by day 30 post-discharge for most diagnoses). PTSD at the end of stay and new diagnoses of PTSD were unique in that the cumulative outpatient acute care rate for residents with PTSD was lower than that of residents without PTSD for more than two weeks post-discharge but similar or higher by the end of 30 days (p>0.05). 3-18 Figure 3-10. Outpatient Acute Care Use of Residents with and without New Diagnoses of Psychotic Disorder N oa NO oO oa Cumulative Percent of Resident oS 0 012345 67 8 9 1011121314151617 18192021 22 23 24 25 26 27 28 29 30 Days since Discharge =--With new diagnosis* Without new diagnosis */**/*** = Difference is statistically significant based on a p-value cutoff of 0.05/0.01/0.001. Residents with Impairments Requiring More Staff Time Key Takeaways e Residents discharged live with impairment-related risk factors had higher hospitalization rates than residents discharged live without these risk factors. Most hospitalizations happened directly after discharge. e Residents with changes to severe incontinence and cognitive impairments had higher outpatient acute care rates and residents with changes to the most severe level of ADL dependency had lower outpatient acute care rates, compared to residents without changes in impairments. Residents with incontinence at end of stay had higher hospitalization rates than residents without these impairments, with relatively large differences. As shown in Figure 3-12 below, hospitalization rates were much higher for residents with severe urinary incontinence at end of stay than for residents without urinary incontinence (51.60% compared to 32.84%, p<0.001). Most hospitalizations happened directly after discharge. Hospitalization rates for residents with severe bowel incontinence were similarly higher than residents without bowel incontinence (p<0.001). 3-19 Figure 3-11. Hospitalization of Residents with and without Severe Urinary Incontinence, at End of Stay a ao oO oOo 'KR ol [ IS oO Ww oO NO on rm Cumulative Percent of Residents NO oO oO o - oa 01234567 8 9 1011121314151617 18192021 2223 24 25 26 27 28 29 30 Days since Discharge e-_With incontinence* --Without incontinence */**/*** = Difference is statistically significant based on a p-value cutoff of 0.05/0.01/0.001. Residents with changes to severe impairments also had higher hospitalization rates than residents without changes to severe impairments (p<0.001). However, there were smaller differences in hospitalization rates between residents with and without changes to severe impairment than residents with and without impairments at end of stay, as shown in Table 3-4, Table 3-4. Hospitalization of Residents within 30 Days of NF Discharge, by Change to Severe Impairments, Percent Impairment Cognitive Incontinence Most Severe Level of nips Urinary*** Bowel* ** ADL Dependency*** With Change 51.30 52.08 58.39 54.45 Without Change 44.63 44.46 43.88 42.43 */**/*** = Difference is statistically significant based on a p-value cutoff of 0.05/0.01/0.001. In contrast, residents with urinary and bowel incontinence at end of stay and changes to the most severe level of dependency for ADL had lower 30-day outpatient acute care use rates than residents without these impairments (p<0.001). A representative graph (Figure 3-13) is shown below. 3-20 Figure 3-12. Outpatient Acute Care of Residents with and without a Change in ADL to the most Severe Level of Dependency 20 15 10 Cumlative Percent of Residents 0 012345 67 8 9 1011121314151617 18192021 2223 24 25 26 27 28 29 30 Days since Discharge =--With change in ADL* === Without change in ADL */**/*** = Difference is statistically significant based on a p-value cutoff of 0.05/0.01/0.001. Residents with changes to severe urinary and bowel continence and cognitive impairments had higher outpatient acute rates across all 30 days than residents without these changes in impairments (p<0.001) (Appendix E, Table E-3). Residents with Transition to Medicaid Eligibility Key Takeaway e Residents who with risk factors related to transition to Medicaid eligibility had higher hospitalization rates and lower outpatient acute care use than residents without these risk factors. Lastly, as shown in Table 3-5, residents who transitioned to Medicaid eligibility in the last 3 months and 6 months had higher hospitalization rates (p<0.001) but lower outpatient acute care use than residents who did not transition to Medicaid eligibility; only outpatient acute care findings for transition to Medicaid eligibility in the last 6 months were statistically significant (p<0.05). In contrast, residents with Medicaid eligibility at the end of stay had lower hospitalization rates but higher outpatient rates than residents who did not have Medicaid eligibility (p><0.001). 3-21 Table 3-5. Hospitalization and Outpatient Acute Care of Residents within 30 Days of NF Discharge, by Medicaid Eligibility Status Percent, % eMedicaid re etigibility in Last End of Stay 3 Months 6 Months Hospitalization |With Characteristic 47 .86*** 37.5 7*** 36,.36*** Without Characteristic 35.50 44.89 45.39 Outpatient With Characteristic 16.01*** 15.17 14.72* Acute Care -_| without Characteristic 13.93 15.52 15.56 */**/*** = Difference is statistically significant based on a p-value cutoff of 0.05/0.01/0.001. 3.3.2 Mortality Of all residents discharged live from nursing facilities, 35.68% died within 30 days. As with acute care use, mortality is not unexpected based on the age and health conditions of residents and does not necessarily reflect inappropriate discharges. A discharge to a hospital followed by death may be appropriate, for example, if a resident is experiencing a health crisis that cannot be treated in the nursing facility. Likewise, a resident might be discharged to the community followed by death because they prefer to receive hospice care or die in a home setting. As shown in the Figure 3-14, mortality rates varied by the timeframe of acute care use. Residents who had acute care on the day of discharge or the following day had a 30-day mortality rate of 59.42%, comprising 72% of all deaths within 30 days of residents discharged live. Residents with no acute care within 30 days had the next highest mortality rate at 19.94% and constituted 26% of all deaths within 30 days. Lastly, resident who were discharged live to community and experienced acute care 2-30 days after discharge had a mortality rate of 6.64%, making up only 2% of deaths within 30 days. For both residents who had immediate acute care and residents who had no acute care in 30 days, most deaths occurred in the days immediately following discharge. 15.1% of residents with immediate acute care and 5.7% of residents with no acute care in 30 days died within one day of discharge, equivalent to approximately a quarter of deaths in their respective groups. 45.5% of residents with immediate acute care and 16.1% of residents with no acute care died within 10 days, equivalent to over three-fourths of death in their respective groups. In contrast, the deaths were spread out over 30 days for residents who received acute care within 2-30 days of discharge. 3-22 Figure 3-13. Mortality by Acute Care Usage 70 60 40 30 20 10 /- 012345 67 8 9 1011121351415161718192021222324252627282930 Cumulative Percent of Residents Days since Discharge eee Acute care within 0-1 days No acute care within 30 days ==<_ Acute care within 2-30 days Difference between acute care usage groups statistically significant at the p=0.001 level. 3.3.3 Analysis of Discharge Location Key Takeaways e Behavior-related risk factors and psychiatric and mood disorders were most prevalent among residents discharged live to psychiatric hospitals. e Residents with changes in impairments were most prevalent among residents discharged to acute care, other nursing facilities and psychiatric hospitals. We examined MDS data to gain further insight into the discharge location of residents. We found that 44.29% of all residents were discharged live to acute care hospitals, which is similar to the 43.13% of hospitalizations found using claims data.' Of all residents discharged live, 34.24% were discharged live to community,® 17.06% to another nursing facility, 1.87% to a psychiatric hospital, and 2.54% to other locations such as an inpatient rehabilitation facility, intellectual disabilities/developmental disabilities (ID/DD) facility, hospice, and long-term care hospital. Severe behavioral symptoms at end of nursing facility stay and changes to more severe behavioral symptoms, for all behavioral types and when aggregated across behavioral ? This discrepancy is to be expected due to differences in methodology and acute care definition. The MDS discharge location is also considered to be less accurate than claims data and should be considered supplementary. 8 Community indicates places of permanent residence (e.g., private homes, assisted living facilities and group homes). 3-23 types, were most prevalent among nursing facility residents discharged live to psychiatric hospitals. For example, 54.3% of residents who were discharged live to a psychiatric hospital had severe verbal behavioral symptoms at the end of their nursing facility stay, compared to 10.3% or less among residents discharged live to any other settings. Likewise, psychiatric and mood disorders at end of nursing facility stay and new diagnoses of psychiatric and mood disorders were highest among nursing facility residents who were discharged live to psychiatric hospitals. Severe urinary or bowel incontinence at end of nursing facility stay was most prevalent among residents discharged live to an acute care hospital, followed by residents discharged live to another nursing facility. Changes to severe urinary incontinence was most prevalent among residents discharged live to psychiatric hospitals while changes to severe bowel incontinence was most prevalent among residents discharged live to acute care hospitals. Severe cognitive impairments were most prevalent among residents discharged live to acute care hospitals. Residents whose cognitive impairment score worsened by one point (on a three-point scale) had the highest prevalence among residents discharged live to acute care hospitals while residents whose cognitive impairment score worsened by two points had the highest prevalence among residents discharged live to psychiatric hospitals. Full results are presented in Table E-3. 3.3.4 Outcome of Residents Discharged Live from Terminated Facilities Key Takeaway e Residents discharged live from terminated facilities had lower acute care use and mortality than residents discharged live from active facilities. Residents discharged live from terminated facilities had lower acute care use and mortality than residents discharged live from active facilities. Specifically, 17.6% of residents discharged live from terminated facilities on or before the termination date and 24.4% discharged live after termination date? experienced acute care within 30 days of discharge, compared to 53.0% of residents discharged live from active facilities (Appendix E, Figure E-1), The higher acute care use rate among residents discharged live from active facilities may be partially attributed to their higher prevalence of risk factors associated with live discharges, relative to residents discharged live from terminated facilities. Residents discharged live from terminated facilities also have lower mortality rates than residents discharged live from active facilities. When comparing residents who had acute care on the day of discharge or the following day, 33.4% of residents from terminated facilities who 9 After losing Medicare and/or Medicaid certification, a facility may continue to care for private pay and/or Medicaid residents (depending on the certifications lost). Some of these residents discharged live later in the year, after termination date. 3-24 were discharged live on or before the termination date and 26.0% of residents who were discharged live after the termination date died within 30 days of discharge, compared to 60% of residents discharged live from active facilities. 