531
NeuroRehabilitation 34 (2014) 531–540 DOI:10.3233/NRE-141056 IOS Press
Transitions between SNF and home-based care in patients with multiple sclerosis Katia Noyesa,b,∗ , Alina Bajorskab , Erin B. Wassermanb , Bianca Weinstock-Guttmanc and Dana Mukameld a Department
of Surgery, University of Rochester School of Medicine, Rochester, NY, USA of Public Health Sciences, University of Rochester School of Medicine, Rochester, NY, USA c Jacobs Neurologic Institute, State University of New York at Buffalo, New York, NY, USA d Department of Medicine and Health Policy Research Institute, University of California, Irvine, CA, USA b Department
Abstract. OBJECTIVES: To assess patient characteristics and risk factors associated with care transitions between skilled nursing facility (SNF) and home care for patients with multiple sclerosis (MS) in an effort to improve outcomes and optimize patient care pathways. BACKGROUND: MS is a chronic neurologic illness of younger adults that is associated with physical disability, cognitive impairment and a high need for supportive services. METHODS: The study was based on the 2005 Nursing Home Minimum Data Set and the Outcome and Assessment Information Set data (n = 10,064). We performed multivariate evaluation of patient risk factors for skilled nursing facility (SNF) admission and disposition while controlling for potential patient self-selection and other characteristics that affect care utilization. RESULTS: MS patients with recent history of home care use were less likely to be admitted to an acute care hospital and had higher physical disability at SNF admission than SNF patients who did not use home care. Insurance type (Medicaid) and availability of informal caregivers were associated with the use of homecare services after a SNF stay, while patient demographic and clinical characteristics did not explain SNF disposition. CONCLUSIONS: Future studies should explore the association between the local availability and affordability of home-based services and physician attitude about community-based care management of disabled patients. Keywords: Multiple Sclerosis (MS), disability, cognitive impairment, home health care (HHC), Skilled nursing facility (SNF)
1. Introduction Multiple sclerosis (MS) is a chronic neurologic illness of younger adults that is associated with progressive physical disability and cognitive problems. As a result, many MS patients require increasing levels of supportive services, including home health care (HHC) and short- and long-term skilled nursing facility (SNF) care, both for rehabilitation and custodial care. ∗ Address
for correspondence: Katia Noyes, PhD, MPH, Departments of Surgery & Public Health Sciences, University of Rochester, Box SURG/SHORE, 265 Crittenden Blvd., Rochester, NY 14642, USA. Tel.: +1 585 275 8467; Fax: +1 585 276 1305; E-mail: katia [email protected].
It is well documented that most community-dwelling older adults want to remain in their own homes (AARP, 2000) and if a skilled nursing facility stay is medically necessary (e.g., as a part of post-acute care or rehabilitation), to return home as soon as possible (Kane, 2008; Nishita et al., 2008). To address this need, a large body of long-term care research and health policy efforts in the last 20 years have been focused on identifying and providing community-based alternatives to institutionalization (Kane, 2008; Arling et al., 2010). In 1999, these efforts became a federal imperative with the Olmstead Decision, in which the Federal Supreme Court determined that unnecessary institutionalization violates the Americans with Disabilities Act of
1053-8135/14/$27.50 © 2014 – IOS Press and the authors. All rights reserved
532
K. Noyes et al. / Transitions between SNF and home-based care in patients with multiple sclerosis
1990 (ADA) (Williams, 2000). The New Rebalancing Initiatives to Transition Persons Out of Nursing Facilities (2003) and the Deficit Reduction Act (2005) further promote home- and community-based services for transitioned individuals (Anderson, 2006). In the absence of reliable evidence, it is difficult to predict what impact these initiatives may have on the care and quality of life of these patients. The purpose of this study is to assess the current patterns of care transitions between home and institutional settings for patients with MS and to identify patient characteristics associated with higher need and potential problems of access to care. We hypothesized that patients with a recent history of home health care enter SNF with a higher disability than patients who did not use home care. We also expect that patients with bladder or bowel incontinence and moderate to severe cognitive problems are more likely to be admitted to a hospital or SNF compared to similar patients without these problems. We believe that this relationship is mediated by patient age and availability of an informal caregiver at home. Finally, we hypothesized that the decision to discharge patients home after a rehabilitative nursing home stay is driven by patient health status or availability of informal care at home.
