A Study of Nursing Facility Transitions:
Who Leaves? Who Stays?Presentation to Olmstead Advisory Committee
November 5, 2009
Kathryn E. Thomas, Ph.D.Kathleen H. Wilber, Ph.D.
University of Southern CaliforniaDavis School of Gerontology
Outline
A Fresh Look at Nursing Home Transition Minimum Data Set, Episodes of Care Transition outcomes vs. discharge outcomes
Goal of Study #1: Identify characteristics associated with successful community discharge
Goal of Study #2: Identify barriers to transition among residents with a community living preference
Implications
A Fresh Look:Nursing Facility Transition 40% of people 65+ likely to spend some time
in a nursing home. Targeting candidates is a critical component
of NF transition programs. Transition as “Conversion Diversion” -
Getting individuals out before they convert to long-stay
Focus on “Transition Outcomes” instead of traditional “Discharge Outcomes”
A New Approach: Episodes of Care
MDS: A wealth of data, generally underutilized in NF Transition efforts 10+ million records/year Combination of short & long-stay residents Difficult to extract meaningful data
Transition Outcomes vs. Discharge Outcomes Transition outcomes require broader
perspective. Need to track person across settings.
Episode of Care
Episode Start Date
All variables taken from full Admissions Assessment.
Episode End DateDischarge date and Discharge Status taken from Discharge Tracking Form
Based on the work of Fisher et al., Medical Care, 41(12) 2003.
HOME NH ACUTE ACUTE NHNHNH
0 10 20 30 40 50 60 70 80 90 100
100 Day Episode
Study #1: Successful Transitions
Unit of Analysis: Episode of care Sample: MDS from SCAN/Medicare
(n = 4635) What did we do? We compared…
Community Discharge w/in 90
days
NF placement 90+ days
Community discharge = home w/ home health, home w/o home health or board & care/assisted living
vs.
Who was in the sample?
Who we included: MDS records for SCAN and Medicare individuals who entered a NF in Los Angeles, Orange, Riverside, or San Bernardino between 1/1/01 and 12/31/03.
Who we excluded: Episode length < 14 days, those who died, residents discharged to the hospital w/in 90 days, MR/DD, persistent vegetative state
What did we look at? Predisposing
Age, gender, marital status, race & education Need
Cognitive functioning, depression, comorbidities, social engagement, behavior, ADLs, incontinence, recent fracture, recent fall, admitted-from location
Enabling Generic: Living situation before admission, legal
responsibility, payment source, type of insurance Transition-Specific: Community living preference,
presence of support person positive toward discharge, discharge prediction timeframe, receipt of community living skills training
Study 1: Questions
We were interested in individual characteristics associated with successful transition to the community Which transition-specific variables affect
transition? Does SCAN membership affect
transition?
Results: What Supports Transition?
Preference (Q1a) increases the likelihood of transition by 28%
Presence of support person (Q1b) increases the likelihood of transition by 250%
Discharge prediction (Q1c): Those predicted to stay 30+ days are 43% - 84% less likely to transition than those predicted to stay < thirty days
Community living skills training (P1ar) increases the likelihood of transition by 42%
SCAN membership increases the likelihood of transition by 50%
Study #2: Barriers to Transition
Subsample: Only residents who expressed/indicated preference to return to the community (n = 2935)
Question: Who gets stuck in the NF and
why?
Results
Supports Transition Female Married Recent fracture SCAN (44%) Support person
(269%) Community living
skills training (133%)
Barriers to Transition ADL limitation Bowel incontinence Medicaid (- 43%) Discharge prediction >
30 days (-36% to -76%)
Characteristics & Barriers Key Issues
Support person most important factor More research needed on discharge prediction
and community living skills training variables.
Insurance Medicaid a consistent barrier, SCAN positive SCAN members are less likely to become long-
stay Reconsider S/HMO models?
MDS 3.0
Q1a, Q1b and Q1c removed Replaced by general question about goals and
desire to talk to someone about community transition
Based on this research, it is unfortunate that ‘Presence of a Support Person’ and ‘Predicted Discharge’ were removed
Transition question is still at the end of the assessment and only on full assessments, not quarterly.
MDS 3.0 (Cont)
New Return to Community CAT Triggers could be informed by results of this
study. CAT may shift responsibility for transition
from NF to agencies. Potential to overwhelm. Results could be used to help prioritize list from
CAT CNFTS can also be used by transition
advocates/agencies
Modernizing the MDS Process
NF personnel vs transition advocates/consumers NF Administrators – occupancy Nursing staff – light care need patients are easier
Mandatory referral processes based on CAT will circumvent some of these issues
Consumer Involvement Resident/family should be informed about how
community CAT triggers are filled out Should be able to talk with ombudsman/advocate if
they disagree with assessment. Should be given opportunity to opt into community
living training.
Targeting Strategies Discharge prediction & accrued length of stay
Compare predicted vs actual and assign likelihood level
Protocol in place to check in with resident/family when approaching predicted discharge
Reactive, but easy and could be used with all ages Preference & discharge prediction
Pref & pred < 30 – minimal assistance needed Pref & pred 30-90 or uncertain – midlevel
assistance Pref & pred > 90 – intensive assistance level. Not recommended for 85+
Targeting Strategies (Cont)
Discharge Probability Score Use study results to create discharge
probability score Assign individuals to different transition
assistance level based on probability score. Ultimately CAT could automatically
calculate probability score
Tiered Transition Interventions
Minimal Assistance (residents w/ high likelihood of transition) Review predicted discharge estimate with
resident/family Let them know transition assistance available if
they get off track for discharge. Offered basic information about HCBS and option
to participate in community living skills training for resident and/or family
Tiered Transition (Cont)
Mid-Level Assistance All of the above plus additional community
living preference & feasibility assessment Could use CNHTS around day 30 for
efficiency Intensive Assistance (residents w/ low
likelihood of transition) All of the above plus dedicated transition
counselor