3-25 4. Discussion 4.1 Summary of Findings 4.1.1 Identification of Risk Factors Our analysis provides empirical evidence for literature-suggested risk factors for FIDs and identifies several additional general risk factors for live discharge by comparing the prevalence of resident characteristics between residents discharged live and not discharged. We then investigated these risk factors further by analyzing their prevalence across states and facility types. Lastly, we examined the outcomes of residents discharged live from nursing facilities with and without risk factors. We found several risk factors that may increase the risk of live discharges. We observed that severe behavioral symptoms at the end of nursing facility stay may be a risk factor for live discharges. Verbal behavior appears to be the strongest risk factor for live discharges among all types of behavioral symptoms, as indicated by verbal behavior having the largest differences in prevalence between residents discharged live and residents not discharged. Severe and very severe behavior aggregated across behavioral types and changes to more severe behavioral symptoms over the past year may also increase the risk of live discharges. Although not indicated as potential risk factors for FIDs in the literature, our findings also observed that psychiatric and mood disorders may increase the likelihood of live discharge. Our results suggest that new diagnoses of psychiatric and mood disorders may be stronger risk factors for live discharges than pre-existing diagnoses. Our analyses show that impairments requiring more staff time such as urinary and bowel incontinence, severe functional dependencies and cognitive impairment are also associated with higher rates of live discharge but only when these conditions changed from a less severe to more severe state. For example, urinary and bowel incontinence in themselves may not be a risk factor for live discharge; the majority of residents have these conditions and most nursing facilities should be capable of managing these conditions as part of standard care. We observed a higher percentage of residents with Medicaid eligibility among residents not discharged than among residents discharged live. This could be because with residents with Medicaid eligibility are sicker overall, requiring nursing facility care, and many of them are being cared for by nursing facilities. However, we see a higher prevalence of residents transitioning to Medicaid eligibility in the last 3 months and 6 months among residents discharged live than among those not discharged. This indicates that transitioning to Medicaid eligibility may put residents at higher risk for live discharges. Additional characteristics such as Alzheimer's disease and other dementias were identified in the literature as risk factors for FIDs but not found in our analyses to be more prevalent among residents discharged live. We found the same risk factors for live discharge among nursing facility residents across years with few exceptions. We showed that the majority of differences in the prevalence of behavioral symptoms and new diagnoses of psychiatric and mood disorders between nursing facility residents discharged live and those not discharged widened from 2012 to 2017. The differences in the prevalence of other risk factors, such as changes to more severe Incontinence and transitions to Medicaid eligibility, were relatively stable over time. 4.1.2 Variation of Risk Factors Across State and Facility Types There was large variation across states in the differences in the prevalence of risk factors between residents discharged live and residents not discharged. For each risk factor, several states also showed lower prevalence among residents discharged live than residents not discharged, suggesting the particular resident characteristic is not a risk factor for live discharge in those states. Other states showed higher prevalence on characteristics such as severe incontinence at end of stay, which were not risk factors at the national level, suggesting additional characteristics may appear as risk factors for live discharges at the state level. We similarly found large variation across types of nursing facilities. For-profit and government facilities had larger differences in the prevalence of most risk factors between residents discharged live and those not discharged than non-profit facilities. Rural facilities and facilities that were part of a chain had larger differences in the prevalence of most risk factors than urban and non-chain facilities, respectively. For other facility characteristics, such as size and ownership changes, the relative size of the differences in prevalence varied by risk factor and facility type. Residents discharged live from terminated facilities had lower prevalence of most risk factors associated with live discharges, compared to residents in active facilities. Since terminated facilities must discharge all residents and active facilities only discharge some residents, this finding may indicate differences in discharge patterns based on resident characteristics in active facilities. 4.1.3, Outcomes of Nursing Facility Residents After Discharge We observed high rates of acute care use and mortality among residents discharged live. Acute care usage and mortality are to be expected in the nursing facility population due to the age and health conditions of residents. Overall, out of all residents discharged live, 52.79% of residents experienced acute care and 35.68% of residents died within 30 days of nursing facility discharge. 43.13% of all residents discharged live, equivalent to 82% of all residents experiencing acute care within 30 days, went directly from the nursing facility to 4-2 the hospital and had a mortality rate at 59.42% within 30 days. These rates partially reflect nursing facilities appropriately discharging residents to the hospital because they are having a health crisis and require higher levels of care. An additional 9.66% of all residents discharged live went to the community or a non-acute care setting followed by acute care within 2-30 days and had a 6.64% mortality rate within 30 days. We found that acute care use, both hospitalizations and outpatient acute care visits, was higher for residents discharged live with a risk factor associated with live discharges than residents discharged live without these risk factors. These patterns are to be expected as many risk factors are associated with declining health and increased risk of acute care. The differences in hospitalization rates between residents with and without severe behavioral symptoms and residents with and without changes to severe impairments were positive and relatively large, which suggests residents with these risk factors are likely to have negative outcomes post-discharge. Residents with and without new diagnoses of psychiatric and mood disorders had higher rates of hospitalization than residents without these disorders but differences in rates were relatively small. These results are not surprising as severe behavioral symptoms and changes to severe impairment are more likely to be indicative of health crisis than psychiatric and mood disorders, resulting in more appropriate discharges to hospitals. We found that the majority of hospitalizations of residents with risk factors happened on the day of discharge or the day following; this pattern was present but less pronounced among residents discharged live without risk factors. Furthermore, outpatient acute care use rates were also higher among residents with risk factors. However, the differences in outpatient acute care use rates between residents with and without risk factors were less pronounced than the differences in hospitalization rates. 4.1.4 Insights into FIDs Some discharges of residents displaying risk factors for live discharge may reflect nursing facilities discharging residents appropriately, either because the facilities cannot meet their increased care needs or because families and residents prefer to move. However, media has raised concern that some discharges may reflect nursing facilities avoiding the increased burden of meeting the care needs of these residents, which would be inappropriate (Bernard & Pears, 2018). Although we cannot directly define such live discharges as "FIDs" or determine whether they were appropriate, we nonetheless can use the risk factors for live discharge identified in this study to gain insights into the nature of such discharges. We explored how discharge patterns for certain conditions (or risk factors) could inform concerns that have been raised about live discharges. For example, we observed associations between live discharges and conditions that should be commonly treated in nursing facilities, such as severe verbal behavior or severe other disruptive behavior. We note that while there was an association between discharges and severe physical behavior, discharges associated with this risk factor may be appropriate in certain circumstances. For the same reason, live discharges of residents with psychiatric and mood disorders should be carefully evaluated. We also examined the discharge location of residents discharged live with risk factors. Discharge of a resident with a new diagnosis of a psychiatric or mood disorder or severe behavioral symptoms may be appropriate if the facility cannot meet the resident's new care needs, and the resident is discharged live to a psychiatric hospital or other settings more equipped to care for residents with these conditions. Our analyses of discharge location showed that both residents with new diagnoses of psychiatric or mood disorders and severe behavioral symptoms are most prevalent among residents discharged live to psychiatric hospitals. However, our analyses also showed that residents with psychiatric or mood disorders and severe behavioral symptoms are among those discharged live to community. Overall, 1.87% of all nursing facility residents who are discharged live are discharged to psychiatric hospitals, which is lower than the prevalence of new diagnoses of psychiatric or mood disorders and severe behavioral symptoms among residents discharged live. These findings raise some concerns related to live discharges of residents with these risk factors. One potential cause is the shortage of psychiatric beds, the number of which have decreased almost 97% as of 2016 from its highest levels in 1955 (Treatment Advocacy Center). Likewise, residents who recently transitioned to Medicaid eligibility may be more likely to experience live discharges from nursing facilities. Since facilities typically receive lower payments for Medicaid residents than for non-Medicaid residents, they may have a financial incentive to discharge Medicaid residents (Bernard & Pears, 2018). For example, a recent news article found that financial incentives may be influencing nursing home behavior: it discovered that Medicaid residents were less likely to be admitted to highly ranked nursing homes in New York and noted a resident who was denied a long-term bed by her nursing home after switching to Medicaid (Michel & McAndrew, 2019). Thus, we should pay particular attention to the discharges of these residents. We likewise acknowledge that differences in the prevalence of risk factors among states and facility types do not necessarily indicate certain states or facility types have more inappropriate discharges than others. Nursing facility discharges at the state level are impacted by a wide range of state-specific policies that may impact the likelihood of live discharges of residents with risk factors. Similarly, the appropriateness of discharges of residents with risk factors by a facility type may be determined by resources in the surrounding community. For example, the higher prevalence of discharges of residents with a psychiatric or mood disorder from facilities of a given type may be the result of that facility type having greater access to psychiatric hospitals. However, our insight into the association between discharges of residents with risk factors and states and facility types can suggest which discharges may warrant further inspection. For example, careful attention should be given to discharges of residents with certain risk factors in for-profit or 4-4 chain facilities. Furthermore, these findings may help target future investigations of live discharges that could possibly be indicative of inappropriate FIDs. State-specific analyses of live discharges may want to focus on the characteristics which appear as risk factors for live discharges in that state. Lastly, our analysis indicates that residents with risk factors consistently have higher rates of hospitalization and mortality post-discharge. These outcomes may either be the result of appropriate discharges or negative consequences of inappropriate discharges, with certain combinations of resident characteristics and outcomes being more problematic than others. For example, a discharge of a resident with severe impairments followed shortly by death may be appropriate if it reflects the patient's preference to die in a hospital or at home. In contrast, the discharge of someone with psychiatric and mood disorders to the community who requires acute care shortly thereafter may indicate discharge to a location with insufficient support. Our study found a large percentage of residents being discharged live directly to acute care hospitals. These patterns have been questioned by prior research and news outlets as potentially attributable to nursing facilities sending residents to acute care settings and refusing to readmit them, a practice referred to as patient dumping which was recently ruled upon in federal courts (Egelko, 2019). Likewise, residents going to the community and then acute care may reflect inappropriate discharges to community without appropriate discharge planning. 