2. Methods 2.1. Data sources The study was based on the 2005 Nursing Home Minimum Data Set (MDS) and the Outcome and Assessment Information Set (OASIS) data. The MDS is a standardized screening and assessment tool of health status that forms the foundation of the comprehensive assessment for all residents of nursing homes certified by Medicare or Medicaid (Morris et al., 1990). The MDS assessments start at admission (within 14 days) and are performed at specific times afterward, as well as at times of significant change in the residents’ heath statuses (e.g., after acute hospital admission). The MDS contains resident demographic and personal information, diagnoses, infections, health conditions, medications, nursing and rehabilitation services, special treatments and procedures as well as information on resident ADL deficiencies, a range of physical functioning, cognitive patterns, accidents, and other symptoms (e.g., incontinence). To determine prior use of HHC, we linked the MDS individual patient records with the informa-
Fig. 1. Study data flow chart. MDS = Medicare Minimal Data Set; MS = multiple sclerosis.
tion from the Outcome and Assessment Information Set (OASIS). OASIS describes each episode of home care for Medicare-certified home care agencies and includes data elements that are integrated into a comprehensive assessment for all adult home care patients. The information collected by OASIS encompasses socio-demographic, health and functional status, environmental and support system attributes of adult (non-maternity) patients receiving home health services, as well as selected characteristics of health service utilization (e.g., therapy and procedures). 2.2. Analytic data sets MS patients were identified based on MS diagnosis, reported at the baseline (initial) MDS assessment (n = 10,064). Because the OASIS database includes information on only Medicare- or dual-eligible patients, our merged analytic dataset only includes SNF residents with Medicare or Medicaid payment source (78% of all admissions). If an individual had multiple admissions to SNF during 2005 (31%), we included only the first SNF admission. We excluded 265 records with entry date more than 1 month before the assessment reference date. We also excluded 164 SNF stays with incomplete information (admission assessments were not followed by periodical assessments or discharge records). Because we used a two week time window to define “HHC use prior to SNF entry”, we limited the study population to MS patients admitted to
K. Noyes et al. / Transitions between SNF and home-based care in patients with multiple sclerosis
SNF after January 14, 2005. The final sample included 6,890 patients with MS who had at least one nursing home admission during 2005 (see Fig. 1 for additional information). By linking the MDS records to the OASIS database, we found discharges or transfers from HHC agencies preceding the SNF admission. If multiple HHC episodes were reported within the year, we selected the last HHC episode of care before SNF admission. To determine Medicaid eligibility, we used payment source information at admission as reported by the MDS. Because MDS payment source field has been shown to be unreliable in reporting Medicare as a payment source for SNF admissions, we used the Medicare Prospective Payment System (PPS) assessment records from the MDS for indication of Medicare as a payment source (Buchanan et al., 2006; Cai et al., 2011). For each admission, we searched for the mandatory Medicare PPS assessments within two weeks after the SNF entry date. The presence of PPS assessment indicated Medicare as the payment source for the admission. To examine factors associated with specific care transitions between SNF and home, with and without home services, we defined three mutually exclusive groups of MS patients. The first group included MS patients who did not use HHC services within the last two weeks before SNF admission. The remaining MS patients with a recent history of HHC use (within the two-week time window prior to index SNF admission) were further divided into two groups — those who were temporarily transferred from a HHC agency to an inpatient facility (acute care hospital or a skilled nursing facility, HHCTR) and those who were permanently discharged from HHC agency (HHC-DIS). We chose to analyze the last two groups separately since “discharged” status may be a proxy for otherwise greater unobserved severity of disability compared to “transferred”. 2.3. Analysis 2.3.1. Assessing differences in MS patient groups at the time of skilled nursing facility admission To explore the differences in personal characteristics and health status of MS patients comprising the three care pathway groups described above, we first performed bivariate analyses using comparison of the means and Chi-square tests. Using the information obtained from the bivariate analysis to guide the model development, we then developed a multinomial logistic regression model with a three-level dependent vari-
