Transcript nursing home placement

Can Care Managers Assist Older Adults
to delay Nursing Home Placement?
Development of a Risk Index to Predict Transfers from Home
and Community-Based Waiver Programs to Nursing Homes
Sandra L. Spoelstra, PhD, RN
Midwest Nursing Research Society
April 14th, 2012
Dearborn, Michigan
Acknowledgements
 Dr. Barbara Given, PhD, RN, FAAN University
Distinguished Faculty
 Charles Given, PhD & Raza Haque, MD
 Mei You, MS Statistician
 Mike Daeschlein, MDCH Long-Term Care
Division
 20 Waiver Agencies, 356 Care Managers, &
20,000 Clients
Objectives
 Participants will be presented factors related to
nursing home placement in frail vulnerable
elderly in the community setting.
 Participants will develop an understanding of
how to assist care managers to identify nursing
home placement risk.
Background
 With adults aged 65 and older currently
comprising 15% of the population and growing
exponentially.
 concern is mounting as to how to care for this
growing demographic group.
 It will be important to find ways to deliver highquality care tailored to the needs of clients in
order to allow these individuals to remain living
in the community.
Purpose
 This research examines the risk of nursing home
placement (NHP) an inception cohort of
vulnerable group of low-income elderly persons
in the State of Michigan Home & CommunityBased Waiver program between 2002—2007.
 Focus: developing a risk index to identify waiver
clients who transferred to NH <2 years and
compared them with clients who remained in the
program >2 years.
The Present Study
 From literature review factors were examined in
the Minimum Data Set-Home Care (MDS-HC).
 Examined how change between next to last
assessment and last assessment increases risk
of NHP.
 Different gradations of change in each variable
were examined.
Data Source
 From the State Medicaid data
warehouse:
 MDS-HC assessments
 Medicaid claim files
 Michigan death certificate information
Sample & Setting
 Federal 1915(c) HCBS waiver program in the
State of Michigan
 Clients must meet Medicaid-defined nursing
facility level-of-care criteria (ADLs/IADLs, <300%
of poverty level, & a caregiver).
 Age >65
 Between 2002—2007
14,568
Eligible & had an
MDS-HC
12,839 Had
2+ MDS-HC
Assessments
2,426
NHP in
2 Years
4,099 Stayed
in the Waiver
Program >2
Years
Sample Analyzed
N=6525
Development
Group
Confirmation
Group
N=3263
N=3262
1,567 Stayed
in the
Program <2
Years
1,729 Had
One MDS-HC
Assessment
4747 Other
(3,983 died in the program;
764 lost follow up)
Variables Examined
 Age, sex, race
 Physical function
 dressing, eating, toileting, transferring, and bathing




Cognitive function
Falls
Caregiver informal support hours
Nursing home placement (from claim files)
Two Level Model
 Deterioration was defined as increased number
of ADL dependency, increased the cognitive
scale, and increased falls comparing next to last
assessment and last assessment.
 Limitation: two clients might both be defined as
having no deterioration, if one remained
independent and the other was fully dependent
at both assessments, as deterioration in
condition is the issue in our risk model
Three Level Model
 Deterioration further divided into whether clients
had ADL dependencies or 2+ cognitive
performance deficits or had falls at last MDS.
 Cases that remained were independent at both
MDS assessment; or improved at the final MDS
when compared with their second to last MDS.
 Few cases reported improvement in any
dimension.
Validation of Risk Model
 The sample was split in half using a simple
random sampling.
 N=3263 to develop the risk index (development
sample).
 N=3262 to validate the risk index (confirmation
sample).
Development Sample
 Risk factors were entered in model as predictors
 NHP <2 years was the dependent variable in the
logistic model.
 Risk factors that were not significant were removed 1-
by-1 until all were p <0.05.
 Two risk indexes based on the summed beta weights
multiplied by the risk factors for either deterioration
alone (the 2-level model) or deterioration and
dependency (the 3-level model) were developed.
 5 points were added to each index score so that all
scores were positive.
Confirmation Sample
 Applied same estimated beta weights from the
development sample to the confirmation sample.
 Computed risk indices, and compared the
Association of Predicted Probabilities and
Observed Responses.
 Mann-Whitney non parametric method was used
to compare statistical differences between the 2level model and the 3-level model.
Results
 N=6525
 2426 (37%) transferred to NH in <2 years.