4.2 Policy Implications Our findings have several implications for policy. There has been increased attention recently on the extent to which state survey agencies have been able to investigate inappropriate FIDs (OIG, 2019). Our findings can aid the efforts of identifying discharges that may raise concern. Our research identified behavioral symptoms, new psychiatric and mood diagnoses, transitions to Medicaid eligibility and impairments requiring more staff time as the conditions with the highest risk for live discharges. An additional finding was that the changes in resident characteristics may be stronger indicators of risk for live discharges than characteristics at the time of discharge. Our research also found differences in discharge patterns by facility type, which also deserves more attention. CMS is also encouraging states to pursue projects funded by the Civil Money Penalty Reinvestment Projects Assistance aimed at preventing FIDs that violate federal regulations (CMS, 2018). This research may help inform the design and targeting of these projects. For example, our state-level analysis compares discharge patterns across all states. Second, our analysis indicates residents with risk factors experienced higher rates of hospitalization and outpatient acute care use, with the caveat that these outcomes could reflect overall worse health among residents with risk factors. These findings provide additional information to inform discussions about how to improve post-discharge outcomes. We found that, while most residents went directly to acute care, a significant portion of residents were discharged live to the community and required acute care shortly thereafter. This may indicate lack of appropriate support in the discharge location and a need for better discharge planning. Lastly, as a policy consideration, it may be helpful to refine discharge items on the MDS assessment to indicate whether a discharge was facility- or resident-initiated, and whether it was involuntary. An MDS assessment item including these additional elements would help assess whether the live discharge of residents displaying risk factors was appropriate and provide another level of protection for this vulnerable population. Such MDS items would also support monitoring of this issue and augment future research. 4.3 Limitations MDS data contains several limitations which must be considered when interpreting our findings. It does not indicate whether a discharge was initiated by the resident or facility and whether a discharge was voluntary or involuntary. Because of these limitations, we could not ascertain whether live discharges were facility-initiated or determine their appropriateness. As a result, we could not determine the number of potentially inappropriate FIDs among nursing facility residents. Instead, our research focused on describing the discharge patterns of residents displaying risk factors for live discharge. By describing these discharge patterns, our research seeks to aid understanding of nursing facility discharges in general and guide future discussion and research of FIDs. To enhance the knowledge base on FIDs, future research should focus on the discharge locations of residents with risk factors for live discharge. Additional research should also examine the impacts of state policies, facility closures and the COVID-19 pandemic on FIDs. References Centers for Medicare & Medicaid Services (CMS). (2017a). An initiative to address facility- initiated discharges that violate federal regulations. Available at: https://www.cms.gov/Medicare/Provider-Enrollment-and- Certification/SurveyCertificationGenInfo/Downloads/Survey-and-Cert-Letter-18- 08.pdf. Centers for Medicare & Medicaid Services (CMS). (2017b). State operations manual, appendix PP. Guidance to surveyors for long-term care facilities. Available at: https://www.cms.gov/Regulations-and- Guidance/Guidance/Manuals/downloads/somi07ap pp guidelines ltcf.pdf. Egelko, B. (2019). Court reinstates suits by patients nursing homes refused to readmit. San Francisco Chronicle. Available at: https://www.sfchronicle.com/news/article/Federal- court-reinstates-suits-by-patients-14106558.php?psid=dnUOj. Lepore, M.J., Yuen, P.K., & Zepeda, S. (2019). Nursing home facility-initiated involuntary discharge. Journal of Gerontological Nursing, 45(8), 23-31. Lepore, M.J., Zepeda, S., & Yuen, P. (2018). Facility-initiated involuntary discharges (FIDs): Final environmental scan. Michel, L., & McAndrew, M. (2019). New York's top nursing homes less likely to admit poor. Buffalo New. Available at: https://buffalonews.com/2019/07/14/new-yorks-top- nursing-homes-less-likely-to-admit-poor-than-worst-homes-data-shows/, Office of Inspector General (OIG). (2019). Involuntary transfer and discharge in nursing homes. OIG. Available at: https://oig.hhs.gov/reports-and- publications/workplan/summary/wp-summary-0000331.asp. Roberts, L. (2019) New data highlights the long-term care ombudsman program's legacy of service, Administration for Community Living. Available at: https://acl.gov/news- and-events/acl-blog/new-data-highlights-long-term-care-ombudsman-programs- legacy-service. Serres, C. (2020). Complaints of eviction in senior care homes up nearly 30% during pandemic. Star Tribune. Available at: https://www.startribune.com/complaints-of- evictions-in-minn-senior-care-homes-up-nearly-30-during- pandemic/571609422/?refresh=true. Silver-Greenburg, J., & Harris, A.J (2020). 'They just dumped him like trash': Nursing homes evict vulnerable residents. New York Times. Available at: https://www.nytimes.com/2020/06/21/business/nursing-homes-evictions- discharges-coronavirus. html. Task Force on Aging and Families. (2020). U.S. House Democratic Caucus Task Force on Aging and Families LTR to HHS CMS to stop nursing home evictions. Treatment Advocacy Center. Psychiatric bed shortages. Treatment Advocacy Center. Available at: https://www.treatmentadvocacycenter.org/key-issues/bed-shortages. Verdana, T.S., & Pear, R. (2018). Complaints about nursing home evictions rise, and regulators take note. New York Times. Available at: https://www.nytimes.com/2018/02/22/business/nursing-home-eviction- regulators.html. Appendix A: Technical Details For each year, we looked through the stream of MDS assessment for each resident in that year and distinguished stays for that resident. We first distinguished between-facility stays for each resident by grouping their assessments by facility ID. We then looked through the stream of assessments for each unique resident and facility pairing to determine within- facility stays. We defined the beginning of the first within-facility stay as the first assessment occurring that year at that facility. We considered the end of the first stay to be the nearest discharge assessment with no subsequent assessments within 30 days. If a discharge assessment was found which had a subsequent assessment within 30 days, we continued the stay and evaluated the next discharge assessment, repeating this process until we found a discharge assessment with no assessments within 30 days. The second and subsequent stay began with the first assessment after the discharge of the previous stay and likewise ended with the nearest discharge assessment with no assessments within 30 day. If a stay had no discharge assessment with no subsequent assessments within 30 days during that year, the stay was continued until the end of the year. Next, we determined whether each stay was a SNF or nursing facility stay by looking at the type of assessment which occurred during that stay. For the first stay for each resident and facility pairing, we also looked through assessment types in the last 4 months of the prior year (September 1 to December 31) in order to gain a more comprehensive understanding of nursing facility usage. We considered a SNF stay to be indicated by the presence of a Medicare Prospective Payment System (PPS) 5-day assessment and a nursing facility stay to be indicated by the presence of a Quarterly or Annual MDS Assessment. For purposes of analyses, we were only interested in new SNF stays, defined as those not preceded by a nursing facility stay. We flagged a stay as SNF stay if it contained 5-day PPS assessment in 2017 that was not preceded by a Quarterly or Annual MDS Assessment, looking back to September 1 of previous year for the first stay and back to the start of the stay for the second and subsequent stays. We flagged a stay as nursing facility if there is a Quarterly or Annual MDS Assessment during the stay, also looking back to the last 4 months of the prior year for a Quarterly or Annual Assessment in the case of the first stay. A.1 Stay Construction " Process: (1) Look for a discharge assessment. (2) Look forward from the discharge date to the discharge date+30 days for any assessment. (3) If an assessment is found within 30 days, the discharge assessment is "not true" and the stay does not end. If no assessments are found within 30 days, the discharge assessment is "true" and the stay ends. (4) The next stay begins with the first assessment after a "true" discharge. © Assessment of any type (PPS, Quarterly, Annual) @ Discharge Assessment - CY 2017: reference year of interest Stay Assessment found in 30 days. "Not true" No Assessment found Stort of Stoy 1 discharge. Stay 2 within 30 days. "True"* Stort of Stay 2 No Discharge Assessment. continues discharge. Stay 1 ends Of1/17 Stoy 2 continues "Lodk farward" period: 4/1/17- "Look forward" period: 7/1/17- ' 5/1/17. Target Date on Discharge 7/31/17. Target Date on Discharge Assessment+30 days Hesessmentel days. \ oy | 62-day gap gap First Assessment Discharge First Assessment 2™ Discharge First Aasecsiment for resident in Assessment After Discharge Assessment rer Disthane 2017. 1/15/17 Identified. Assessment Identified. 9/4/17 Target Date 4/26/17 Target Date ; afifi7 aay A.2 Determination of SNF and NF Stays General logic: «" A stay is considered a nursing facility if there is a Quarterly or Annual MDS Assessment during the stay or Previous Period. Subsequent stays are considered a nursing facility if there is a Quarterly or Annual MDS Assessment during the stay. " A stay is considered a SNF if there is a PPS 5-day and no Quarterly or Annual MDS Assessment between the Lookback Period start date (September 1, 2016) and the entry date recorded on the PPS 5-day. @ = PPS Assessment @ = MDS Quarterly or Annual Assessment mums Lookback period for MDS assessment (Quarterly/Annual) [| CY¥ 2017 | SNF Stay [ NF Stay - " Example 1 below is considered a nursing facility stay because there is a Quarterly or Annual MDS Assessment during the stay. Example 1: Quarterly/Annual MDS assessment found; No PPS 5-day (SNF=0, NF=1) 9/1/16 Previous Period afa7 CY 2017: reference year of interest 12/31/17 je i a | First Assessment of 2017 NF Discharge Date for resident Of1/17 " Example 2 is considered a SNF because there is a PPS 5-day and no Quarterly or Annual MDS Assessment between the Lookback Period start date (September 1, 2016) and the entry date recorded on the PPS 5-day. Example 2: PPS 5-day found in CY 2017; No Quarterly/Annual MDS Assessment found (SNF=1, NF=0) Entry date on PPS 5-Day S116 Previous Period MAE 2fift7 CY 2017: reference year of interest aE Lookback period for MDS assessments PPS 5-Day Target Date SNF Discharge Date 2yafzo1y,- (2/1 " Example 3 is not considered a SNF because there is a PPS 5-day but also a Quarterly or Annual MDS Assessment in the Lookback Period. Example 3 is considered a nursing facility because there is a Quarterly or Annual MDS Assessment in the Previous Period. Example 3: PPS 5-day found in CY 2017; Quarterly/Annual MDS Assessment found (SNF=0, NF=1) Entry date on PPS 5-Day S/1/16 Previous Period a afa/t7 CY 2017: reference year of interest 12/31/17 Lookback period for MDS assessments PPS 5-Day Target Date SNF Discharge Date 3y3fzoi7 = Bfar/a? e Example 4 is considered a SNF because there is a PPS 5-day and no Quarterly or Annual MDS Assessment in the Lookback Period. Example 4 is also considered a nursing facility as there was a Quarterly/Annual MDS Assessment during the stay. A-3 Example 4: PPS 5-day found in CY 2017; no Quarterly/Annual MDS Assessment in Lookback period; Quarterly/Annual MDS Assessment found later in CY 2017 (SNF=1, NF=1)--"SNF to NF" Entry date on PPS 5-Day Previous Period aiLily afifi7 CY 2017: reference year of interest 12/31/17 O/T/16 | | Lookback period for MDS assessments PPS 5-Day Target Date SNF to NF Transition NF Discharge Date 2/3/2017 Date 2/27/17 5/11? e Example 5 is considered a nursing facility as there was a Quarterly/Annual MDS Assessment during the stay. It is not considered a SNF because there is a PPS 5-day but also a Quarterly or Annual MDS Assessment in the Lookback Period. Example 5: PPS 5-day found in CY 2017; Quarterly/Annual MDS Assessment found in lookback period (SNF=0, NF=1)--"PPS assessments within NF" Stay 1 /1/16 Previous Period LAT: CY 2017: reference year of interest ~" 12/31/17 OQuerterhy Annual PPS S-clay First Assessment of 2017 found in "look Assessment NF Discharge Date for resident back" period: FAS /17 a/ifla S/1/16-7/15/17 A.3 Multiple Stays Example 6: Two NF only stays 32-day gap Stay 1 \ Stay 2 9/1/16 Previous Period a/a/i7 CY¥ 2017: reference year of interest 12/31/17 | | | | --I First Assessment of 2017 NF Discharge First Assessment for resident Date 7/20/17 after NF Discharge B/21/17 Example 7: SNF stay followed by NF Stay Entry date on PPS 5-Day 9/1/16 . Z 1/1/17 Stay 2 Previous Period a 2/1/17 staya_ «CY. 2017: reference year of interest il i | s - r 108-day gap \ Lookback period for MDS assessments PPS 5-Day r Target Date SNF Discharge Date First Assessment NF Discharge Date 2/3/2017 2/27/17 after SNF discharge 11/1/17 6/15/17 Example 8: Two NF only stays Entry date on PPS 5-Day 9/1/16 = - AAs? Stay 2 Previous Period ae 2/4/17 stay1 «CY 2017: reference year of interest ad | i i * N } |_-- f L \ 108-day gap Lookback period for MDS assessments PPS 5-Day Target Date SNF Discharge Date First Assessment NF Discharge Date 23/2017. 