533
able (one level for each group/care trajectory). Using the model, we conducted a multivariate evaluation of the odds of being in each of the care pathway groups after adjusting for patient risk factors (i.e., patient demographics, insurance status, admission source, cognitive and functional limitations, living situation before admission, and pain). 2.3.2. Evaluating SNF discharge destination for residents with MS We used the Cox proportional hazard model with the event (outcome) defined as a discharge/transfer from the SNF to any of the available destinations (e.g., discharge home with HHC services or assisted living facility, discharge home without HHC services, discharge or transfer to an acute care hospital, or continue SNF stay). To estimate the proportion of MS patients discharged or transferred to a specific destination, we employed survival-type sub-distribution hazard models. The time origin was admission to SNF and the event of interest was the specific type of discharge as described above (e.g., discharge home with HHC services or assisted living facility, discharge home without HHC services, discharge or transfer to an acute care hospital, or continue SNF stay), while all other discharges were considered competing risks. Based on the sub-distribution models, we estimated Cumulative Incidence Function (CIF) for each care pathway group. The value of CIF at a given point of time was the estimated proportion of people with a specific outcome (e.g., discharge home with HHC services or assisted living facility, discharge home without HHC services, discharge or transfer to an acute care hospital, or continue SNF stay) by that time. The proportion of people who remained in SNF beyond four months was estimated using a survival function. To explore the impact of patient characteristics on the choice of SNF discharge destination, we first developed four sub-distribution hazard models (one for each discharge destination, same as described above) with an indicator for care pathway. This way, we obtained group-specific proportions, not adjusted for any differences between the groups. Next, we incorporated additional patient characteristics in the hazard model including payment source, an indicator whether the patient was admitted to the SNF directly from an acute care hospital, as well as socio-demographics, disability status and other health status indicators. Using these models, we estimated unadjusted and adjusted proportions of patients discharged to each destination as a function of time since admission.
534
K. Noyes et al. / Transitions between SNF and home-based care in patients with multiple sclerosis
Fig. 2. Multiple sclerosis patient journey: care transitions between community and institutional care. Sixteen percent of all multiple sclerosis patients in skilled nursing facilities had a home care episode within 2 weeks of being admitted to an institution. Out of those, a quarter were discharged from (completed) the home care episode before being transferred to an institution. Among patients who did not complete the episode of care, a much greater proportion were transferred to an acute care hospital before going to SNF (87% compared to 27% in the discharged group). HHC = Home Health Care; MS = Multiple Sclerosis, NH = Nursing Home, Synonymous with SNF = Skilled Nursing Facility.
3. Results 3.1. Prior homecare use and patient functional status at the time of SNF admission Type of insurance, functional status, cognitive problems and incontinence determined which patients would continue with HHC while having an acute health care episode, and which patients would be permanently discharged from HHC. Based on unadjusted bivariate analysis, the vast majority of patients (87%) who were transferred rather than discharged from a HHC agency were Medicare beneficiaries admitted to SNF right after an inpatient stay (Fig. 2). Patients in this group were more physically disabled (as measured by “bed mobility” and “locomotion off unit”) but had fewer cases of bladder incontinence and cognitive impairment compared to MS patients who did not use HHC prior to SNF entry (Table 1). These patients were more likely to be women, over age 65, and more likely to be living with a spouse compared to patients who did not use HHC prior to SNF admission. Among patients who were discharged from their last episode of HHC, a lower proportion (about
