One Way Out? A Multistate Transition Model of Outcomes After Nursing Home Admission

Patterns of Transient and Terminal Transitions in Activities of Daily Living Performance Levels among Long-Term Care Residents: A Multistate Markov's Model Analysis of Population-Based Longitudinal Data in Canada

OBJECTIVE

We examined how long-term care (LTC) home residents transition between different activities of daily living (ADL) performance levels, and to eventual terminal clinical outcomes.

DESIGN

We conducted a longitudinal retrospective analysis of population-based data among institutionalized older adults within 3 Canadian provinces.

SETTING AND PARTICIPANTS

LTC home residents within 3 Canadian provinces of Alberta, British Columbia, and Ontario placed between January 2010 and December 2020.

METHODS

We fit a Markov-chain multistate transition model to the data to obtain transition probabilities, sojourn times, as well as the adjusted odds of each transition.

RESULTS

Three distinct transitions were commonly experienced by residents from this analysis. Most LTC residents stayed unchanged in their ADL performance level between 90-day assessments, a substantial proportion transitioned to worse performance level, and only a small proportion improved to a better performance level. Residents spent on average between 21 and 29 months on admission before finally transitioning out of the setting to 1 of 4 terminal states that include mortality, hospitalization, home, or other setting discharges. Within 5 years of admission, between 63% and 72% died, 18% to 19% were hospitalized, and 2% to 4% were discharged back home. The odds of transitioning to different states were strongly affected by factors such as Index of Social Engagement, Cognitive Performance Scale, Changes in Health, End Stage Disease, and Signs and Symptoms score, age, as well as province where the LTC home is located, but varied depending on the admission ADL status.

CONCLUSIONS AND IMPLICATIONS

Evidence from this study shows that it does not always have to be one way out for LTC residents. LTC home administrators could use the findings to identify residents who could be provided the right intervention to facilitate ADL performance improvement and prevent further decline.

A systematic review of reasons and risks for acute service use by older adult residents of long-term care

AIMS AND OBJECTIVES

To identify the reasons and/or risk factors for hospital admission and/or emergency department attendance for older (≥60 years) residents of long-term care facilities.

BACKGROUND

Older adults' use of acute services is associated with significant financial and social costs. A global understanding of the reasons for the use of acute services may allow for early identification and intervention, avoid clinical deterioration, reduce the demand for health services and improve quality of life.

DESIGN

Systematic review registered in PROSPERO (CRD42022326964) and reported following PRISMA guidelines.

METHODS

The search strategy was developed in consultation with an academic librarian. The strategy used MeSH terms and relevant keywords. Articles published since 2017 in English were eligible for inclusion. CINAHL, MEDLINE, Scopus and Web of Science Core Collection were searched (11/08/22). Title, abstract, and full texts were screened against the inclusion/exclusion criteria; data extraction was performed two blinded reviewers. Quality of evidence was assessed using the NewCastle Ottawa Scale (NOS).

RESULTS

Thirty-nine articles were eligible and included in this review; included research was assessed as high-quality with a low risk of bias. Hospital admission was reported as most likely to occur during the first year of residence in long-term care. Respiratory and cardiovascular diagnoses were frequently associated with acute services use. Frailty, hypotensive medications, falls and inadequate nutrition were associated with unplanned service use.

CONCLUSIONS

Modifiable risks have been identified that may act as a trigger for assessment and be amenable to early intervention. Coordinated intervention may have significant individual, social and economic benefits.

RELEVANCE TO CLINICAL PRACTICE

This review has identified several modifiable reasons for acute service use by older adults. Early and coordinated intervention may reduce the risk of hospital admission and/or emergency department.

REPORTING METHOD

This systematic review was conducted and reported following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) methodology.

PATIENT OR PUBLIC CONTRIBUTION

No patient or public contribution.

A multistate transition model of changes in loneliness and carer stress among community-dwelling older adults in Aotearoa New Zealand

OBJECTIVES

To identify changes in loneliness and carer stress between two time points for older people of different ethnicities who had repeated interRAI home care assessments.

METHODS

Participants consisted of community-dwelling older adults across New Zealand who received two interRAI-HC assessments between 5 July 2012 and 31 December 2019. Two multistate models were developed: the first model was not lonely versus lonely, and the second model was no carer stress versus carer stress. The one-year transition probabilities were calculated. Mean sojourn times were calculated for each state except death. Paired t-tests assessed the differences in transition probabilities between the different ethnic groups.