2/27/17 after SNF discharge 11/1/17 6/15/17 Example 9: "SNF to NF" stay followed by NF only stay Entry date on PPS 5-Day 'Stay1 : = 1/1/17 z Previous Period en afa/li | C¥ 2017: reference year of interest | 12/31/17 | -t KN a | aged a t i 1 Stay 2 9/1/16 | \ Sd ay gap 63-day gap Lookback period for MDS assessments PPS. 5-Day Target Date SNF Discharge Date NF Discharge 27/17 Date 6/1/17 2/3/2017 alee yf First Assessment NE Discharee First Assessment after NF discharge Hote E after SNF discharge 8/3/17 10/20/17 3/4/17 Fa Appendix B: Methods B.1 Data Sources and Analytic File Construction A single resident-level analytic file was created that includes all variables needed to conduct this research by merging stay-level MDS data with corresponding MBSF and FFS data for that resident and POS data for the facility at which the stay occurred. Data from CY 2012 to 2017 was included for each source. We defined the population for each year as all residents with at least one nursing home stay in that year. A nursing home resident can have multiple stays. For each stay, we assigned discharge status as not discharged, discharged live or death, based on information in the MDS discharge assessment. MDS data was used to identify the characteristics of residents discharged live and not discharged from nursing facilities. We included MDS items in our analysis pertaining to behavioral symptoms, psychiatric and mood disorders, Alzheimer's and other dementia diagnoses, suicidal thought and mood distress. We also constructed an Aggregate Behavioral Score, ADL Score and Cognitive Function Score for each resident. We created one set of variables capturing the value of characteristics at end of stay (assessed at discharge or the last MDS assessment). We then compared the value of each MDS item at the end of stay to the earliest value observed, up to 1 year prior to discharge. We created a second set of variables for each characteristic indicating whether the resident increased in the severity or acquired a new diagnosis. To supplement MDS data, we used MBSF to capture Medicaid eligibility at end of stay. We also noted whether a resident change from none or partial Medicaid eligibility to full Medicaid eligibility over the 3 months and 6 months prior to discharge. MBSF data was also used identify residents that only had health maintenance organization. We used POS data to obtain facility characteristics, including state, profit status, urban-rural status, chain status, bed count and changes in ownership. Using FFS inpatient and outpatient claims data, we identified hospitalizations post- discharge, defined as the presence of an inpatient claim in the year following discharge. We identified outpatient acute care use by the presence of an outpatient claim with a revenue charge code indicating emergency department visits or observation stays in the year following discharge. Claims dates were used to determine when outcomes occurred relative to discharge. Lastly, we examined death date of the beneficiary, identified from MBSF data, to determine the presence and timing of death. B.2 Data Analysis by Research Question e Unit of Analysis: Stay e Analysis: Descriptive RQ1: Nationally, what are the characteristics of residents who are discharged live from nursing homes, compared to residents who remain in the facility? What percentage of residents who are discharged live demonstrate such risk factors for live discharge, relative to the percentage among residents who remain in the facility? How have such risk factors for live discharges changed over time? The first objective for this work was to identify the association between resident characteristics and live discharge. To address this objective, we conducted descriptive analyses comparing the prevalence of characteristics among residents discharged live and not discharged. For nursing facility analysis, we were interested in the characteristics associated with two outcomes at the end of the nursing facility stay in each year: being discharged live or remaining in the facility. We compared the characteristics of residents discharged live at point of discharge and residents not discharged at point of last assessment. We looked at both the prevalence of characteristics at end of stay (point-in-time analysis) and changes in the characteristics in the time leading to discharge (longitudinal analysis). A diagram of nursing facility analyses is shown below. 1. Nursing Facility stay alone (SNF=0, NF=1): 1-year lookback DSCHG ASSMT Discharged live: vn or as ony i a . PRIOR PRIOR . PRIOR ASSMT i ASSMT 2 ASSMT n 1-year lookback LAST ASSMT Not Discharged: Cg oa S - 4 vo, Te _ PRIOR PRIOR om PRIOR ASSMT 1 ASSMT 2 ASSMT na = Point-in-time: Pick and compare resident characteristics available from e/ther the discharge/last assessment or nearest prior assessment. « Longitudinal: Measure and compare changes in resident characteristics from first available assessment in the 1-year lookback to last available assessment. For SNF analysis, we were interested in the characteristics associated with two potential outcomes at the end of SNF stays: being discharged live or transiting to nursing facility care B-2 in the same facility. We compared the characteristics of residents discharged live at time of discharge and residents transitioning to nursing facility care at transition point. We defined the transition point as the last MDS PPS assessment (required as part of Medicare-covered SNF care) before the first Quarterly assessment (indicating that the resident has been in the facility for at least a quarter and thus likely to have started nursing facility care). For SNF, we only conducted point-in-time analysis due to the short duration of SNF stays. A diagram of SNF analyses is shown below. 2. Transition Point Analysis (SNF=1, NF=0 versus SNF=1, NF=1). DSCHG ASSMT SNF only & o- Fant @ discharged live " a 1 or not 5-DAY PPS PPS 3 PPS discharged SSMT 2 ASSMT A - DSCHG ASSMT | | SNF to NF stay fan rat os e@ discharged live o ao = 7 , 7 or not 5-DAY . | = discharged: PPS LAsT pps 2 QTRLY/ tT oTRLy; = ASST ANNUAL ASSMT Point-in-time: 1. For SNF only stays, pick resident characteristics available from either the discharge/last assessment or nearest prior assessment (End Point). 2. For SNF to NF stay, pick resident characteristics available from the last PPS before the 1% Quarterly or Annual assessment, whichever is earlier (Transition Point). 3. Compare residents characteristics at the SNF only End Point and SNF to NF Transition Point. For both nursing facility and SNF analyses related Medicaid eligibility, we limited residents to those with MBSF data, an approach continued in subsequent analyses. We considered a characteristic to be a risk factor for potentially questionable discharges if the characteristic was identified in literature and was more prevalent among residents who were discharged live than among those who remained in the facility. Likewise, we also considered an undesirable change in a characteristic (such as a new disease diagnosis or transition to Medicaid eligibility) to be a potential risk factor if the characteristic was identified in literature and showed higher prevalence among residents who were discharged live than not discharged. Characteristics were considered general risk factors for live discharge if they did not appear in the literature as a potential risk factor for FIDs but were more prevalent among residents discharged live. B-3 We calculated the difference in prevalence of each risk factor between residents discharged live and not discharged and compared the direction and size of the difference across years. A positive difference indicates a characteristic is more prevalent among residents discharged live than not discharged, while a negative difference indicates the characteristic is less prevalent among residents discharged live. The size of the difference in prevalence indicates how much more or less prevalent the characteristic is among residents discharged live than residents not discharged. RQ2: What are the differences in characteristics among residents discharged live across all states and types of facilities (e.g., for-profit vs. non-profit; chain vs. non-chain)? Our second objective was to assess how relationship between resident characteristics varied by state and facility type. To address this objective, we limited analyses to the risk factors identified in RQ1. For each risk factors, we examined both version (end of stay and change over time), even if only one version was found to be more prevalent among residents discharged live than not discharged. We also limited analyses to nursing facility residents since no risk factors were identified among SNF residents. For each year of analysis, we included only stays that had matching POS data for that year. We conducted descriptive analyses to assess the prevalence of risk factors within each state and for each facility type characteristic. Using the same method as our across years analysis, we compared the difference in prevalence of each risk factor between residents discharged live and not discharged and compared the direction and size of the difference across states and facility types. Lastly, we compared the prevalence of risk factors among residents in active and terminated facilities. We focused on the characteristics of residents discharged live from each facility type as the majority of residents were discharged live from terminated facility, and thus the populations of residents discharged live and not discharged were not comparable. RQ3: Do post-discharge outcomes (e.g., risk of death, worsening health condition leading to hospitalization or emergency department visits) differ among residents discharged live with and without risk factors for live discharge? The last objective of this research was to determine the outcomes of residents discharged live with and without risk factors. To answer this objective, we focused on hospitalization, outpatient acute care and mortality. We examined three timeframes: from zero to one day after discharge, from 2 days to 30 days after discharge and across the full 30 days after discharge. First, we examined hospitalization, outpatient acute care and mortality of all residents discharged live over the three timeframes. We combined hospitalizations and outpatient acute care to determine overall acute care use across each timeframe to distinguish resident who directly to acute care and residents went first to a community or non-acute setting and then acute care. We also examined mortality for residents experiencing acute care within each timeframe and compared the acute care use of residents discharged live from active and residents discharged live from terminated facility. Next, we examined the characteristics of residents experiencing hospitalization and outpatient acute care. We compared the rate of hospitalization among residents with and without each risk factor across the three timeframes. We repeated this analysis looking at outpatient acute care. Lastly, we analyzed the MDS item on discharge location, comparing the prevalence of each risk factor among discharges to each setting. "1BA9] TOO'O=d ay} Je JUBDYIUBIS Ajjeosijzeys sdnoi6 aBesn sjed aynoe UsaMjeq sdUsIAYIG OAT] POSIeYOsIC M posieyosip JON @ sxx OLOADS AIO A, x OTOAIS xg OLARYO xx TOTARYO x OLARYO 'qyeSoissy 'qyeSo1ssy IOYIO WdA9S [RGI9A, dI9A05 [eoiskyd O10A96 %0 a va %C %E %ov %S %9 YL %8 %6 IOIARYIG YPM SJUApIsay JO JWIdI19g ZT0z 'snjzes ebseyosig Aq AejS IN JO pug je SwojydwAs je1iolAeyag "T-D o41nbi4 $10}9e4 4SIY 10} Sa|qGeL e7eG pue Sydeiy |euOCIIppy :D xipusddy C-1 [BA] TOO'O=d 3y} We JUeDYIUBIS Ajjennsijeys sdnoiB abesn sued aynde usemjeq sdUsIBYIG DAT] POSILYOSI MH =S« Pos IeYOSIP ION x~4AoUdpuadag [RO], 0} osueyD x%xAouapusdag [e310 J, x4AoUdpuadag [e}0], IeaN %0 %l %C YE sJUOpPIsoY JO JUDII9g %v %S %O Lt0z 'snje1s sbieypsig Aq 1e39A {Seq BY} A2A0 BHueyD pue Aejs 4N JO pug 3e TGV 103 Aquepuedseg *Z-5D a1nb14 C-2 (panuljuo5) xxx "ARDA se] 943 Ul AUIqIBIe WLEST WSSST MWOT'ST WLL'VT MWELPT MWOG'HT | %GL'OT WLS'OZ WLE'OZ MVWPT'OZ WSL'0ZT %WIG'ET pleoipay, oO} uoNsued, a &SUYJUOLU 9 yse| BY} Ul AQUIqIBIe LBL WEYL WL WEL WL WOOL | WOE BWGE WEP MVWSLE *WLL'E %r9'e PleoipaW 0} uoNsued, 2k « SYJUOW € 3se] 9Y3 Ul AGI %ENZ WBS WIS BWLEZ WOT MWSS'Z | WZIT WIT *WrIT VWr'T WIS'T %6v'T PledIpaW 0} UONISUeI | xxxAB4S JO pug je WIGEL W2STL WBLTL WI9TL WSL TL Wp7'7L | WOG'PB WE'VE WES PS %WOGPS Wr rv %WLZ'rs Ayqi6ije presipaw ling ABS Jo pug ye Ayqibye WS7T %WLOT %6ZT %6TT %bT'T %LO'T | %6E0 %B8EO %ZEO WED %OED %rr'0 plesIpaw [eed sn}eis jeng WIT? WET BW BOP Bb MVE | WSERS WLS BEB MVWEEB *WTE'B %SZ'S8 S6= aby