60%), were Medicare beneficiaries admitted directly from a hospital. MS patients who were permanently discharged from HHC agency before SNF admission were younger and less cognitively impaired, but needed more assistance in bed mobility and off unit locomotion than MS patients in other groups (Table 1). Based on the results of the multivariate analysis (Table 2), being female, on Medicare, having no bladder or cognitive problems, being admitted directly from an acute care hospital and needing assistance in bed mobility were significantly associated with continuing HHC episode while staying in SNF for rehabilitation. Living with a spouse and advanced age were not significantly associated with prior HHC use. On the contrary (Table 3), we found that MS patients who were permanently discharged from HHC prior to SNF admission were less likely to have a prior acute care hospital admission and had higher physical disability than SNF patients who did not have prior HHC use. 3.2. Association between being discharged home with HHC services and custodial SNF stay Almost half of MS patients who used HHC before SNF admission returned home with HHC after a short
K. Noyes et al. / Transitions between SNF and home-based care in patients with multiple sclerosis
535
Table 1 Characteristics of nursing homes MS patients at admission, based on prior home care use Characteristic
Prior use of home care No prior HHC use
Medicaid payment source Medicare payment source admitted from hospital age ≥ 65 group, percent male Living situation living alone married, living with spouse not married, living with sibling in other facility Race white black Hispanics/Others Cognitive impairment- (CPS) none to mild moderate moderate/severe to very severe Bed mobility independent or supervision limited assistance extensive assistance total dependence Locomotion off unit independent, supervision or limited assistance extensive assistance total dependence activity did not occur Incontinence (frequent or all the time) Bladder Bowel Pain intensity no pain – reference mild moderate excruciating
Prior HHC use
5,758
Transferred (HHC episode open) 827
Discharged (HHC episode closed) 305
28.4 71.6 75.1 39.0 29.1
% 14.8 85.3 86.1 43.9 24.7
40.3 59.7 55.7 32.5 29.2
23.0 31.1 33.8 12.1
24.0 38.9 32.6 4.6
25.9 30.2 36.7 7.2
83.4 13.1 3.5
84.0 13.4 2.5
83.6 13.8 2.6
78.4 14.2 7.5
82.1 11.1 6.8
86.6 8.9 4.6
16.6 16.8 42.7 23.9
9.0 15.5 47.6 27.9
11.8 16.7 41.6 29.8
34.6
29.4
36.1
17.1 37.3 11.1
15.6 40.2 14.9
10.8 40.0 13.1
28.2 41.2
21.9 44.1
26.6 45.3
38.7 14.9 38.0 8.4
34.5 15.5 41.4 8.7
36.1 13.4 42.0 8.5
HHC = home health care; MS = multiple sclerosis; CPS = Minimum Data Set Cognitive Performance Scale.
SNF stay (47% of patients return home with HHC services within 120 days of SNF admission, compared to 30% in other groups (p < 0.001)) (see Appendix Tables A1 and A2 for additional information). Factors positively associated with HHC discharge included Medicare payment source (HR = 2.169, p < 0.001), admission to SNF directly from an acute care hospital, and living with a spouse. Factors negatively associated with being discharged home with HHC services
included being male, living with somebody but not spouse or living in another facility at the time of SNF admission. Additionally, high stages of cognitive impairment, need for extensive assistance, total dependency in bed mobility and bowel incontinence were associated with lower probability of being discharged home with HHC. Results for other SNF disposition locations are presented (see Tables 4 and 5 for additional information).
536
K. Noyes et al. / Transitions between SNF and home-based care in patients with multiple sclerosis Table 2 Patient characteristics associated with high rate of transfer/discharge from home care setting
Patient characteristics Transfer (temporary leave) from HHC Medicare payment source Admitted from hospital Age ≥ 65 below 65 -continuous age above 65-continuous Male
OR
SE
P-value
1.702
0.202
0.000
1.318 1.038 0.999 1.005 0.782
0.158 0.141 0.006 0.009 0.070
0.021 0.783 0.852 0.600 0.006
1.137
0.115
0.205
Table 2 (Continued) Patient characteristics
OR
SE
P-value
Joint test
Transfer (temporary leave) from HHC Bed Mobility Limited assistance extensive assistance total dependence
1.703 1.738 1.995
0.390 0.367 0.479
0.020 0.009 0.004
0.030
Locomotion extensive assistance total dependence activity did not occur
0.645 1.089 1.214
0.137 0.171 0.257
0.039 0.588 0.359
0.056
Incontinence (frequent or all the time) Bladder Bowel
0.845 1.129
0.123 0.164
0.245 0.403
0.202
Pain intensity Mild Moderate Excruciating
0.956 1.222 1.122
0.182 0.170 0.258
0.812 0.148 0.617
OR = odds ratio; SE = standard errors; CI = confidence interval; HHC = home health care.