• Clients at high risk of NHP are over 75 years of age, of
Caucasian race, were in a NH before, wished to reside in
another setting, were more likely to have been hospitalized in
the past 90 days, and reported behavioral problems at the
last assessment.
• Informal caregivers & living arrangement did not impact the
model.
 Each factor produced between 10-25% greater
rates of NHP than clients without those factors.
Changes between the 2-MDS Assessments
 In the 3-level model, deterioration in cognitive
status and physical function is a more sensitive
indicator of NHP than the level of dependence
alone at the last observations.
 No deterioration in falls, but reported falls at the
last MDS produced rates of NHP’s (45.7%) to
(46%) that were similar in the group who
remained at home or those with NHP.
Association of Predicted Probabilities &
Observed Responses
 Area under Curve (AUC) of Receiver Operating
Curves (ROCs), % of concordance plus a half
percentage of ties:
 Development sample
• 2-level model was 0.72
• 3-level model was 0.73
 Confirmation sample
• 2-level model was 0.70
• 3-level model was 0.72
 Compared ROCs 3-level & 2-level model to see
which was better at predicting NHP.
Examining Change in Index on NHP
 Categorized risk index into 7 levels, increasing
each level by a magnitude of 0.5.
 Summed each level according to the proportion of
clients who entered a NH.
• Relationship between a 0.5 unit increase in the risk index and
the probability that clients transfer to a NH.
• Beginning with scores of 4.0 to 4.49, each half unit increase
in the risk index produced around a 10% increase in the rate
of NHP.
 For example, as the risk index score increases
(from 2.5 to 6.0), the rate of NHP increased from
21% to 77%.
Correspondence Between NHP & Risk
Index
 Correspondence of probability, sensitivity, and
specificity.
• Assuming a score of 5 on the risk index, then the probability
of transferring to a NH is approximately 50%.
• Using this 50%, we examined sensitivity & specificity
• A sensitivity of approximately 0.4 & specificity close to 0.9.
 This means that for a risk index score >5, we will be
able to correctly identify 40% of those clients who will
actually go to a NH.
 For those clients with a score <5 correctly identifying
90% of those clients who are not going to transfer to a
NH.
Discussion
 The utility of this risk index for waiver program
staff comes from the fact that this model can be
easily produced from information that is already
being collected in the MDS-HC assessments.
 If collected on a laptop computer, the risk index
could be calculated in real time.
 Information from the prior observation, paired with
current risk index scores, could be used to produce a
risk score that would reveal the rates of deterioration
over consecutive assessments.
Discussion
 Waiver staff could target education and/or
services for clients and their caregivers towards
those areas with greater vulnerability.
 Example: if cognitive status was declining, then
caregivers could be informed about how to manage
persons with these declines.
• If hospitalizations existed in the past 90 days, then waiver
agents could examine the reasons for the hospital
admissions and determine what might be done by waiver
staff, and to engage the client, and/or their caregiver, and the
primary care physician in actions to prevent hospitalization.
Discussion
 The index might guide resource allocation needs
to be tested.
 The data does indicate how it might be used to
address this decision.
• Decisions should not rely solely on the overall score, but on
the individual changes in each of the risk index components.
 The 3-level model was superior to the 2-level
model that focused solely on deterioration.
Limitations
 To assure that floor effects were addressed,
examined the prevalence all factors.
• Age was divided as 65-75, and 76+ and then clients were
classified according to no change, or change in one, two, or
all three measures.
• This change score was then compared with the number and
percent of cases with a maximum score on each measure (a
score of 5 on the ADL index, 6 on cognitive performance, and
9+ on falls) at the next to last contact.
 Only 9% of all patients with no change had a
maximum score on ADLs, and 1% had maximum
scores on cognitive performance and falls at the next
to last contact.
Research Implications
 Future research could focus on development of
the laptop application so that this risk index
could be used in the home setting.
 Future research could also focus on how the risk
index functions in a real world setting, and what
actions care managers are able to take to delay
or prevent NHP, and what cost saving are
experienced.
Conclusions
 This index defining risk of transfer to a nursing
home could be a valuable adjuvant to clinical
observations.
 If waiver agents are aware of those clients at
greater risk, they could target services or
intervene to delay or prevent NHP.
Conclusion
 With the increasing pressure to lower costs of
health care, especially for the dually eligible,
efforts such as this capitalize on existing
information, and deliver it to agencies so that
they can make more informed decisions with
respect to how to service clients in wavier
programs.