WSS'7Z WIVES VWISZ W66'PZ MVWGL'SZ WL99T | WHS'TE WE9ZE WI'ECE Wp ve WE PE YWPpPry'Se 56-8 aby %SO'ST WEE WISZ WIL'GT MWEO9Z WEITZ | %99SZ WLE'ST W6ZSZ MWISSZ WEST M%WIT'9Z §8-SZ aby WILTZ VWOTS WTZOZ WLy6T %W%69ST WSILT | WLL'LT WGG'OT WEST WLS'ST WOG PT W8Z'PT SZ-S9 aby WLTIZ WLYIZ WIB'SZT WEISZ WIEST WIZ | WL9'9T WOG'9T VWES'OT WIT'S9T VWvrO'9T WLB'ST xx4959> OY WSES WES WSS BPS BLb> RII | WPI Wr *WSr VW br *WZ0'r %ZO'D dueds|H-UOU 18410 %69 989 W2L9 %999 %8Z9 %099 | %69S WSS MEPS %BES %8Z'S %60'S auedsiy %rT'6T VWEO'GT WE9'ST WIS'BT MVWSE'ST WI9ST | WAS'ST WOE'ST WGO'ST WZBPT WELT VWvG'rT auedsiH-uou Ajuo 4De/q sex xOIUBUSIH %LS°B9 WeLB9 WET'E9 WG6E6I %80'OL %WBO'OL | M80'PL WSS PL WBS PL WZE'SL WLGSL M%WSE'OL -uou Ajuo az WIIG WIIG WIZLS Why LS WEBS %BB'SS | %50°99 %/49°99 %STLI %06'LZ9 W889 %96'89 ajewie4 WES EY WEEP Welty WITH WITH WITH | WSG'EE WEEE WGSZE WOT'ZE WCSTE MV*PO'TE «xx2/2W %~I9T WIS OT WIEST WEEOST VWb7'9T %WLZ'9T | %58'S9 %95°99 %0S5°99 %88'99 %/699 %B8T'L9 S]UAPISA/ JN IIE JO % TZS'ETZ_-EOT'ZZZ_-GPB'TZ@Z ~Bz9'~7z = LEO'T7Z =7AT'EztZ | 8BZ'S8B8 PryS'S68 770706 099'806 S99'TT6 L8P'TZ6 4aquinn so1ydesBoweg 4toz | 9toz | stoz | vtoz | etoz | ztoz | ctoz | 9toz | stoz | vtoz | etoz | ztoz Al] pebseyosig poBseyosip JON (uawssessy SQW 1Se7 10 abueydsig }e pessessy) s}USpISOY 4N LT0Z-ZT07Z 'SoSl9}De1eYD WSpIseY IN *T-9 SIGGL C-3 (panuljuo3) SZ0°0 8Z0°0 SZ0°0 x«x(SJUlOd ZT JO no) 060°0 680'0 9Z0°0 TOO"O z00'0- T00'0- €00'0 ZT0°0 TTO'O aBueyd julog ueaW SAV #kKDIBAIS AlaA 0} B1BABS %68'0 %66'0 %60'T %SE'T %C'T %OP'T %9S'0 %€9"0 %EL'0 %E8'0 %S6'0 %TO'T JO SUON 'PIIW Woy Sav KKKBIBARS 0} %~ZE %0S'E WE9'E %ZO'b %bT'b %L0'b %bS"Z WELT %96'Z WLT'E Mbv'e %0S'E SUON JO PIIW Wold Sav #KKBIBAVS WBE WS %8L'b %E0'S %CT'S %P0'S %0S'E %SL'E %60'P WTE'b %PS'P %ZO'b 0} JOIABYag 4240 *kKBIBAVS %TT'9 %0E"9 %S2'9 %Tb'9 %bb'9 %86'S %TS'b %T9"'b %98'b %TE6'b %00'S %66'V 0] JOIARYSq |EQUOA #KKBIBAVS %OLE TEE %68'E bev %OT'b WLLE MELT %L8°C %C6°7 %60°E WLT'E %ST'E 0} JOIAeYyag |BoISAUd 41e9A JSeq dy} 49A0 Huey) JOIAeYyog xxx0] 26016 Bby 'solAeuag %90'T WIZ T WLE'T NZL T %Z3'T %vs'T %vs'O %v6'0 %60'T %8C'T %Ybr't %vs'T anisseibby auaaes Ala, «xx 2] e@Be15By "olAeuag YI" voy YEG' %SS'S %SL'S %88'S %98'E NEC %99'P %60'S %TS'S %IL'S aaissaibby a1aAas x2) e60166y 'solaeyag MZTCT WSCCT WHECT WECCT WECCT WESCT | WLETT %W6CCT WEBCT WE6'CT %WOT'ET %EBT'ET aalssaiHby ajesepow xxx 2 1ebo15 By JO|IAeyaq SAIssolbby %SS°78 WETS WITS WLv'08 WTZ'08 WSL'6L | WEES WGS'7S WISTS %S9'08 %88'64 %0S°62 JEWIUIW 40 SUON *kABUIO %L0°9 MEv'OD %06'9 WIGL %BL'L %Bs'Z MLL'S %ST'9 %T6'9 KIL ZBL %9IT'3 'JOIARY®Y B418A8S #* kOIABYSG %T9'8 %86'8 %Z0'6 %0S'6 %09'6 %EV'S %Ie"L %65'°2 %80'°8 %ZE'8 %Z9'8 %6ZL'8 JPQJSA 51BA0S #& x OIABYSG %ZLb WIP %Z0'S %LS'°S %9P'S %8Z'S %60°F WSE'Y %6v'v %~8'P %TO'S %CT'S [2d1SAYd B48AaS JOIAeyag 4toz | 9toz | stoz | wtoz | etoz | ztoz | ctoz | otoz | stoz | wtoz | Etoz z10z 3A] pobseyosig poBseyosip JON (quewssessy SQW }Se7 10 oB1eUudsIq }e pessessy) s}USpISoY 4N (panuiuos) ZT0Z-ZT07 'SoISI4a}DeIeYD WSpISeY 4N "T-D9 PIGeL v Oo (penuruos) %870 W970 %WIZO0 %0Z70 %8T0 %9T'O | %9TO %ST'O %ET'O %IT'O %OT'O %O0T'0 **xxCS1ld %Z9'T WIT WEZT ETT %TTT %00'T | %ZET %LZT %6T'T %68°0 %T6'O %T6'0 x««21UBIYOZIYIS xxx (PIUBIYAOZIYIS UU} 18430) WBZ WIE WAVE BWELE BWSEb BOrry | MLZ MVNEET MVWRG'E WAVE MWOG'E WEL'b 4apsosiq 290Y49Asq x«x(49psosiq Jejodig) %LLT %W8BET WELT WELT PST LIT | MOTT MITT %ST'T %60'T %SO'T %80'T ualsseidaq J1ueW WST6 VWbr'6 W086 W686 WENO %WZ9OT |] %B8L'L WELL MWOER WSR WETS %EI'S xxxUolssaidaq %r0'S %W%lCB WIS VNETBS 678 WIL | WZS9 %WSL9 %W60L %*WT89 %BL'9 %T6'9 xxx AJOIXUY enusweq S$ JOWISUZ|/V-UON MESS WTI WED VWSb9 %ZB'9 W689 | HOSS M%86'S %B6'9 VWEV'9 WE'D %S8'9 pue aseasiq s,Jewiayz|y 4123, 1SEd 94} 19A0 SasouBbelig MON %IT'T %560 %f/2'70 %59'0 %850 %WZb0 | %9Z'0 %590 ESO WSO %B8E0 %ZE'O **&O0S1d 2786 WES6 WSS MPR WETS *WHEL | W6S'OT %066 %SE6 ILS NPrG's %ET'S 4 2UBIYCOZIYIS ek (PIUBIYCOZIYIS ue} 48430) %Sr'S %676 WOOT WLIOT WIZTT WS OT | WGP'OT WCT'TT %WGSTT %OL'TT %OT'ET WTE'ET JBPIOSIG) D1}OUIAS_ xx (4apsosiq Jejodig) %8ZL WrTL WSEI MWIN9 W6EO9 MESS | MWLED9D %9ID09 WESS WLS MOPS %OT'S uo|ssoidaq 21ueW WE'6h VWrS'6r %W%O0OOS Wl66h VWES'6h WCL'6P | WGO'ES MWEG'TS MWOZES WCS'ES WHLES WEL'ES xxU0|SS31daq WES'TE WSS'OE WO'OE WHG'ST VWEL'LT WGS'ST | WIL'TE WEE'TE WST'TE WET'VE WOL'ST MW9IT'LZ 4 A}9[XUY x *2UBWWEQ S$ JOWIBUZ|y-UON MIE'GE WI'OK WLEOY WIG Ob WEIR VWEE'Ch | WSIS WST'ZS WOCLS MWBr'LS WILLS %Pv0'8S pue aseasid S,Jawleyziy 4toz | 9toz | stoz | vtoz | etoz | zctoz | ctoz | 9toz | stoz | vtoz | etoz | ztoz 3Al] poBseyosig poBseyosip JON (uawssessy SGW 1Se7 10 abieydsig }e pessossy) s]USpISOY IN (ponunuos) ZTOZ-ZTO7 'soNsePeIeYD WSPISeY IN = "T-D 2IGeL C-5 (panuljuo5) xx 10 J0j Asuapuadaq %90°8T %WSZST WIEST %WE0ST WELT WS9ILT | M8SE6T WIZ 6T WvG'6T %9L6T %%EZ'ET %SE6T aanag Ala; ak 1OV %88'8Z WIESC WpT'82 WSOLZZ WLO9Z WLS Z| WLIZE WICTE VRprte WOrde %86'8zZ %LELT Joy Asuepuadagq ajeaAes x IOV JO} WISZT WESZT WHS'ZT WINZT WIP'CT WLO?ZT | W2V'CT WTL'ZT VW2S'CT W2S'CT WILT %IP'2T Aduapuadsg 3}e49powW ax* IOV %9Z7'6 %68'8 %8Z'8 %b6'8 %ES'S %6r'8 %99'8 %0S'8 %ZE'S %0E'S %EE'S %6T'S Joy Auepusdaq piIW ax IOV %SZ'TT %89'O0T %SL'6 %B8E'6 %T9'S %9T'S %ZL'S %Tr's %90'8 %08°L %Z9'L %OE"L 4oj Aguapuadag |ewiulW axe *k IOV %TS'S %SS'S %bO'S %Z6'S %LE'9 %B8Z'9 %69'P %BL'P %E6'b %C27'S %T9'S %9T'9 Joy souepusdapuy Jean xk LV %LO0'P %Z9O'P %6S'P %06'P %ZE'S %Z8'S %9G'T %L9'Z %68'7 %66'72 %9ITZ'E %G9'E 4oj ssuapusdapul sqyBnouL %0E"0 %ZEO %LEO %br'0 %TS"O %L9'°0 %0E"0 %SE"O %6E°0 %bvr'0 %zS'0 %Z9'0 [Epping a48AeS **x9/29S POOW %9IP'S %9P'S %E8P's %MEL'S %Z6'S %L9'9 MLE'L %MSE'L MWOE'L %ZE'L MEPL %6S'Z a}98|dWo0s 0} ajqeun #&&S5S914519 %LT"O %TZ'O %bz'0 %8Z'0 %0E"0 %EE"O %ET"O %9T"O %8T"O %6T'O %ZZ'0 %GZ'0 POO] d.19A9S #5591951] %76'0 %TO'T WET T %ITT Wve'T %ov'T %LL'O %C8'0 %S6'0 %WEO'T MOTT %WS8t'T Ppoow Bares AjeeJepo~ #**5S914519 HET WES' ISP YWEB'Y HELV %90°S WISE HLT AWGE'v %O0S'b %OG' Pv CS' PpoowW e}]e19poW %IEOT WLB'OT WESTT WSO ZT WOB'CT WSP'ET |] %89'GE MVWPOT WvO TTT WOLTT WLTZT %EL*CT +#445S9.951Q POOW PIIW ***z9S944S1 W~EVBL WIGVLL WSI'I9IL BWIPVGL AWEV'vL WlG7L | WM69NLZL WGSLZ'9L WESSL WO6 VL BWGT vl WMSGE*EL POoOoW JEWIUI,] 10 SUON exe BOUBUIUODUT %MES'ZGS %WSO'OS %WOB'OS %69'6h MRWOE'6h WIC 6h | WIZ'LZG WLP'SS WLE'SS AWLP vS MVWpp'|es %Z9'CS JBMOg S1BA9S exe BOUBULUOSUT %IPVLS WELZ'9G WOOLZS WIIG RWYIELZGS WIE LS | WIE'99 WSZ'99 W%TIO99 %W9L'S9 WEG'SO %GS'S9 Aleulip 848085 DASH SABYD 19410 4toz | 9toz | stoz | wtoz | etoz | ztoz | ctoz | 9toz | stoz | wtoz | Etoz z10z Al] pebseyosig poBseyosip JON (Juawssessy SGW 1Se7 10 abeydsig }e pessoessy) s}USpISoY 4N (panunuos) ZTOZ-zZT0z 'soNSMe}eIeYD JUSPISEY IN "T-9 SIqeL Oo (penuluos) x«x(SJUlOd €-9 abuey) €2T'0 OzT'O OzT'O T2T'O £2T'0 92T'0 bv0'0 £v0'0 €v0'0 Tv0'0 vv0'0 0v0'0 aBGueyd Julog uedW S4D xx x19UIO jo Aduapuadaq NW8b'7~ %69'Z %86'Z MITE %99'E NET'v %T8'T %L6'T %OT'Z %9Z'Z %b6p'Z %LL'T Je}0L 0} 10V xax(Squiod gz-9 aBuey) Z£L9°0- cvs'0- Z6E'0- Ose'0- c8T'0- 290°0 Ssov'0 6¢cr'0 ozs'0 sT9'0 zv9'0 vTs'0 aBueyD Julog uesw dav *kKDIBABS %TZ'0 %CT7'0 %LZ'0 %CE"0 %SE'0 %pbr'0 %TZ'0 %vz'0 %LT'0 %0E'0 %SE'0 %0b'0 0} sqy6nou! jepioinsg xxx (Squlod 22-0 ebuey) eaHueyy Z9€'0- ss¢e'0- cse'0- sse'0- 66€'0- 6ce'0- €oT'0- SOT'0- 680'0- €90'0- €Z0°0- 980°0- JUIOd UBB 2103S pooW **KBIBAIS %WIVCT Wrr'~el WS9'CZT WIG72T MWOM'CT WEP'ZT | WIMTT WLI TT WEL'TT WILTT WGP'TTt WSv'TT 0} BDUSUIUOD [aMog KxXDIBADS %98'0T %S9OT WSS'OT WOGOT WET'TT WISTT | %TS'6 %TIL'S %T8'6 %98'6 %TL'6 %TO'OT 0} B9USUIUOD AUeULI ABSA JSEd 9} 4BAO BGueyD SI1S19}-eeUD 19410 seek JULI ed WIT %L6'D %ST'S %Z9'S %PbL'S %ve'9 %E8'O MWIE'L WELL MITES %68'8 %Cv'6 %E6'6 BaUBoD asaaag kx GUS Wed WT WOT'PE WET'VE WZT'SE WEOOSE WEO'LE WOB'IE | MIG'SE WEE'GE WTB'GE WTE'OP WSGS'OV %TS8'0v BAnUBOD ayesapow sek JUOWUUIed WIT %GLLE WGELE WOTLE WSELE WILE WGESEE | WOE LZ WBL9Z WLG9Z WOT9C %E68'SZ %L0°SZ aanluBbo> PIIW ene UOWUIEC WIT %IZVOC W290 WIISGT WEE'BST WEST Wpv'6l | M2972 WZETZ WES TZ WL6'0C MBWTT'OZ %LTC'0Z BAINUBOD ON ¥x%eIDUIO %9Z'S %S9'S %6E'9 %98'9 %L8'°L %ZL0'6 %Z8'b VWCC'S %EL'S %LE'9 %LO'L %08°2 4o Aquapuadagq [e301 ax IDV je 40) ADuapuadag %0Z2'°S %0S5°S %98'S %6T'9 %Z8'°9 %SEL %B8v'S %BL'S %LT'9 %E9'9 %8T'Z %OL'L jej0, Aeon 4toz | 9toz | stoz | vtoz | etoz | ztoz | ctoz | 9toz | stoz | vtoz | etoz | ztoz 3Al] poBseyosig poBseyosip JON (quawssessy SGW 1Se7 10 abieydsiq }e pessoessy) s]USpISOY IN (ponunuos) ZTOZ-ZTO7 'soNsePeIeYD WSPISeY IN = "T-D 2IGeL C-7 *9/29S uOKOUNY BAIUBOD 3u) Bulsn passasse sem JUsWIedwu! dAbuU0D, "VAIBased Ayidej-UOU JO 1Qy WUOpIed JOU pip JUSpIseJ d}edIpu! Osje pjnoD, "SAW 24} UO BjedS pooW ey) Bulsn passasse sem ssa.jsiq poop, *(Sav) ayeos jes0lIAeyag aaisseiHby a4) Buisn passasse sem JolAeyag aAissaibBby ayeboibby, "(ZT0Z UO peseq) T00°0/T0°0/S0'0 JO yoyNnd anjeA-d e UO paseg JUBIYIUBIs Ajje313S1}29S S] SUBBYIG = xxx/xa/x se SYJUOW Z }se7] Ul JUaLUSSaSsY aBueyD JuesyiuBis %LL'S %T0'°9 4toz | 9toz | stoz | wtoz | etoz | ztoz | ctoz | 9toz | stoz | vtoz | Etoz Z10z 3Al] poBseyosigq poBieyosip JON (quU9WISSeSsy SGW 3Se71 10 aHeydsig }e pessessy) s}UuSpIsoYy IN (panunuos) ZTOZ-ZTOz 'SONSHePeIeYD JUSpPISOY IN = "T-D SIGEL O (panuijuos) Ae3s INS 6buling %00'T %EO'T %bO0'T %v0'T %b0'T %TT'T %bv's %TS"8 %97'8 %9T"S %6T'8 %19°8 = jena Inj 0) sbueyD Aes ANS Bung jeng %O0T'O %O0T'O %TT'O %TT'O %60°0 %60'0 %TZ'0 %8T"O %02'0 %TZ'O %LT'O %6T"O jelj4eg 0} aBueyD Aes %2SLZT WEGLT WIE'LT WILT WEE'LT MWGEE'LT | WS6'6b Wr 6b VWlrSr %I'Sb WETS *Wprr'srv jo pug ye jeng [ind Ae YS Jo %66'0 %80'S %L0'S %ZL0'S %S6'P %S8'P %MEV'E %MLV'E %0S'E %0Z*E %L6'C %06'Z pug ye jend |elved snjej}s eng %ES'E WIZE WITTE %S8'7 %BL'Z %2L'T %TS'b %C9"'b %69'b %8b'b %SS'v %LS'b G6= aby WISLE WEL RWOVLS WSL VWelvl42 ROLLS | WILE VWI6LE WIGE6C WELE6ZC W/BOE MBWOCTE G6-S8 aby HWLVEE VWbvCE AWCLEE WEGEE WEP ve VWCT'SE | WHT ZZ W842 WILLZ WLBLT MWSESZ *WIS8'8zZ G8-G/ aby %0L'SZ %WOLSZ WvOST WBSrZ AWS v2 MVWG6V'EZ | WG6Z'0Z WS66T MWOGST WEVST WEE LT BWLL9T G/-G9 aby %6L'6 %6TC OT %TEOT %00' TT %EOD'TT %EG6 OT %MLT TT MBC TT %TL OT %CC TT %CS"OT %ETC OT S9> aby %09'S %05°S %Cb'S %Pr8'S %rvs'v MIT WEI WLS RIS %86'P %S0'V %0G'E a1UedsIH-UOU 19410 %PpS'C %9IS'E %MLY'E %PpS'e ESE %b6V'C ELV OLY CLV %69'P %6S'P WEL'Y ojuedsiH diuedsiy %Lbv'6 %bS'6 %bz2'6 %9T'6 %OT"6 %S6'8 W%PT'ET WS6CT WHGZT W227'CT WED'CT MBWE'TT -uou Ajuo yoeIg d1uedsiy %6ETS WOVTS *WLVTS WITS VWESTE MVWpr'es | %75'89 %00'69 %62'69 %98'69 WED TL %W%EETL -uou Ajuo ayy %l8'6S %GE'09 %0Z'09 %OZT9 WIS'T9 WEI | WLBVS MWhL'SS %ES'9S WL6'9S MWEILGS %MWPC'SS ajewe, %8T Ob %WS96E BWOE'GE WOBSE VWhrSE VWB9LE |} WISE VWISSE MWZ0SGE WBZ vE WIOvVE YWSGECE aleW 689'E9E'T EZT'ECh'T EZT'ETH'T OPZL'Evh'T TES PSHT peZ'EEb'T| GZp'68T LSO'TOZ ZZe'ZTZ E0E'ZTZ BEZ'STZ €02'8TZ J9quinN so1ydesBoweg Lt0z 9T0Z stoz vtoz €to0z ctoz~ Lt0Z 9T0Z stoz vtoz e€to0z~ ctoz~ 2Al] pobseyosiq 2425 IN 0} pouoisuess (Aeqs 4NS JO pug 043 3e pessassy) s}uapisoy INS LTOZ-ZTOZ 'SONSHewWeIeYD WUSpPIseY ANS = "Z-D ZIGeL c-9 (panuljuo3) %0S'0 %Sb'0 %8E'0 %EE'0 %8Z'0 %PrT'0 %89'0 %9S5'0 %6r'0 %8EC'0 %SE'0 %9Z'0 dSld %bO'T %SS'T WLE'T %TE'T YLT %TCT %~2I'V VvE'V %SV'E WEV'E MpE'e %CCE B|UasYdOZIYIS %EO'T %89'T %b0'C %67'C %6E'T Mbpe'~ %8B'E %S0'V %S6'P %be'S %Sb'S %E9'S Japsosiq I10YIASq (4apsosiq sejodig) GET WvE'TS WIEST HEE TST WlC TS %8T'T %68'E CBE %E9'E Wve %MBv'E %M%PpT'E uolssaidagq 21UeW WBS'OE WST'OE MWEZ'OE MWSO'OE %9G'6Z M%WIZ'6C | "WIT'EE WL'6E VWOL'6E WPO'OWV MVWEL'6E W%WE0'OvV uoisssidaq %T9'6T %6L6T %6S'6T YT? 6T %SS'ST %6S'LT WEOES WG ET %08'ETS %HSVEZ WBVCC WSBT Ajoixuy enuaweaq S,J@WIaYZ|\y-UON pue WMLS9T YWLE'OT WEV'9T WEE'OT WLG'OT MWLG'OT | WL6'9E WLT'LE MBWLL'LE 8T'8e %SS'6E Wy9I'6E sS5e9SIC SjASWISYZIV sosouBbeig ajebalbby 'IO|lAeyaq SAIssobby %CC'0 %ET'O %L7'0 %TE'O %SE'0 %Tv'O %vS'0 %29'0 %69'0 %06'0 %86'0 %OT'T auaAes Ala, aqvebaubBby 'uolaeyag %Ppt't MET T %OE'T %Gr'T %IG'T %LI'T W6L'T %00°E Wve'e WELE %LB6'E MvE'y salssolbBy a1a00¢ a7ye60166y 'solaeyeg %OE'S %GE'S %Ybr's %ID'S %Tv's %~ZG'S %S8T'OT %LEOT %6P'OT %28'OT %r6'OT %SO'TT ealsseibby aj}erapow 72 ]eba166y "lo|jaeuag aaissalbby %SE'EG WET'EE %WOO'ES WBL'CE %%WE9N'TE WOP'CE | %OS'98 %ZO'9S MWS8r'S8 VWSS'rS MAWIT'v8 MWTS'Es JEWIUIW] JO SUON 49410 WILT %ES'T %T6'T %S0°Z %80'°7 %MCT'TS %8TC'V MIS vb %S8'v %8E'S %8S'S %T6'S 'JOIARY®Y B18A8S JO|lAeyog %~S'TS %vS'7Z %Z9'TST %v9'Z %99'T MLS Mbv's %LS'S %BL'S %LT'9 %8Z'9 %6b'9 JPQ4I9A S49A0S JolAeyag TET %CE'T %IC'T Ly't %b6p'T %CG'T %90°E LTE LTE WCLE %TG6'E MEDD Jed1SAUd S18A3S Jo1aeyueg Lt0z 9T0Z stoz vtozd €to0z ctoz~ Lt0Z 9T0Z stoz vtozc e€to0z~ ctoz 2Al| pebseyosiq 42) IN 0} pouolsuesy (Aes 4NS JO puy 24} 3e pessessy) s}uapisey INS 10 (panunuos) ZT0Z-ZT0z 'SoSHEP~eIeYD JUSpISOY INS "Z-DeIqeL = (penuluos) Tay 4045 MWIv'OE WOOTE MWZOTE MWL6'6Z WEST %E60'8Z | WSZ'Sh WSS VWLOSh WISEV WE Try VITOR AdUapUad|d aleAas Jay joj Asuapuedagq %ZSTC WOT WITS WS2TC MBWETC MAWHTI'TZ | WIEST MWES'ST MWGST WOG'ST %WOC'9T MWCE'OT 27e13poW 10V %IZLT WLTLT %86'9T WEZLT WSELT WIPLT %ET'B %9T'S %9IT'8 %SE'S %br3's %G6°S Jo, Anuapuadad pi!W Jay 10) Anuapuadaq WLO'PT YWEP'ET WCE'ZT MWESCT WLS'ZT *WpHz'ZT | %9P'S %ST'S %ZO0'S WLT'S %6T'S WES JELUIULW Tay 405 %rb'D rrp %99'r %90°S %BE'S %9L'S %S8'T %68'T %S6'T %be'Z %9b'S %BL'~ souUapuedapuT IPSN lav %OE'Z %ET'Z WSr'Z %98'Z %9Z'E %09'€ %bZ'0 %9Z'0 %08'0 %96'0 %IT'T %672'T doj eouapuedapur syyBnoul %ST'O %8T'0 %0Z'0 %rz'0 %8Z'0 %SE'0 %67'0 %LE'0 %Z'0 %8r'0 %SS'0 %69'0 lepising ajaaas 3]29S poow %09'T %ES'T %TS'T %ITS'T %9S'T %69'T %98'E %08"E MIs' %ZB'E %E6'E %IT'b sBadwioD 0} ajqeun ssaijsiq %TT'O %ZT'O %ET'O %ST'O %T'O %rz'0 %9T'O %8T'O %TZ'O %vTZ'0 %8Z'0 %EE'O Poo 3120985 sseujsiq poow %0S°0 %SS'0 %Z9'0 %TL'O %E8'0 %TO'T %Z8'0 %v6'0 %EO'T %ET'T %8Z'T HLT BUaAes Aja}e18poW ssajsiq %66'T %BT'7Z WMLETS %G9'T %ve'Z MICE MWES'E %9I6'E YOST COV %ES'Y %LTS PpooW 9}e0poW %CC'S %8L'6 %TIZ'OT %WC8'OT WPS'TT WSP'CT | %90'CT %OB'CT %CV'ET %ET'PT %E8'PT %BZL'ST $S019SIQ POOW PiIW 25S01SIG poow %v0'r8 MWETES WSS'78 WSIS *WS8'O08 WCl'6Z | WSZ'6L w66L4L WZOLL WIGSL ROG PL MAWLL°CL JewIUIW] 40 SUON a@UaUUODUT %CO'SC WLVEC WESCC WETC Wlp'O0C WIEST | %9G'O0S NME9'SH WELD BREr'Sh WrGEv BWLE'Crv JOMOg S1BAaS @9UaUUODUT %CO8C %669C BwWHS9C WOLSC WLHSC MWOVSt | WSESS MBWITLZS WOLIG %BW8ESGS %W80SS %6S'bS AleULI)) 8480985 SOSLSPeILUD 19410 Lt0z 9T0Z stoz vtoz €to0z ctoz~ Lt0Z 9T0Z stoz vtoz e€to0z~ ctoz~ 2Al] pebseyosiq 2425 JIN 0} pouoisuesy (Aeqs 4NS JO pug 043 3e pessassy) s}uapisoy INS (penunuos) ZT0Z-ZT0Z 'solsi4e7}De1eYD WUSPISeN ANS "Z-D F1GeL C-11 3]29S uONDUN SAIIUBO>D 34} Bulsn passasse sem JuaWIeduu aAibuo05, 49aiBased Ayploej-uou 410 Tqy Wuojiad JOU pip JUapIsas syeo!pul Ose pinod, SCW 84} UO aj/23S pooW ay} Bulsn passasse sem ssaijsiq poow, (Sav) a]e9s jesolaeyog daissai6by ay} Buisn passesse sem solAeyag oAissoibby ajebaubby, que wuiedwit %9IC'T %0E'T %IE'T %SP'T %SS'T %MTL'T %CC'E %Sv'E %0S'°E MLE %E6'E %0C'V BANIUBOD a1aAaS quewuedwit WLL'7T WIVTT WIWG'TT MVWI'CT WLTET VWCEET | WIZ'ST MVWEV'8Z *WL9'GZ %IG'ET WES'OE MVWGE'TE aaAnuBoD ajyesapow quewuledwit %Z6'8Z WOLBZ MVWSE6C WeS6Z VWCE'OE WCCOE | M2ZZ9C WTI VRWGE'9IC WC6'SC WE6SZ %BS'SZ aanluBOD PIIW lusueduwy MWEC'VS WEEVS WLS'ES WLSES %W69CG WLIZS | WLS'OP WST'OV MWEL'BE VWLI'BE WZLE MWEL'IE @AUBOD ON -49U}O 10 %PT'T WPT %6T'T %ST'T %WLE'T %TS'T %S6'T %ST'Z %MTETS %LS°Z %L3'°7 %60°E Aduspusdaq e301 10V je 40) ADUBpuadaq %8T'T %PC'T Wbe'T %Sb'T %9S'T %LL'T %Z0'E CTE CEE ESE %08'E %LC'V je30, AjWeen Jav 10; Asuapuedagq %S9°L %06'L %90'8 %Z0'8 %SZ'8 %9b'8 %SO'ZT MLILZT BwWEVLT BWSLLZT WES'ZT %EBLT auanas Ala, Lt0z 9T0Z Stoz vtoz €t0z cZtoz Lt0z 9T0Z stoz vtoc €toz Ztoz al] pebseyosig 341209 4N 0} pouolzsueiL (Aeqs 4NS JO pug 94} 3e pessassy) sjuapisoy INS (penunuos) ZT0Z-zZT0Z 'SONSMa-eIeYD WUSpPISeEY INS "Z-D FIGeL i2 Oo Table C-3. Differences in Risk Factors for NF Residents Difference (Discharged Live-Not Discharged), % 2012 2013 2014 2015 2016 2017 Dual Status Full Medicaid eligibility at End of = -12.03 -12.71 -12.89 -12.81 -11.91 -11.39 Stay Transition to Medicaid eligibility in 1.06 0.79 0.89 0.77 1.02 1.07 last 3 months Transition to Medicaid eligibility in 3.56 3.30 3.52 3.37 3.74 3.57 last 6 months Behavior Severe Physical Behavior 0.16 0.45 0.75 0.53 0.61 0.63 Severe Verbal Behavior* 0.64 0.98 1.18 0.94 1,39 1.35 Severe Behavior, Other** -0.28 -0.14 0.05 -0.01 0.18 0.30 Severe Aggressive Behavior, 0.12 0.24 0.46 0.27 0.42 0.40 Aggregate Very Severe Aggressive Behavior, 0.30 0.38 0.44 0.28 0.27 0.22 Aggregate Behavior Change over the Past Year Physical Behavior to Severe 0.62 0.93 1,15 0.97 1,04 0.97 Verbal Behavior to Severe* 0.99 1.44 1.50 1,39 1.69 1.60 Other Behavior to Severe** 0.42 0.58 0.72 0.69 0.77 0.88 ABS from Mild or None to Severe 0.57 0.70 0.85 0.67 0.77 0.68 ABS from Mild, None or Severe to 0.39 0.47 0.52 0.36 0.36 0.33 Very Severe Diagnoses Anxiety -1.31 -0.97 -1.19 -1.10 -1.48 -1.13 Depression -4.07 -3.91 -3.90 -3.20 -3.39 -3.71 Manic Depression (Bipolar Disorder) 0.79 0.99 0.94 1.02 1.18 0.91 Psychotic Disorder -2.47 -1.89 -2.03 -1.81 -1.83 -2.00 Schizophrenia -0,29 -0.35 -0.32 -0.81 -0.37 -0.77 PTSD*** 0.15 0.20 0.20 0.24 0.30 0.40 New Diagnoses over the Past Year Anxiety 0.90 1.51 1.38 1.17 1.47 1.52 Depression* 1.99 1.80 1.84 1.50 1.71 1.37 Manic Depression (Bipolar 0.59 0.79 0.64 0.64 0.77 0.67 Disorder) Psychotic Disorder (other than -0,39 0.45 0.06 -0.30 0.42 0.14 Schizophrenia) Schizophrenia 0.09 0.20 0.23 0.03 0.34 0.30 PTSD*** 0.06 0.08 0.09 0.08 0.11 0.12 Severe Urinary Incontinence -7.64 -8.17 -9.12 -9.01 -9.50 -8.90 Severe Bowel Incontinence* -3.46 -4.14 -4.78 -4.57 -5.42 -4.73 Total Dependency or Other for all 1.27 0.80 0.49 0.66 0.43 0.44 ADL (continued) C-13 Table C-3. Differences in Risk Factors for NF Residents (continued) Difference (Discharged Live-Not Discharged), % 2012 2013 2014 2015 2016 2017 Changes in [Impairments Urinary Continence to Severe 1.50 1.48 1.04 1.04 0.94 1.35 Bowel Continence to Severe 1,04 0.95 0.80 0.92 0.77 0.96 ADL to Total Dependency or 1.42 1.17 0.90 0.88 0.72 0.67 Other** CFS Mean Point Change (Range 0- 3 points) 0.086 0.079 0.080 0.077 0.077 0.079 */**/*** = Difference is statistically significant based on a p-value cutoff of 0.05/0.01/0.001 Figure C-3. Difference in Bipolar Disorder and PTSD at End of Stay 1.40% 1.20% 1.00% 0.80% 0.60% 0.40% a 0.20% harged live and not discharged ISC 0.00% Percent difference between residents d 2012 2013 2014 2015 2016 2017 - Bipolar ---PTSD*** */** /*** = Difference is statistically significant based on a p-value cutoff of 0.05/0.01/0.001. C-14 Figure C-4. Difference in Urinary and Bowel Incontinence 2.00% ee 0.00% -2.00% -4.00% LL -6.00% -8.00% Se Percent difference between residents discharged live and not discharged -10.00% 2012 2013 2014 2015 2016 2017 2012 2013 2014 2015 2016 2017 Symptom Present at End of Stay Symptom Developed in the Past Year - Bowel Incontinence* = Urinary Incontinence Change in Bowel Incontinence ====Change in Urinary Incontinence */**/**X = Difference is statistically significant based on a p-value cutoff of 0.05/0.01/0.001. Figure C-5. Difference in the Most Severe Level of Dependency for ADL 25 1.50% o 0 0D m @ 1.30% 25 c 2 1.10% O 2 2 0.90% Zo ® § 0.70% eo ® 2 0.50% Lo 5 & 0.30% € & 2012 2013 2014 2015 2016 2017 2012 2013 2014 2015 2016 2017 oO 5 2 Symptom Present at End of Stay Symptom Developed in the Past Year oO -- Most Severe Level of Dependency in ADL* = Change to Most Severe Level of Dependency in ADL** */**/*** = Difference is statistically significant based on a p-value cutoff of 0.05/0.01/0.001. C-15 Appendix D: Data Tables for State and Facility Characteristics (penuluos) BIBABS 0} SUON 46T T9O0 TET 9EO 82T ved O40 9240 4b0 O80 eF0 LZL0 E50 E€2'% SET B8bO- SLT 40 PIIW Wo. Sav BJBAVS cot €S0 ELT TRO €9T TSO 8Z0 950 O6T L402 Ee€T BHT 690 69% LEO g882- 720Z 0} 4OlAeYyag 19410 BJBAIS 9272 v6T E27 OFT Let SET PLT OFT *@ST Cet sTT Ove LT 99b T6% SGHO9- CTE 0} JOIAeYad [EQU9A BIBAVS Guz €70 T9T EST ZT £4270 OOT 980 ZrO 920 TEO 90T 860 PEE SET "ZT €0'E 3 4OIAeYag JBDISAUd 1B@9A JSC, 94} J9A0 DHueyD JOIAeyoOg ayebaibBy 'solaeyog aaissalbby 0Z'70 )6€©609'0) Ss Z@EO0-Cs«é#PBT'O:-S- 8Z'0-s SS"0-séOS0-CsésaTZ@'~=ss« GE"O-C*GT'O- ~vS'O) =6S8'0) Ss bO'0- «S280 90'0--s«éOOwW'Z-_-s- G00 auaAss AlaA a7ebaibby 'uolAeyag eve | 06TLZ00C CT O90- «PO «6cSO- 400 «6©€S0)068z0- «9ST O- 70 «#20F0 «6900 «Tze LZTO- 6ZT- 8S'T aAISseIBBy a19A9S 49410 OT'Z c30 Z49°T Svt- soz zT'0 ZT0- 8£°0 CoC 98°0- 90°T blL0 po'oO- E82 ET'¢c- ZOS- EST "JOLAeYa B1BAaS JoIaeyag eve "Z45T €9% O60 S9E 69T ETT POT SEO *vZ40- 640 TSE 890 brS GET TrE- POE JEQ48A SABA8S JOIAeYyag Te Br O- O8T BET vOT Z4z0- £€9°0 990 ZT O- OvO0- SZO0 s90 SEO OFF FvZ0 TO0'Z- PTE [2d1SAYd B18A9S JOIARYOg SUJUOW 9 JSP] Ul Ayqi6ye presipayw 96°0 Ob'T oO9'T- €8'0 OVO T6'O0 O6E 868 ELS BT'S Sse 9Z'E TST "4H'S 9ST'% 60°% OFT 0} UOHISUeA] SUIUOW © SP] Ul AyqiBye presipayw st'O0- 820 7¢S'€ 8sT'O vp8Tt €2°0 95°70 OOF SO? pre LTT %TZ'O0 OS'O %72'T SEO 90°0- Le'T 0} UOHISUBAL AeS Jo pug ye Aq Ze'9- ZE'9- 6'6- LL'ET- 66'7T- 90°F- B8bTT- 8T'b B8b'OT- 68'bP- EO'TT- 62°8- Ec'bpT- 90°6- 60°8- EE'OT- LE'9- PledIPEW [INA snjejs jeng Sy VI NI WI aI IH v9 Td od id Lo (ojo) wo uv ZV AV TV (poebseyosig JON-2AI] pebseyosiqg) sdusa19jJ1G Lt0z '(s)-1v) $9}e81S Aq 'sjuapisoy IN 40} S10}0e4 ISIY Ul SOOUSIDJIG JUlOd 26e]uU99190q "T-qd 91geL N 1 a (panuljuo3) aUaUNnUODUT Z6'T- 82Z°9- 9T'9- QZL'Z- BZL'9- GO'S- PE'O- BT'S- OS'2 LO'E- O6'6- 66'6- 66'Z- 82'0 90°F SGPH'TZ- 86°0 JeMOg B1BA9S BDUBUIUODUT 68'9- 9b'ZT- PS'TT- OT'Z- OP'ZT- BST'6- B86'C- O09°9- BOT 9EZ- GE6'CT- @H'ET- TZ°OT- TL'O- ZZ2'8- 60°ZZ- 8T'?C- AJQULIN B18A98S syusweduy Z40°0 §=6st'O.) 0=-(000-Ss«ésTT'O)=3-hfb9'0Ss €0'O- «=22Z0'O0)=SdzTT'O)3F-l Z@T*0- «=220'O)=S- G6 E"0-CsEZO-C*OT"0O-«CéPBT'OC ETO. Ss 2Z"0--s ET'0- aS1d Ovo 6 TTO «69T0 6 «6627006400 @S'0- ~=T2Z0- O40 68'0- SPO- STO SO0';O 60 O20 Oc d- 2z'0- ZS'0 ejuatydoziyas (eluasydoziyas uey} 42410) ss'0 §699v'0- O2'0- Z42°0 OS'0 G6rO- ZZ'0- 69'0 62'0- PT'T- 99°0 92°0- vrO 82'0 €2°0 OT'O SEO JepJOsiq IOUIASd (Japsosiq sejodig) T60 #SO';O €9'0 E40 Sz'Oo- 8r0- E770 99'0 g85°0 09'0 Z2ro 9T'0 S80 EO ZZ'0 O0'O0) O0'T uo|sseidaq J]UeW 69T TOO- O@T 4@T "LbT SE?C- 90T ECC Che CEO ST £90 v60 €6'0 TT'? 8s89°€- E69'T uoissaidaq v6O OOT 80T 69T 99% 800- 60T 4272 vod T9E Tet 920 S86T T6O0 64T SET- OFT Agaixuy 429, JSeq 9y} 19AO0 SosouBbelg MON TSO §6©6TZ'O0)0=-dzTZ7O-s-«CSE*O-s«*OOG'O-'éEE'O- «=s GE"O--Cs-C*STEO-sCsEO'O- «=EZ'O0 )=62SO'T) SS 82@"T) = =-@2'0-s«*OZ0-«édE'O---s HB'Z- -SsCOOEO aS1d 99°0- cvO- G6O'T- 92°¢2- 9OT'T- v8'O- Z6'T- 6S'0- €0'P- G6T'c- TrT- 69'0- ZTE ZCO'T- E€0°S- PTC 8E'T- ejuasydoziyrs c8'T- 6b'T- 99°C- PpZ'T- CO'T- 98'0- 68°C- Pp6'T- ZLZ'E- CE'E- QE'T- B8B'O- OE'T- Tr'c- 9E°Z- GO'T- 9'T- 4apsosiq d1}0UDASd (4apsosiq sejodiq) 690 v£eO0 8240 T60 v6'0- SZT- poo 680 s00- 6710 O90 69°70 940 80 GZ0- S9T- 02470 uolssaidaq s1ueW Z4S°T- v2@'€- CTT- O€ E€- BS'E- CZ6b- E€99- Ob POE- OL T- OTS- pET- GOZ- VEO LG L97Z- T9I'O- uolssaideq Z9°0- pee- p60 98'0- OT'T- PpOT- Z6E- ESO- BL'E- LYO- Ef€- LET- LET B8EO- ECT- BIT ZCO'T- Aaixuy sosouBbeig auaaas Ala; 0} BUBAVS JO SUON 920 6S5'0 sGzOo ¢cf€0 céo0 £90 €80 9270 00';0 9T'O0- 95°0 T6'0 9T'O0 €6'0 9E°0 Z6'T- TO "PIIW Woly Say Sy VI NI WI aI IH v9 Td od id Lo (ojo) wo uv ZV AV TV (paBseydsig JON-2AI] pebseyosig) sdus19jJIG sajeis Aq 's]uapisey AN 10} $10}9e4 Sly Ul SODUdJaYIG JUIOg 96e}U9910¢q (penunuos) ZT0z '(Sy-TVv) *T-d SIqel D-3 (sjulod ¢-0 eBbuey) eHueyD TT'O }3=69€"0- ~3S0'O)3=6-OT'O0sC*#9E"0--=-«~TO'O.-Ss«COT'O0-s 80'0--s-«éoT'O0~-S G60'0-sZO'0-sazTt'0=-s Z0'0--CéC*éaTT''[ =--« 6d'N-«#BN'O.-s ZT'0 qlod UBBW SAD 9TO Z0°0- p20 95°0 ZO'O- SEO- BSZT vTT She 620 OVO 220 sg0T STT 960 Z8T- SLT 82 93 10V B1BAaS 96T 79'0 860 €0'T 790 840 98T €0'% 65'T 69% TS'O- O40 vE'O TI'b 8€0 THE LI OF BQUBUI]UOD [eMog BIDAIS 0} 65°C v8'o 190 s8ET v8'O0 61T0- P8T £6 95° E80 €2'0 cot O60 e¢ve E€OT SbO SZC aouauju0D Aun) syusuedwy ul sobfueyD Sy VI NI IW aI IH v9) Td od id Lo (o}e) vo av ZV AV TV (poebseyosig JON-2AI7 poebseyosiqg) sdus19jJIG sa}]e1S Aq 's]uapisey AN 10} $10}90e4 4SIY Ul SODUdJBYIG JUIOg 36e}U9I10q (penuljuos) ZT0z '(S>-1V) *T-d 31qeL vt a (penuluos) BIBADS 0] SUON 6r0 sT'O- 8rTt sso 720 T2TtT 400 Z2Tt g0% vOT SHO brO SOT ST'O- 95°09 LOT ZOT 40 PIlIW WOW SAV BBAVS zeo. 06 TT'O- «€©TT?)~=6E9O0 COST CO88T OUT8O0)0 CEST OSC) =6Z60_- KONC*OT TT) OLB OCOO- «=-8c'c S60 HOT 0} JOJAeyag 49410 BdaAaS Ect @T'O- 96° S2TtT ZTT OL OOT O6'T SPE S2t G6rT LZyT 69'T vS'O SE'O SOT GET 0} JOlAeYag [ECHaA B4BAaS Of0 prO- 720 SOT O40 S07 £0 "4Z8Z 90E PST B800- ZrO EO Z2TO0 900 9ST FZ'0 9} 4OIAeYag JeIISAUd 4129, JSlq 94} 19A0 BHueys 101Aeyog aqyebaibBby 'solaeyag aalssaibby co'0- ~OT'O- @TO SzoO T80 sOO- STO- 65'T E80 9b0 %vO0- 620 6590 vod STO 620 TT'O aaAag Ala, aqebaubby 'uolaeyag T7'O 868€'0- OO'T ZS'°0 680 Z2T ZO'T- GET Let vO LEO STO G6E'O0 €E2'0- 92'°0- 86°0 60'T aAIssatbBy a19A9S 49410 Ty'0- 8S'0- 80°0 9€°0 O92 9ST p8'O- wHT 99% 62T vS'O- £60 920 sp 0- 9270 T9'0 92'0 'JOIARYag B12A8S JOlIAeuag sot os'o- Zee ZS'T TZ'0 OTS SE'O- OTC Ev TSC SST 880 6S9°T 95'0 v9'0- B8Z'T O8'T JEqJ8A S48A98S Jo|IAeyag 90°0- 68'0- p80 480 ZdE'O 2O'C E4'0- LEE 92'E SGS'T ZS'T- 90'0- 80'0 v2'0O- v9'0- TZ'T 82'0 [2d1SAUd B18ABS JOIAeYuog SYUJUOLU 9 4S] Ul AQuiqi61j8 plesipew O67 TSh 69T ZEST SOE TZ'8 86% 819 OFC YOGE SBE Cb EFC BCT SEL LBP GED 0} UORISUBIL SYJUOWW £ JSP] Ul Aynqi6i@ plesipaw os'tT 82'T cO;O v0'O 90'T 85h 80°70 Occ cro OS'O0 Set vet EeT'0- OVO 62°F OZ'T Z5°0 0} UONISUeIL Aeqs Jo pug ye Aqiqibye T2TT- T6°8- CS'OT- v6'6- LLET- 66°2T- 90'F- Br TT- 8T'b- Sr OT- 68'F- EOTT- 62'8- E2@ PT- 90°6- 60°8- CE'OT- PIEIPEW I1N4 snjeis jeng ON AN WN CN HN AN JN LW OW SW NW IW VW aw AW v1 AM (poeBbseyudsig JON-2AI] pebseyosig) sdus19jJIG £t0z '(DN-Ay) soje1s Aq 'swuapisoy -IN 10} S10}2e-+4 Js Ul SodUSJaIG JUIOg B6eus2zI9q = "Z-d JIGeL D-5 (penuuos) @DUBUIUODUT S6'S- TE'9- GT'8- S8'O- O8'S- cS°Z- €0°9- ZT'6- Z@BE- 69% TC6- Z9'E- CG'Z- v2'8- LT'bT- 96°T ZS'0 J8MOg B1BA2S @9UAUIJUOSUT vS'8- OTTT- T9°8- Z6'7- CEET- 6VE- PE?CT- 66°CT- 62°9- OTT ZE9T- L9°6- QZ TT- @S5°2T- pZ'8T- E€S'0 8S°C- AJQULIN B418A98S syUSsUWIeduiy vo'o )§=6Z0°0-S ss «OOO-C'és«SCQ'D- «ss ZO'O- «= CE'O=SC ETO SCO p'0sCZZ'0-CsésdCO-Ci-="os'0) Ss SZ'0-s-C«éd6Z'0.