Joint test
Living situation Married, living with spouse Not married, living with sb. Living in other facility Black Hispanic/Others
0.991
0.104
0.933
0.449 1.115 0.700
0.085 0.129 0.174
0.000 0.349 0.151
Cognitively impaired moderate severe
0.776 0.801
0.095 0.128
0.038 0.165
0.064
Bed mobility limited assistance extensive assistance total dependence
1.583 1.969 2.117
0.249 0.284 0.347
0.003 0.000 0.000
NeuroRehabilitation 34 (2014) 531–540 DOI:10.3233/NRE-141056 IOS Press
Transitions between SNF and home-based care in patients with multiple sclerosis Katia Noyesa,b,∗ , Alina Bajorskab , Erin B. Wassermanb , Bianca Weinstock-Guttmanc and Dana Mukameld a Department
of Surgery, University of Rochester School of Medicine, Rochester, NY, USA of Public Health Sciences, University of Rochester School of Medicine, Rochester, NY, USA c Jacobs Neurologic Institute, State University of New York at Buffalo, New York, NY, USA d Department of Medicine and Health Policy Research Institute, University of California, Irvine, CA, USA b Department
Abstract. OBJECTIVES: To assess patient characteristics and risk factors associated with care transitions between skilled nursing facility (SNF) and home care for patients with multiple sclerosis (MS) in an effort to improve outcomes and optimize patient care pathways. BACKGROUND: MS is a chronic neurologic illness of younger adults that is associated with physical disability, cognitive impairment and a high need for supportive services. METHODS: The study was based on the 2005 Nursing Home Minimum Data Set and the Outcome and Assessment Information Set data (n = 10,064). We performed multivariate evaluation of patient risk factors for skilled nursing facility (SNF) admission and disposition while controlling for potential patient self-selection and other characteristics that affect care utilization. RESULTS: MS patients with recent history of home care use were less likely to be admitted to an acute care hospital and had higher physical disability at SNF admission than SNF patients who did not use home care. Insurance type (Medicaid) and availability of informal caregivers were associated with the use of homecare services after a SNF stay, while patient demographic and clinical characteristics did not explain SNF disposition. CONCLUSIONS: Future studies should explore the association between the local availability and affordability of home-based services and physician attitude about community-based care management of disabled patients. Keywords: Multiple Sclerosis (MS), disability, cognitive impairment, home health care (HHC), Skilled nursing facility (SNF)
1. Introduction Multiple sclerosis (MS) is a chronic neurologic illness of younger adults that is associated with progressive physical disability and cognitive problems. As a result, many MS patients require increasing levels of supportive services, including home health care (HHC) and short- and long-term skilled nursing facility (SNF) care, both for rehabilitation and custodial care. ∗ Address
for correspondence: Katia Noyes, PhD, MPH, Departments of Surgery & Public Health Sciences, University of Rochester, Box SURG/SHORE, 265 Crittenden Blvd., Rochester, NY 14642, USA. Tel.: +1 585 275 8467; Fax: +1 585 276 1305; E-mail: katia [email protected].
It is well documented that most community-dwelling older adults want to remain in their own homes (AARP, 2000) and if a skilled nursing facility stay is medically necessary (e.g., as a part of post-acute care or rehabilitation), to return home as soon as possible (Kane, 2008; Nishita et al., 2008). To address this need, a large body of long-term care research and health policy efforts in the last 20 years have been focused on identifying and providing community-based alternatives to institutionalization (Kane, 2008; Arling et al., 2010). In 1999, these efforts became a federal imperative with the Olmstead Decision, in which the Federal Supreme Court determined that unnecessary institutionalization violates the Americans with Disabilities Act of
1053-8135/14/$27.50 © 2014 – IOS Press and the authors. All rights reserved
532
K. Noyes et al. / Transitions between SNF and home-based care in patients with multiple sclerosis
1990 (ADA) (Williams, 2000). The New Rebalancing Initiatives to Transition Persons Out of Nursing Facilities (2003) and the Deficit Reduction Act (2005) further promote home- and community-based services for transitioned individuals (Anderson, 2006). In the absence of reliable evidence, it is difficult to predict what impact these initiatives may have on the care and quality of life of these patients. The purpose of this study is to assess the current patterns of care transitions between home and institutional settings for patients with MS and to identify patient characteristics associated with higher need and potential problems of access to care. We hypothesized that patients with a recent history of home health care enter SNF with a higher disability than patients who did not use home care. We also expect that patients with bladder or bowel incontinence and moderate to severe cognitive problems are more likely to be admitted to a hospital or SNF compared to similar patients without these problems. We believe that this relationship is mediated by patient age and availability of an informal caregiver at home. Finally, we hypothesized that the decision to discharge patients home after a rehabilitative nursing home stay is driven by patient health status or availability of informal care at home.
2. Methods 2.1. Data sources The study was based on the 2005 Nursing Home Minimum Data Set (MDS) and the Outcome and Assessment Information Set (OASIS) data. The MDS is a standardized screening and assessment tool of health status that forms the foundation of the comprehensive assessment for all residents of nursing homes certified by Medicare or Medicaid (Morris et al., 1990). The MDS assessments start at admission (within 14 days) and are performed at specific times afterward, as well as at times of significant change in the residents’ heath statuses (e.g., after acute hospital admission). The MDS contains resident demographic and personal information, diagnoses, infections, health conditions, medications, nursing and rehabilitation services, special treatments and procedures as well as information on resident ADL deficiencies, a range of physical functioning, cognitive patterns, accidents, and other symptoms (e.g., incontinence). To determine prior use of HHC, we linked the MDS individual patient records with the informa-
Fig. 1. Study data flow chart. MDS = Medicare Minimal Data Set; MS = multiple sclerosis.
tion from the Outcome and Assessment Information Set (OASIS). OASIS describes each episode of home care for Medicare-certified home care agencies and includes data elements that are integrated into a comprehensive assessment for all adult home care patients. The information collected by OASIS encompasses socio-demographic, health and functional status, environmental and support system attributes of adult (non-maternity) patients receiving home health services, as well as selected characteristics of health service utilization (e.g., therapy and procedures). 2.2. Analytic data sets MS patients were identified based on MS diagnosis, reported at the baseline (initial) MDS assessment (n = 10,064). Because the OASIS database includes information on only Medicare- or dual-eligible patients, our merged analytic dataset only includes SNF residents with Medicare or Medicaid payment source (78% of all admissions). If an individual had multiple admissions to SNF during 2005 (31%), we included only the first SNF admission. We excluded 265 records with entry date more than 1 month before the assessment reference date. We also excluded 164 SNF stays with incomplete information (admission assessments were not followed by periodical assessments or discharge records). Because we used a two week time window to define “HHC use prior to SNF entry”, we limited the study population to MS patients admitted to
K. Noyes et al. / Transitions between SNF and home-based care in patients with multiple sclerosis
SNF after January 14, 2005. The final sample included 6,890 patients with MS who had at least one nursing home admission during 2005 (see Fig. 1 for additional information). By linking the MDS records to the OASIS database, we found discharges or transfers from HHC agencies preceding the SNF admission. If multiple HHC episodes were reported within the year, we selected the last HHC episode of care before SNF admission. To determine Medicaid eligibility, we used payment source information at admission as reported by the MDS. Because MDS payment source field has been shown to be unreliable in reporting Medicare as a payment source for SNF admissions, we used the Medicare Prospective Payment System (PPS) assessment records from the MDS for indication of Medicare as a payment source (Buchanan et al., 2006; Cai et al., 2011). For each admission, we searched for the mandatory Medicare PPS assessments within two weeks after the SNF entry date. The presence of PPS assessment indicated Medicare as the payment source for the admission. To examine factors associated with specific care transitions between SNF and home, with and without home services, we defined three mutually exclusive groups of MS patients. The first group included MS patients who did not use HHC services within the last two weeks before SNF admission. The remaining MS patients with a recent history of HHC use (within the two-week time window prior to index SNF admission) were further divided into two groups — those who were temporarily transferred from a HHC agency to an inpatient facility (acute care hospital or a skilled nursing facility, HHCTR) and those who were permanently discharged from HHC agency (HHC-DIS). We chose to analyze the last two groups separately since “discharged” status may be a proxy for otherwise greater unobserved severity of disability compared to “transferred”. 2.3. Analysis 2.3.1. Assessing differences in MS patient groups at the time of skilled nursing facility admission To explore the differences in personal characteristics and health status of MS patients comprising the three care pathway groups described above, we first performed bivariate analyses using comparison of the means and Chi-square tests. Using the information obtained from the bivariate analysis to guide the model development, we then developed a multinomial logistic regression model with a three-level dependent vari-