-« PBT'O=-s ZO"O---s- 80'0~-s_-«€=000 aS1d Ze'0- TO'O- 9T'O- TO;O £€£0°0 SPr'O- 90°0- 2'°0 7zS''0 O0°';O s£'0 9'0 92'°0 99'0 82'0 S'0 ¥0'0- ejuatYdozIYyIS (eluasydoziyos UB} 18430) 97°70 §=6-os'0- ~=6E'0- ~TO'O- 46'°0 vE'O- cS'0- ZO'O- 6T'O- OVO €€'0 O20 8sT'O- 6T'O- 90°0- ET'O- T9'D- Japsosiq 20Y9ASd (4apsosiq sejodig) £80 TS'O OTT O@O €20- €6T E0T TOO E@T S60 sso e80 950 90 ETO vS'0 6€°0 uo|sseideq a1UeW 69° s0T 49% Trt TT'O- 822 ste 2cZTt o90 O40 90% 960 eST 8c tT SES G2ZTt 90°T uo|ssaidaq Tec p60 £60 CZT E€0T TZ? T9T OTE ODOT %CTO- CTT Tre vet so0- 670 ZIT OT Ajalxuy 423d, JS2q 9y} 10AO0 SosouBeIg MON STO =6hpT'0-sCZT'O--CsCd C's Z8°0--sz@T'O-S so SOOs«O9ZT--_-sG6'ON-sC«édTO'O = Z@E'TSs BH'O.CCGSE'T-iCéEO-CiasiaGO-CéC«éTT'D-s«COPE aSld T9'0- T6'Z- vZ'E- LvE- 96'0- OO'T- 99'T- Z6'T- 98'0 OZ'T- TO'O- HT'O- O8'O- 80'O T9'0 9b T- 8Z'T- ejuatYydoziyss S8'T- 6€°2- 96'C- GO'T- TO9'E- E€0'O T9'C- QZLZT- LZ9'T- Lv't- sgat- 90'C- GTE- YOT'C- HT'?- CH'T- TI'E- 4OPJOSIG IIJOUDASd (4apsosiq sejodiq) bro 8 690':0 )3=€6-0s'0 Ss ZE"O-S-COOON-CsGT TFT) «368600 orth COewE 6 60CB8LZT OCO6T SCOTT) OL 990-CsEG'O-SCET'O- 80 uolsseidaq s1ueW TG'?- O@Z- 82°9- GEC- BbS- OZE- POT- 620 B9°%- Ly?- 99T LZ2S- THS- p2S- vET- CS ?- BT'E- uolssaidaq 66'T- 68T- pOE- ECT 69°E- OOT 49 %- S60 gsS°0- 890- EFC ZEC- TrT- 86T- €80 Z6°0- 8S'T- Ayaixuy sosouBbeig auaaes AlaA 0} BJDABS JO SUON soo 6©€0'0)«€69E'0) |- Z270--s 8Z'0-sCé#*OOT0- «=62'0-s-sC«dE ET = «6Z'0-s 8@'O-C'=*TO#"O~=CsCé#STE'O=-s-«étOZ0N-CiéT'-«=s ZE"O-C'TEEO-s« 810°0 "PIIW WO. Say ON AN WN CN HN AN JN LW OW SW NW IW VN aw AW v1 Ay (paBseyudsig JON-2AI] pebseyosig) sdus19jJI1G (penunuos) ZT0z '(DN-A») saj1e1s Aq 'sqUuapisey AN 10} $10}9e4 JSIy Ul SODUdJAYIG JUIOg BHezUSII9g «=-Z-q BIGEL wo a (sjulod €-0 s6uey) sbueyD 8006 40'0-~=- 60'0--Ss 80'0--S ss zTT'O)=36ctO0) 06900 )3=-oTO0 SC 60'0--s-CéioTO0 SC f00S 400s 4O'O-sC=990'0--s wON-C'=#wB'O-Cé'*GSCV|D qwlod UBSW SAD 900 040 2TO0 6TT %ZL0- TZ0 920- SPO 90 LZLT TZ O- 98°09 SOO T2O0 §€S'0- E2T 96°70 82 9) 10V a1BA0S SO'Z Z22'°0 OT'Z@- 9O0'E ELT Z2P'O- OZ'T 9O'T- TOT 72S'2 9£0- p90 LT'Z 60'0- ZZ'0- 9'% GS°E 97 BOUBUIRUOD jeMog BIBABS 0} wZe@ 86€S'0)0 = (OZ 0- ~O9' «=6-LZO- «62% ~= «609 T: CTET)h6UETT)«36U 88°C) SC oT 0 CO ZOT) =669T) =660T) «6666S COLETTE soueunUOD AIeULN syuswuiedwy ul sobueyD ON AN WN CN HN AN IN LW OW SW NW IW VN aw IW vi AM (poeBseyudsig JON-2AI] pebseyosiqg) sdus19jJIG (penunuos) ZT0z '(DN-A») saje1s Aq 'squapisey AN 10} $10}9e4 4SIY Ul SODUdJaYIG JUIOg BHezUSII9g «=-Z- BIGEL D-7 (panuluoz) BIBADS 0} SUON ve'O- 89°0- 6T'O €S'0- T6'O0 T9'2- 480 LZTT TET 95'0- €0°0- LZZ'C 820 G2'O- 2@T'E Z2°0 Z2Z'O- 40 PIlIW WOuy SAV DIBAIS £9'0- O€'O- 6T'O0 ST'O- 4670 Z9'T- O@'2 cht 98'T O9'0- Z8'O OTE 92'0 ZE'O- SE'E T9'O0 OT'O 0} 4olAeyag 18410 BIBAIS Ze'T- TO 60'T 6O'T OZ4T ZPO- OO'E SBC OF? SGT'O- SPO B8T'h p20 2TI'T OLb Ev'T ¥S'0 0} JOlAeYag JECHaA BIBAVS ve0- T6O- ZITO 400 EO G6S'T- ZLT 67% SEZ GOO PTO 9S TOO- 6O'O- OE TSO EF'0 9} 4OIAeYag JeDISAUd 1B29A JS 94} 19A0 BHueyD JOIAeyOg aqebaibBby 'solaeyag daalssoibby ZT'0- sO';O- dO';O- T20- ST''O 89'0- 9T'T- TyO 8990 T9'0- SE'0- OS'O 2@z0 eEv0d- O@T Z40°0 sg2°0 aaAes Ala, aqyeboubby 'aolaeyag c6'0- 8z'T- TO'O- 6rT- TSO E€9'E€- G6T'O- ST'T ZE'T 6Z4'0- TS'O- 88'C zTt'dO T9'0- 6T'E TS'O- ET'O- eAIssetbBy 319095 49410 6y'T- pt'c- p9'0- L9T- E80 ZEb- TS'O- 8T'T 99'T Z8O0- €270 98°C 820- PIT- POE Y9P'0- HTD "OIARYSG B12ABS JOIAeYyag 6c'c- ET'T- 880 OS'O- Z6'T 9S'E- 2480 6T'E 8BS'C 88'0- 66°0- 60'S 92'0 €£'0 62Z'b 86'0 98°0 JEq48A B48A9S JoIAeyag 6by'T- Z6'T- 98'0- bZ'0- ZE'O vT'E- 7270 Z4EC SE? ZCL'O- SE'O- ZS'€ ZS5°0- vO'T- ZOE E20 T9'O [2d1SAUd B18ABS JOIARYOg SYUJUOLU 9 4S] Ul Aquiqi61j8 plesipew os'é S09 T8b 629 6rE 8T'S TdOT SBE EC BHT cece L6T ED PTS 8sB9T OVE ELZ 0} UORISUBIL SYJUOWW £ JSP] Ul Aynqi6i1@ plesipaw TO ceo s86T €90 Ee€v0 O92 T62 Set L490 E20 6€0 £20 p90 O90 OFO0 SOT EBT 0} UONISUeIL Aes Jo pug ye Aqqibye vE'CT- ST'OT- 69°2T- TS TT- 6E°OT- PLZOT- OT'O ST TT- SE°CT- CT'LZ- LT ET- PB CT- ES'OT- VECT- Ch y- LOG6T- PT'TT- plesipew [Ind snjejs eng AM IM AM VM VA LA in XL NL as oS Ta Vd uo Oo HO aN (poaBbseyosig JON-2AI] pebseyosig) sdus19jJIG saj1e1g Aq 'squapisey 4N 10} $10}9e4 YS!IY Ul SODUdJaIG JUIOg 96e}U9910¢q Ltoz '(AM-GN) "e-d sige a (panuljuo3) aouaUNnUOsUT 6S°2- Ty'OT- €8°C- S6'ET- SE*P- OGS'OT- 66°2T- TS'Z- 96'P- G6rT- TH9- LO'CT- 96°9- LO'OZ- 8rT v0'6- ZT'L- JeMOg B1BA9S BIUBUIJUODUT 88'Z- €bv'6T- 6T'9- P6'ZT- bO'B- OG'ZT- OT'OZ- 6S°b- 69°9- C6OT- OS'Z- 8Z'ST- EE'ET- HT'@Z- GG'T- G6T'bT- TT'ST- AJQULIN B18A98S syusweduy ZT'O- ZT'O TO'O- OS'O ZT''O 9S5'0 90 Z0'O 6T'0 ZO0'O- SO'O- Z4O'O 2T'O €2'°0 80°0 220 v2z'0 aS1d oo;o)§6sz'0-s-s«éeETO- ~«220'0)«=6-0E0-s-sCédTO- «=SCTE'O)=-2z'0- =2EO0)S=CZZ'0-CiGE'O--CéTO =F OT'0-i'"EV@WO--=«sOOS"0- ~=~CTEO = ZG ejuatydoziyas (eluasydoziyas ueyy 42410) 8T'O 7zs'0- GSbp'0o- 99°0- O£'0 Z2Z'0 Z2'0 ST'O ShO 9S5'0 9T'O- 6T'O- O0'O 9P'0- SS'°O0 SEO ZE'O- JepJOsiq IOUIASd (4apsosiq sejodig) 6c'0- Tr'O S80 sgrO0 co'o- 729'0 7290 89'0 9E'T vE'O cot 8T'T 9€'0 ZE'O ZrO 909'0 80°0 uolssaidaq J1ueW T90- 49T E€8'E coc SEC 820 480 s8O0 sre 624T Z6T O8T @ST OFT OTT STT Tz'0 uolssaidaq 68T 6€0 Zee p9T OT 6TC 62T SOT L4EC L240 98T bLO0 6HYT COT vert 68T 6r'0 Agaixuy 429, JSeq 9y} 19AO0 SosouBeIg MON ps0 §=68S'O) = SS*0s-Cé*OS'O-C=sGSH'O-Ci=éT"''Z~=-=-<i'EOT OZT'O. SCOL'O0s GT'O- +~60°0- ZO'T 2E'O T9'0 cro 820 T9'0 aS1d Z6'T- SO'O 9Z'T- OZT- £9°0- 96'2- ZT'O- 86'T- Sh'O- CTC O€'T- €9'0- PO'T- 68'C- 89'0- 66'T- SPH'T- ejuasydoziyas ZO';O)3=- 69'C-_ B86'C- BHE- EC'?2- OQLT- CT'T- Gz2'c- TO'E- S9'O- ZS'2Z- PE'T- LY E- EO'E- EL'O- GT'b- 8C'T- Japsosiq d1j0UDASd (4apsosiq sejodig) e9'0- €f&'T 9690 SpyO- 620 SS'0- €€'2 O20 OFT vot TOO e"4Tt ODOT 9O0T 880 cé60~ vE'O- uolssaidaq j1ueW L4y?- 6&'?- 840 bZ?@- OT T- CLLZ- Thb LZ8b- 7@L9- GSTT veb- 92'b- EVE- LLb- 640 SS'S- GED- uolssaideq 690 O6°2- O27 v80- OTO- ET'?C- 89S B9'2- 809- EVT TSE- TSE- T2T- 6TT- TT? g80°2- SS'0- Ajaixuy sosouBbeig a1aaas Ala, 0} SJDABS JO SUON 90°70 4O0;O0 TT'O0 90°0 T2'0 «S€'O- S90 ZyO 690 9F'O0- cO';O- EVO OED Sc'O- O@T OF0O0 S2'0 "PIIW Wo. Say AM IM AM VM VA LA Ln XL NL as oS Tu Vd uo > [e) HO aN (poaBseyudsig JON-2AI] pebseyosig) sdus19jJIG sajeis Aq 's]uapisey AN 10j $10}0e4 4S!Iy Ul SODUdIAJJIG JUIOg 96e}U9910¢q (penunuos) ZT0z '(AM-AN) D-9 "E-d s1gel (sjulod ¢-0 eBbuey) eHueyD soo ¢€0'0 sO';0 40°0 60°0 Z40°0 sO';O0) 6€6©80':0)6«6©660'0)6€640':0)6Z0'0)62@T'0-Ss(90°0-s 80'0-Ss ZT'0.-s 80"0_~-s' E100 qulod UBBW SAD ect 9€°0- 490 Sz0- 9F0 TOT- TOO O40 90 v20 86T ETO B8SrO §E9'0- 420 22';0 SGZ2T 8z 9 10V BIBAaS 880 «(O8S'T- 8ht ZZT- 62'T 96'0- vE'O OZT Te'? SGT vET L6°0- ZT'O O8'T- CER e2'0- LZG'Z 0} BoUBUIUOD jeMog BIDAVS OF OZ'T- TZ'0- G6'% zcro- 8STT Z8T €60 £0 TST v8'0- OTC ETT SET ETT- 82% YbO'O- OSD aouauljuoD AeUuL/) syusuedwy ul sobfueyD AM IM AM VM VA LA in XL NL as 3S Ta Wd uo > [e) HO aN (poabseyosig JON-2AI7] pebseyosiq) sdusd19jI1G sa}e1S Aq 's]Uapisey AN 10} $10}90e4 4SIY Ul SODUdJAYIG JUIOg 96e}U9910q (penuluos) ZT07 '(AM-AN) "E-d s1gel 10 a Table D-4. Percentage Point Differences in Risk Factors for NF Residents, by Profit Status, 2016 Difference (Discharged Live-Not Discharged) For-Profit Non-Profit Government Dual Status Full Medicaid eligibility at End of Stay -12.89 -11,19 -10.65 Transition to Medicaid eligibility in last 3 months 1.12 0.87 1.35 Transition to Medicaid eligibility in last 6 months 3.84 3.02 2.60 Behavior Severe Physical Behavior 1.00 -0.30 0.39 Severe Verbal Behavior 1.68 0.42 0.53 Severe Behavior, Other 0.40 0.13 -0.26 Severe Aggressive Behavior, Aggregate 0.50 0.06 0.34 Very Severe Aggressive Behavior, Aggregate 0.23 0.31 -0.01 Behavior Change over the Past Year Physical Behavior to Severe 1.19 0.30 1.16 Verbal Behavior to Severe 1.83 0.86 1,22 Other Behavior to Severe 0.98 0.58 0.80 ABS from Mild or None to Severe 0.78 0,29 0.77 ABS from Mild, None or Severe to Very Severe 0.32 0.38 0.22 Diagnoses Anxiety -1,25 -2,11 0.72 Depression -3.96 -3.76 -1.46 Manic Depression (Bipolar Disorder) 0.62 0.69 1.05 Psychotic Disorder -2.47 -1.42 -1.29 Schizophrenia -1.55 -0.25 -1.37 PTSD 0.43 0.26 0.28 New Diagnoses over the Past Year Anxiety 1.55 1.15 1.78 Depression 1.33 1.59 1.12 Manic Depression (Bipolar Disorder) 0.66 0.45 0.75 Psychotic Disorder (other than Schizophrenia) 0.13 -0.01 0.14 Schizophrenia 0.23 0.18 0.31 PTSD 0.12 0.11 0.21 Impairments Severe Urinary Incontinence -8.15 -10.82 -7.93 Severe Bowel Incontinence -4,96 -5.74 -2,58 Changes in Impairments Urinary Continence to Severe 1.49 1.12 1.50 Bowel Continence to Severe 1.14 0.76 1.68 ADL to Total Dependency or Other 0.64 0.67 0.92 CFS Mean Point Change (Range 0-3 points) 0.081 0.073 0.086 D-11 Table D-5. Percentage Point Difference in Risk Factors for NF Residents, by Rural and Urban Facilities, 2016 Differences (Discharged Live-Not Discharged) Rural Urban Dual Status Full Medicaid eligibility at End of Stay -9,50 -12.06 Transition to Medicaid eligibility in last 3 months 1.02 1,08 Transition to Medicaid eligibility in last 6 months 3.44 3.60 Behavior Severe Physical Behavior 0.98 0.59 Severe Verbal Behavior 1.96 1.27 Severe Behavior, Other 0.82 0.23 Severe Aggressive Behavior, Aggregate 0.86 0.32 Very Severe Aggressive Behavior, Aggregate 0.46 0.19 Behavior Change over the Past Year Physical Behavior to Severe 1.29 0.94 Verbal Behavior to Severe 2.03 1,54 Other Behavior to Severe 1.30 0.80 ABS from Mild or None to Severe 1,02 0.63 ABS from Mild, None or Severe to Very Severe 0.56 0.27 Diagnoses Anxiety 0.23 -1.28 Depression -1,28 -4,18 Manic Depression (Bipolar Disorder) 1.36 0.77 Psychotic Disorder -1.36 -2.16 Schizophrenia -0.46 -0,94 PTSD 0.45 0.40 New Diagnoses over the Past Year Anxiety 1.92 1.41 Depression 1.56 1.30 Manic Depression (Bipolar Disorder) 0.64 0.67 Psychotic Disorder (other than Schizophrenia) 0.45 0.06 Schizophrenia 0.23 0.30 PTSD 0.13 0.12 Impairments Severe Urinary Incontinence -8.03 -9.34 Severe Bowel Incontinence -3.53 -5,52 Changes in Impairments Urinary Continence to Severe 1.87 1.27 Bowel Continence to Severe 1.75 0.79 ADL to Total Dependency or Other 0.64 0.67 CFS Mean Point Change (Range 0-3 points) 0.08 -0.32 D-12 Table D-6. Percentage Point Differences in Risk Factors for NF Residents, by Chain Status, 2016 Difference (Discharged Live-Not Discharged) Not Chain Chain Dual Status Full Medicaid eligibility at End of Stay -10.14 -12.45 Transition to Medicaid eligibility in last 3 months 0.74 1,28 Transition to Medicaid eligibility in last 6 months 3.17 3.78 Behavior Severe Physical Behavior 0.49 0.78 Severe Verbal Behavior 1.04 1.61 Severe Behavior, Other 0.25 0.39 Severe Aggressive Behavior, Aggregate 0.37 0.46 Very Severe Aggressive Behavior, Aggregate 0.21 0.26 Behavior Change over the Past Year Physical Behavior to Severe 0.93 1.02 Verbal Behavior to Severe 1,39 1.76 Other Behavior to Severe 0.82 0.93 ABS from Mild or None to Severe 0.69 0.71 ABS from Mild, None or Severe to Very Severe 0.31 0.35 Diagnoses Anxiety -1.22 -1.36 Depression -4.40 -3,54 Manic Depression (Bipolar Disorder) 0.87 0.88 Psychotic Disorder -1.68 -2.28 Schizophrenia -0.38 -1.02 PTSD 0.33 0.43 New Diagnoses over the Past Year Anxiety 1.42 1.52 Depression 1.26 1.42 Manic Depression (Bipolar Disorder) 0.68 0.67 Psychotic Disorder (other than Schizophrenia) 0.15 0.12 Schizophrenia 0.41 0.22 PTSD 0.11 0.13 Impairments Severe Urinary Incontinence -8.02 -9,52 Severe Bowel Incontinence -3.46 -5,75 Changes in Impairments Urinary Continence to Severe 1.45 1,27 Bowel Continence to Severe 1,04 0.87 ADL to Total Dependency or Other 0.92 0.52 CFS Mean Point Change (Range 0-3 points) 0.081 0.077 D-13 Table D-7. Percentage Point Differences in Risk Factors for NF Residents, by Facility Size, 2016 Difference (Discharged Live-Not Discharged) Small Medium Large ($100 beds) (>100-200) (>200 beds) Dual Status Full Medicaid eligibility at End of Stay -11.18 -12.05 -9.50 Transition to Medicaid eligibility in last 3 months 0.96 1.18 0.87 Transition to Medicaid eligibility in last 6 months 3.02 3.98 3.06 Behavior Severe Physical Behavior 0.58 0.82 0.03 Severe Verbal Behavior 1,82 1.46 -0.05 Severe Behavior, Other 0.67 0.33 -0.49 Severe Aggressive Behavior, Aggregate 0.71 0.43 -0,33 Very Severe Aggressive Behavior, Aggregate 0.31 0.25 0.02 Behavior Change over the Past Year Physical Behavior to Severe 1.06 1.04 0.56 Verbal Behavior to Severe 1,92 1.68 0.59 Other Behavior to Severe 1.15 0.85 0.42 ABS from Mild or None to Severe 0.89 0.72 0.19 ABS from Mild, None or Severe to Very Severe 0.43 0.33 0.11 Diagnoses Anxiety -0.83 -1.45 -1.33 Depression -2.88 -3.90 -5.46 Manic Depression (Bipolar Disorder) 1.05 0.84 0.81 Psychotic Disorder -1,66 -2,16 -2,24 Schizophrenia -0.14 -0.93 -1.39 PTSD 0.41 0.42 0.31 New Diagnoses over the Past Year Anxiety 1.56 1.51 1.27 Depression 1.39 1.29 1.58 Manic Depression (Bipolar Disorder) 0.67 0.65 0.73 Psychotic Disorder (other than Schizophrenia) 0.29 0.11 -0.06 Schizophrenia 0.23 0.29 0.51 PTSD 0.15 0.11 0.13 Impairments Severe Urinary Incontinence -9.80 -8.74 -7.53 Severe Bowel Incontinence -5.18 -4.94 -3.13 Changes in Impairments Urinary Continence to Severe 1.08 1.49 1.29 Bowel Continence to Severe 0.59 1.15 0.96 ADL to Total Dependency or Other 0.47 0.68 1.11 CFS Mean Point Change (Range 0-3 points) 0.08 0.08 0.08 D-14 Table D-8. Percentage Point Differences in Risk Factors for NF Residents, by Number of Ownership Changes, 2016 Difference (Discharged Live-Not Discharged) 0 