533
able (one level for each group/care trajectory). Using the model, we conducted a multivariate evaluation of the odds of being in each of the care pathway groups after adjusting for patient risk factors (i.e., patient demographics, insurance status, admission source, cognitive and functional limitations, living situation before admission, and pain). 2.3.2. Evaluating SNF discharge destination for residents with MS We used the Cox proportional hazard model with the event (outcome) defined as a discharge/transfer from the SNF to any of the available destinations (e.g., discharge home with HHC services or assisted living facility, discharge home without HHC services, discharge or transfer to an acute care hospital, or continue SNF stay). To estimate the proportion of MS patients discharged or transferred to a specific destination, we employed survival-type sub-distribution hazard models. The time origin was admission to SNF and the event of interest was the specific type of discharge as described above (e.g., discharge home with HHC services or assisted living facility, discharge home without HHC services, discharge or transfer to an acute care hospital, or continue SNF stay), while all other discharges were considered competing risks. Based on the sub-distribution models, we estimated Cumulative Incidence Function (CIF) for each care pathway group. The value of CIF at a given point of time was the estimated proportion of people with a specific outcome (e.g., discharge home with HHC services or assisted living facility, discharge home without HHC services, discharge or transfer to an acute care hospital, or continue SNF stay) by that time. The proportion of people who remained in SNF beyond four months was estimated using a survival function. To explore the impact of patient characteristics on the choice of SNF discharge destination, we first developed four sub-distribution hazard models (one for each discharge destination, same as described above) with an indicator for care pathway. This way, we obtained group-specific proportions, not adjusted for any differences between the groups. Next, we incorporated additional patient characteristics in the hazard model including payment source, an indicator whether the patient was admitted to the SNF directly from an acute care hospital, as well as socio-demographics, disability status and other health status indicators. Using these models, we estimated unadjusted and adjusted proportions of patients discharged to each destination as a function of time since admission.
534
K. Noyes et al. / Transitions between SNF and home-based care in patients with multiple sclerosis
Fig. 2. Multiple sclerosis patient journey: care transitions between community and institutional care. Sixteen percent of all multiple sclerosis patients in skilled nursing facilities had a home care episode within 2 weeks of being admitted to an institution. Out of those, a quarter were discharged from (completed) the home care episode before being transferred to an institution. Among patients who did not complete the episode of care, a much greater proportion were transferred to an acute care hospital before going to SNF (87% compared to 27% in the discharged group). HHC = Home Health Care; MS = Multiple Sclerosis, NH = Nursing Home, Synonymous with SNF = Skilled Nursing Facility.
3. Results 3.1. Prior homecare use and patient functional status at the time of SNF admission Type of insurance, functional status, cognitive problems and incontinence determined which patients would continue with HHC while having an acute health care episode, and which patients would be permanently discharged from HHC. Based on unadjusted bivariate analysis, the vast majority of patients (87%) who were transferred rather than discharged from a HHC agency were Medicare beneficiaries admitted to SNF right after an inpatient stay (Fig. 2). Patients in this group were more physically disabled (as measured by “bed mobility” and “locomotion off unit”) but had fewer cases of bladder incontinence and cognitive impairment compared to MS patients who did not use HHC prior to SNF entry (Table 1). These patients were more likely to be women, over age 65, and more likely to be living with a spouse compared to patients who did not use HHC prior to SNF admission. Among patients who were discharged from their last episode of HHC, a lower proportion (about