changes 1-2 changes >2 changes Dual Status Full Medicaid eligibility at End of Stay -11.10 -12.33 -13.16 Transition to Medicaid eligibility in last 3 months 0.76 1.12 1.40 Transition to Medicaid eligibility in last 6 months 3.13 3.48 4,17 Behavior Severe Physical Behavior 0.09 0.93 1.16 Severe Verbal Behavior 0.60 1.71 1.89 Severe Behavior, Other 0.10 0.49 0.43 Severe Aggressive Behavior, Aggregate 0.13 0.65 0.52 Very Severe Aggressive Behavior, Aggregate 0.20 0.20 0.33 Behavior Change over the Past Year Physical Behavior to Severe 0.60 1.16 1.30 Verbal Behavior to Severe 1.04 1.87 1.95 Other Behavior to Severe 0.68 0.97 1.02 ABS from Mild or None to Severe 0.51 0.80 0.81 ABS from Mild, None or Severe to Very Severe 0.29 0.31 0.41 Diagnoses Anxiety -1.88 -1.08 -1.01 Depression -4.23 -3.25 -4.06 Manic Depression (Bipolar Disorder) 0.82 0.73 0.62 Psychotic Disorder -1.68 -2,14 -2,89 Schizophrenia -0.33 -1.18 -1.82 PTSD 0.31 0.38 0.49 New Diagnoses over the Past Year Anxiety 1.34 1.44 1.69 Depression 1.44 1.35 1.21 Manic Depression (Bipolar Disorder) 0.58 0.68 0.71 Psychotic Disorder (other than Schizophrenia) 0.04 0.15 0.14 Schizophrenia 0.29 0.28 0.16 PTSD 0.14 0.09 0.15 Impairments Severe Urinary Incontinence -9,02 -8,50 -8,.85 Severe Bowel Incontinence -4.18 -4,94 -6.07 Changes in Impairments Urinary Continence to Severe 1.26 1.38 1.54 Bowel Continence to Severe 0.99 1.09 0.99 ADL to Total Dependency or Other 0.66 0.70 0.67 CFS Mean Point Change (Range 0-3 points) 0.08 0.08 0.08 D-15 Table D-9. NF Residents Discharged Live, by Program Termination Status, 2012- 2017 2012 2013 2014 2015 2016 2017 All Residents Residents discharged from an 216,255 215,224 217,861 219,617 218,886 221,441 Discharged _ active facility 98.69% 99.07% 98.84% 99.18% 98.68% 99.19% NY from @ Residents discharged from a 2,481 1,675 2,253 1,491 2,672 1,631 terminated facility 1.14% 0.77% 1.03% 0.68% 1.21% 0.73% Residents Residents discharged live 1,962 1,276 1,910 1,298 2,412 1,390 Discharged on or before termination date 79.1% 76.2% 84.8% 87.1% 90.3% 85.2% ee rom 2 Resident discharged live 519 399 343 193 260 241 NE after termination date 20.9% 23.8% 15.2% 12.9% 9.7% 14.8% D-16 Table D-10. Prevalence of Risk Factors among NF Residents Discharged Live, by Program Termination Status, 2016 Residents Discharged Live, % Active Facility Terminated Facility Dual Status Full Medicaid eligibility at End of Stay*** 73.45 86.02 Transition to Medicaid eligibility in last 3 months 2.70 2.09 Transition to Medicaid eligibility in last 6 months*** 7.89 4.83 Behavior Severe Physical Behavior 4.72 4.66 Severe Verbal Behavior 8.60 9.93 Severe Behavior, Other 6.06 6.32 Severe Aggressive Behavior, Aggregate 4.25 4.97 Very Severe Aggressive Behavior, Aggregate* 1.07 0.49 Behavior Change over the Past Year Physical Behavior to Severe 3.71 3.43 Verbal Behavior to Severe 6.12 5.27 Other Behavior to Severe* 4.38 3.86 ABS from Mild or None to Severe 3.22 3.25 ABS from Mild, None or Severe to Very Severe* 0.90 0.43 Diagnoses Anxiety 31.57 33.60 Depression 49,38 50.28 Manic Depression (Bipolar Disorder) 7.27 8.52 Psychotic Disorder** 8.44 10.85 Schizophrenia*** 9.79 12.81 PTSD 1.16 0.98 New Diagnoses over the Past Year Anxiety 8.04 7.23 Depression* 9.17 7,11 Manic Depression (Bipolar Disorder)*** 1.78 1.04 Psychotic Disorder (other than Schizophrenia) 2.88 2.45 Schizophrenia 1.62 0.98 PTSD 0.28 0.43 Impairments Severe Urinary Incontinence*** 57.42 62.35 Severe Bowel Incontinence* 52.51 55.61 Changes in Impairments Urinary Continence to Severe** 10.87 9.26 Bowel Continence to Severe** 12.63 10.42 ADL to Total Dependency or Other* 2.48 1,66 CFS Mean Point Change (Range 0-3 points)*** 0.124 0.059 */**/*** = Difference is statistically significant based on a p-value cutoff of 0.05/0.01/0.001. D-17 Appendix E: Graphs and Data Tables for Outcomes After Discharge Table E-1. Outcomes of All Resident, by Timeframe After Discharge Days After Discharge, % 0-1 2-30 0-30 All Acute Care 43.13 9.66 52.79 Hospitalizations 38.54 5.79 44.33 Outpatient Acute Care 8.71 6.69 15.40 Table E-2. Hospitalization by Risk Factor NF Residents Risk Factor, % Without Time Period After Discharge, Days With o-1 | 2-30 |0-30| 0-1 2-30 | 0-30 Dual Status Full Medicaid eligibility at End of Stay 29.05 5.92 34.97) 42.11*** 5.75 47 .86*** Transition to Medicaid eligibility in 39.13 5.76 44.89] 29.87*** 7,.70*** 37,57*** last 3 months Transition to Medicaid eligibility in 39.71 5.68 45.39) 29.18*** 7.18*** 36.36*** last 6 months Behavior Severe Physical Behavior 37.66 5.78 43.44] 53.46*** 6.01 59.4 7*** Severe Verbal Behavior 37.58 5.71 43.29] 47.76*** 6.64*** 54,40*** Severe Behavior, Other 37.70 5.80 43.50) 49.68*** 5,74 55,42*** Severe Aggressive Behavior, 36.66 5.70 42.36) 50.35*** 6,40* 56,75*** Aggregate Very Severe Aggressive Behavior, 36.66 5.70 42.36) 55.50*** 5.88 61.38*** Aggregate Behavior Change over the Past Year Physical Behavior to Severe 37.71 5.80 43.51) 54.62*** 6.05 60.67*** Verbal Behavior to Severe 37.71 5.70 43.41] 47.94*** 7,00*** 54,94*** Other Behavior to Severe 37.75 5.79 43.54) 50.57*** 5.83 56.40*** ABS from Mild or None to Severe 38.06 5.76 43.82) 51.47*** 6.60** 58.07*** ABS from Mild, None or Severe to 38.35 5.79 44.14) 56.72*** 5,70 62.42*** Very Severe Diagnoses Anxiety 37.68 5.66 43.34) 40.37*** 6.08** 46.45*** Depression 37.38 5.66 43.04) 39.66*** 5.92* 45. 58*** Manic Depression (Bipolar Disorder) 38.41 5.66 44.07} 40.21** 7.65*** 47, 86*** Psychotic Disorder 37.78 5.85 43.63] 45.30*** 5,.28** 50.58*** Schizophrenia 38.18 5.74 43.92] 42.26*** 6,.39** 48.65*** PTSD 38.67 5.79 44.46) 23.31*** 5.86 29.17*** New Diagnoses over the Past Year Anxiety 38.19 5.77 43.96| 40.74*** 6.15 46.89*** Depression 38.30 5.76 44.06| 38.36*** 6,35** 44,71*** Manic Depression (Bipolar 38.43 5.75 44.18 | 41.95*** 7,74*** 49, 69*** Disorder) Psychotic Disorder (other than 38.10 5.83 43.93 | 43.90*** 5,39 49,29*** Schizophrenia) Schizophrenia 38.46 5.77 44.23] 41.66*** 7.64** 49.30*** PTSD 38.59 4.14 42.73 | 23.43*** 4.67 28.10*** Impairments Severe Urinary Incontinence 25.59 7.25 32,84 | 46,73*** 4,87*** 51,.60*** Severe Bowel Incontinence 26.99 6.94 33.93 | 47.95*** 4,86*** 52,.81*** E-2 (continued) Table E-2. Hospitalization by Risk Factor (continued) NF Residents Risk Factor, % Without Time Period After Discharge, Days With Hospitalization 0-1 |2-30| 0-30| 0-1 2-30 0-30 Changes in [Impairments Urinary Continence to Severe 38.91 5.72 44.63 46.28*** 5,02*** 51.30*** Bowel Continence to Severe 38.69 5.77 44.46 47.15*** 4,93*** 52.08*** ADL to Total Dependency or Other 38.00 5.88 43.88 55,51*** 2,88*** 58, 39*** CFS 2 Point Change (out of 3 points) 36.22 6.21 42.43 50.74*** 3,71*** 54.45*** */**/*** = Difference is statistically significant based on a p-value cutoff of 0.05/0.01/0.001 E-3 Table E-3. Outpatient Acute Care by Risk Factor NF Residents Risk Factor, % Without With Time Period After Discharge, Days Outpatient O-1 2-30 0-30 O-1 2-30 0-30 Dual Status Full Medicaid eligibility at End of Stay 7.52 6.32 13.84] 9.17*** 6.84** 16.01*** Transition to Medicaid eligibility in 8.84 6.68 15.52] 7.11** 8.06** 15.17 last 3 months Transition to Medicaid eligibility in 8.96 6.60 15.56] 6.71*** 8.01*** 14,72* last 6 months Behavior Severe Physical Behavior 8.44 6.71 15.15] 13.65*** 6.57 20.22*** Severe Verbal Behavior 8.34 6.58 14.92])12.53*** 7.88*** 20.41*** Severe Behavior, Other 8.46 6.67 15.13] 12.37*** 7.08 19,45*** Severe Aggressive Behavior, 8.08 6.61 14.69] 12.49*** 6.90 19.39*** Aggregate Very Severe Aggressive Behavior, 8.08 6.61 14.69] 14.48*** 8.70** 23.18*** Aggregate Behavior Change over the Past Year Physical Behavior to Severe 8.47 6.72 15.19] 13.63*** 6.91* 20.54*** Verbal Behavior to Severe 8.40 6.62 15.02) 12,.42*** 8.06*** 20,48*** Other Behavior to Severe 8.49 6.70 15.19|12.56*** 7.17 19,73*** ABS from Mild or None to Severe 8.58 6.69 15.27|12.47*** 6.84 19,31*** ABS from Mild, None or Severe to 8.65 6.67 15.32] 14.83*** 9,38*** 24,21*** Very Severe New Diagnoses over the Past Year Anxiety 8.70 6.61 15.31] 8.97 7.39* 16.36** Depression 8.73 6.68 15.41] 8.24* 7.25** 15.49 Manic Depression (Bipolar Disorder) 8.70 6.65 15.35] 9.32 8.79*** 18,11** Psychotic Disorder (other than 8.63 6.72 15.35] 9.92*** 6.70 16.62* Schizophrenia) Schizophrenia 8.71 6.67 15.38] 8.27 7.93* 16.20 PTSD 6.21 6.68 12.89] 3.98 9.71** 13.69 Impairments Severe Urinary Incontinence 7.09 9.12 16.21) 9.74*** 5,16*** 14,90*** Severe Bowel Incontinence 7.63 8.80 16.43) 9.60*** 4,98*** 14,58*** Changes in Impairments Urinary Continence to Severe 8.67 6.64 15,.31)10.82*** 5.55*** 16.37** Bowel Continence to Severe 8.59 6.70 15,29/11,13*** 5,23*** 16,36*** ADL to Total Dependency or Other 8.63 6.83 15.46) 11.35*** 2.49*** 13.84** CFS 2 Point Change (out of 3 points) 8.14 7.12 15.26)12.21*** 4,35*** 16,56*** */**/*** = Difference is statistically significant based on a p-value cutoff of 0.05/0.01/0.001. E-4 Table E-4. Discharge Location of NF Residents NF Residents Discharged Live (N=223,522) Another Hospital Nursing Home or Additional Community Swing Bed Acute Psychiatric Settings Demographics Number 76,540 38,142 98,997 4,183 5,660 % of Live Discharges from 34.24% 17.06% 44.29% 1.87% 2.53% Stay Type Male 45.90% 40.90% 42.20% 52.40% 48.18% Female 54.10% 59.10% 57.80% 47.60% 51.92% White only non-Hispanic 68.16% 73.55% 66.59% 73.42% 71.29% Black only non-Hispanic 19.45% 15.44% 20.65% 15.28% 16.25% Hispanic 6.66% 6.42% 7.40% 6.34% 6.80% Other non-Hispanic 5.72% 4.59% 5.36% 4.97% 5.65% Age <65 35.72% 23.20% 19.05% 36.07% 34.49% Age 65-75 23.46% 20.61% 20.81% 28.09% 18.55% Age 75-85 22.01% 26.98% 27.02% 23.02% 20.23% Age 85-95 16.53% 24.83% 27.48% 11.59% 22.40% Age 295 2.27% 4.38% 5.64% 1.22% 4.33% Dual Status Partial Medicaid Eligibility at 1.66% 0.42% 0.87% 0.86% 1.24% End of Stay Full Medicaid Eligibility at End 50.34% 71.61% 70.23% 73.22% 53.60% of Stay Transition to Medicaid 2,66% 2.73% 1.89% 2.77% 1.89% Eligibility in the last 3 months Transition to Medicaid 8.44% 7.29% 5.32% 6.91% 6.10% Eligibility in the last 3 months Behavior Severe Physical Behavior 1.58% 5.45% 5.21% 42.68% 5.87% Severe Verbal Behavior 4.39% 10.27% 9.06% 54.34% 10.17% Severe Behavior, Other 2.83% 7.36% 6.71% 36.87% 7.79% None or Minimal Aggressive 89.86% 79.44% 80.75% 24.34% 79.22% Behavior, Aggregate Moderate Aggressive 7.89% 14.50% 13.40% 34.47% 0.14% Behavior, Aggregate Severe Aggressive Behavior, 1.90% 4.92% 4.69% 29.86% 0.05% Aggregate Very Severe Aggressive 0.35% 1.13% 1.17% 11.33% 1.22% Behavior, Aggregate Behavior Change over the Past Year Physical Behavior to Severe 1.24% 4.18% 4.01% 36.28% 4.84% Verbal Behavior to Severe 3.46% 7.00% 6.35% 40.25% 7.34% Other Behavior to Severe 2.04% 5.15% 4.85% 28.02% 5.74% ABS from Mild or None to 1.45% 3.51% 3.53% 24.58% 4.20% Severe ABS from Mild, None or 0.29% 0.92% 0.95% 10.18% 0.99% Severe to Very Severe (continued) E-5 Table E-4. Discharge Location of NF Residents (continued) NF Residents Discharged Live (N=223,522) Another Hospital Nursing Home or Additional Community Swing Bed Acute Psychiatric Settings Diagnoses Anxiety 27.87% 3.65% 32.00% 47.79% 34.79% Depression 45.27% 54.13% 50.57% 55.50% 47.66% Manic Depression (Bipolar 6.71% 8.63% 6.63% 19.99% 7.93% Disorder) Psychotic Disorder 5.02% 10.96% 9.39% 25.36% 9.05% Schizophrenia 7.64% 11.32% 9.44% 39.95% 13.32% PTSD 1.48% 1.26% 0.80% 2.46% 1.40% Diagnoses change over the past year Anxiety 7.35% 8.35% 8.14% 12.73% 10.21% Depression 9.33% 8.93% 9.06% 9.88% 9.44% Manic Depression (Bipolar 1.64% 1.96% 1.69% 4.21% 1.89% Disorder) Psychotic Disorder (other 1.96% 3.40% 3.08% 9.57% 3.38% than Schizophrenia) Schizophrenia 1.40% 1.80% 1.60% 4.45% 1.64% PTSD 0.39% 0.25% 0.20% 0.53% 0.42% Other Characteristic Severe Urinary Incontinence 34.61% 61.19% 74.43% 42.58% 55.45% Severe Bowel Incontinence 30.53% 53.79% 69.77% 36.74% 51.76% No Cognitive Impairment 40.79% 20.34% 5.61% 9.04% 20.69% Mild Cognitive Impairment 39.32% 30.87% 39.29% 35.74% 37.61% Moderate Cognitive 17.46% 40.45% 43.94% 50.47% 32.05% Impairment Severe Cognitive Impairment 1.60% 5.26% 7.52% 1.96% 6.06% Other Characteristics Change over the Past Year Urinary Continence to Severe 6.20% 11.57% 13.98% 14.45% 11.94% Bowel Continence to Severe 7.03% 13.03% 16.69% 13.76% 13.72% ADL to 28 0.66% 1.56% 4.24% 0.55% 3.78% Cognitive Impairment worsen 11.30% 13.95% 21.98% 19.58% 18.68% by 1 point out of 3 Cognitive Impairment worsen 0.80% 1.56% 2.63% 3.75% 1.80% by 2 points out of 3 1Community includes private home/apartment, board/care, assisted living and group home 2 Additional settings include the following settings: Inpatient rehabilitation facility, ID/DD facility, hospice, long-term care hospital and other. E-6 Figure E-1. Acute Care Use, by Program Termination Status 60.0% 50.0% n & 40.0% A - oS = 30.0% 20.0% a 10.04 -=- idents 0.0% 012345 6 7 8 9 1011121314151617 18192021 222324 25 26 27 282930 Cumulative Perce --Residents discharged from active facilities -- Residents discharge from terminated facility, after termination date -- Residents discharge from terminated facility, on or before termination date Difference between resident groups statistically significant at the p=0.001 level. E-7