60%), were Medicare beneficiaries admitted directly from a hospital. MS patients who were permanently discharged from HHC agency before SNF admission were younger and less cognitively impaired, but needed more assistance in bed mobility and off unit locomotion than MS patients in other groups (Table 1). Based on the results of the multivariate analysis (Table 2), being female, on Medicare, having no bladder or cognitive problems, being admitted directly from an acute care hospital and needing assistance in bed mobility were significantly associated with continuing HHC episode while staying in SNF for rehabilitation. Living with a spouse and advanced age were not significantly associated with prior HHC use. On the contrary (Table 3), we found that MS patients who were permanently discharged from HHC prior to SNF admission were less likely to have a prior acute care hospital admission and had higher physical disability than SNF patients who did not have prior HHC use. 3.2. Association between being discharged home with HHC services and custodial SNF stay Almost half of MS patients who used HHC before SNF admission returned home with HHC after a short
K. Noyes et al. / Transitions between SNF and home-based care in patients with multiple sclerosis
535
Table 1 Characteristics of nursing homes MS patients at admission, based on prior home care use Characteristic
Prior use of home care No prior HHC use
Medicaid payment source Medicare payment source admitted from hospital age ≥ 65 group, percent male Living situation living alone married, living with spouse not married, living with sibling in other facility Race white black Hispanics/Others Cognitive impairment- (CPS) none to mild moderate moderate/severe to very severe Bed mobility independent or supervision limited assistance extensive assistance total dependence Locomotion off unit independent, supervision or limited assistance extensive assistance total dependence activity did not occur Incontinence (frequent or all the time) Bladder Bowel Pain intensity no pain – reference mild moderate excruciating
Prior HHC use
5,758
Transferred (HHC episode open) 827
Discharged (HHC episode closed) 305
28.4 71.6 75.1 39.0 29.1
% 14.8 85.3 86.1 43.9 24.7
40.3 59.7 55.7 32.5 29.2
23.0 31.1 33.8 12.1
24.0 38.9 32.6 4.6
25.9 30.2 36.7 7.2
83.4 13.1 3.5
84.0 13.4 2.5
83.6 13.8 2.6
78.4 14.2 7.5
82.1 11.1 6.8
86.6 8.9 4.6
16.6 16.8 42.7 23.9
9.0 15.5 47.6 27.9
11.8 16.7 41.6 29.8
34.6
29.4
36.1
17.1 37.3 11.1
15.6 40.2 14.9
10.8 40.0 13.1
28.2 41.2
21.9 44.1
26.6 45.3
38.7 14.9 38.0 8.4
34.5 15.5 41.4 8.7
36.1 13.4 42.0 8.5
HHC = home health care; MS = multiple sclerosis; CPS = Minimum Data Set Cognitive Performance Scale.
SNF stay (47% of patients return home with HHC services within 120 days of SNF admission, compared to 30% in other groups (p < 0.001)) (see Appendix Tables A1 and A2 for additional information). Factors positively associated with HHC discharge included Medicare payment source (HR = 2.169, p < 0.001), admission to SNF directly from an acute care hospital, and living with a spouse. Factors negatively associated with being discharged home with HHC services
included being male, living with somebody but not spouse or living in another facility at the time of SNF admission. Additionally, high stages of cognitive impairment, need for extensive assistance, total dependency in bed mobility and bowel incontinence were associated with lower probability of being discharged home with HHC. Results for other SNF disposition locations are presented (see Tables 4 and 5 for additional information).
536
K. Noyes et al. / Transitions between SNF and home-based care in patients with multiple sclerosis Table 2 Patient characteristics associated with high rate of transfer/discharge from home care setting
Patient characteristics Transfer (temporary leave) from HHC Medicare payment source Admitted from hospital Age ≥ 65 below 65 -continuous age above 65-continuous Male
OR
SE
P-value
1.702
0.202
0.000
1.318 1.038 0.999 1.005 0.782
0.158 0.141 0.006 0.009 0.070
0.021 0.783 0.852 0.600 0.006
1.137
0.115
0.205
Table 2 (Continued) Patient characteristics
OR
SE
P-value
Joint test
Transfer (temporary leave) from HHC Bed Mobility Limited assistance extensive assistance total dependence
1.703 1.738 1.995
0.390 0.367 0.479
0.020 0.009 0.004
0.030
Locomotion extensive assistance total dependence activity did not occur
0.645 1.089 1.214
0.137 0.171 0.257
0.039 0.588 0.359
0.056
Incontinence (frequent or all the time) Bladder Bowel
0.845 1.129
0.123 0.164
0.245 0.403
0.202
Pain intensity Mild Moderate Excruciating
0.956 1.222 1.122
0.182 0.170 0.258
0.812 0.148 0.617
OR = odds ratio; SE = standard errors; CI = confidence interval; HHC = home health care.
Joint test
Living situation Married, living with spouse Not married, living with sb. Living in other facility Black Hispanic/Others
0.991
0.104
0.933
0.449 1.115 0.700
0.085 0.129 0.174
0.000 0.349 0.151
Cognitively impaired moderate severe
0.776 0.801
0.095 0.128
0.038 0.165
0.064
Bed mobility limited assistance extensive assistance total dependence
1.583 1.969 2.117
0.249 0.284 0.347
0.003 0.000 0.000