RESULTS

The mean age of the cohort was 82.5 years (SD 7.7 years). At first assessment, 14,646 (21%) older people stated they were lonely and 26,789 carers (38%) experienced stress. The most common first transition type was not lonely to not lonely: Māori 42%, Pacific 54%, Asian, 48% and Other 40%. The highest one-year transition probability in the loneliness model was living in aged residential care to death (0.79). The most common first transition type for the carer stress was no carer stress to no carer stress: Māori 35%, Pacific, 46%, Asian, 43% and Other 33%. The highest one-year transition probability in the carer stress model was living in aged residential care to death (0.80). The statuses not lonely and no carer stress had a mean sojourn time of approximately one year, and eight months to one year, respectively.

CONCLUSIONS

Loneliness can change over time due to circumstances and an individual's perception of loneliness at the time of assessment. Carer stress is enduring and has a low probability of improvement.

Factors Associated With the Initiation of Comfort-Focused Nutrition Care Orders for Long-Term Care Residents at End of Life

BACKGROUND

Comfort-focused nutrition orders are recommended to manage eating changes among long-term care (LTC) residents nearing the end of life, though little is known about their current use. This investigation aims to describe current practices and identify resident-level and time-dependent factors associated with comfort-focused nutrition orders in this context.

METHODS

Data were retrospectively extracted from resident charts of decedents (≥65 years at death, admitted ≥6 months) in 18 LTC homes from two sampling frames across southern Ontario, Canada. Observations occurred at 6 months (baseline), 3 months, 1 month and 2 weeks prior to death. Extracted data included functional measures (e.g. cognitive performance, health instability) at baseline, formalised restorative and comfort-focused nutrition care interventions at each timepoint and eating changes reported in the progress notes in 2 weeks following each timepoint. Logistic regression and time-varying logistic regression models determined resident-level (e.g. functional characteristics) and time-dependent factors (e.g. eating changes) associated with receiving a comfort-focused nutrition order.

RESULTS

Less than one-third (30.5%; n = 50) of 164 participants (61.0% female; mean age = 88.3 ± 7.5 years) received a comfort-focused nutrition order, whereas most (99%) received at least one restorative nutrition intervention to support oral food intake. Discontinuation of nutrition interventions was rare (8.5%). Comfort orders were more likely with health instability (OR [95% CI] = 4.35 [1.49, 13.76]), within 2 weeks of death (OR = 5.50 [1.70, 17.11]), when an end-of-life conversation had occurred since the previous timepoint (OR = 5.66 [2.83, 11.33]), with discontinued nutrition interventions (OR = 6.31 [1.75, 22.72]), with co-occurrence of other care plan modifications (OR = 1.48 [1.10, 1.98]) and with a greater number of eating changes (OR = 1.19 [1.02, 1.38]), especially dysphagia (OR = 2.59 [1.09, 6.17]), at the preceding timepoint.

CONCLUSIONS

Comfort-focused nutrition orders were initiated for less than one-third of decedents and most often in the end stages of life, possibly representing missed opportunities to support the quality of life for this vulnerable population. An increase in eating changes, including new dysphagia, may signal a need for proactive end-of-life conversations involving comfort nutrition care options.

IMPLICATIONS FOR PRACTICE

Early and open conversations with residents and family about potential eating changes and comfort-focused nutrition care options should be encouraged and planned for among geriatric nursing teams working in LTC. These conversations may be beneficial even as early as resident admission to the home.

Primary health care service utilisation before and after entry into long-term care in Australia

OBJECTIVES

To examine utilisation of primary health care services (subsidised by the Australian Government, Medicare Benefits Schedule, MBS) before and after entry into long-term care (LTC) in Australia.

METHODS

A retrospective cohort study of older people (aged ≥65 years) who entered LTC in Australia between 2012 and 2016 using the Historical Cohort of the Registry of Senior Australians. MBS-subsidised general attendances (general practitioner (GP), medical and nurse practitioners), health assessment and management plans, allied health, mental health services and selected specialist attendances accessed in 91-day periods 12 months before and after LTC entry were examined. Adjusted relative changes in utilisation 0-3 months before and after LTC entry were estimated using risk ratios (RR) calculated using Generalised Estimating Equation Poisson models.

RESULTS

235,217 residents were included in the study with a median age of 84 years (interquartile range 79-89) and 61.1% female. In the first 3 months following LTC entry, GP / medical practitioner attendances increased from 86.6% to 95.6% (aRR 1.10 95%CI 1.10-1.11), GP / medical practitioner urgent after hours (from 12.3% to 21.1%; aRR 1.72, 95%CI 1.70-1.74) and after-hours attendances (from 18.5% to 33.8%; aRR 1.83, 95%CI 1.81-1.84) increased almost two-fold. Pain, palliative and geriatric specialist medicine attendances were low in the 3 months prior (<3%) and decreased further following LTC admission.

CONCLUSION

There is an opportunity to improve the utilisation of primary health care services following LTC entry to ensure that residents' increasingly complex care needs are adequately met.

A Longitudinal Examination of Post-COVID-19 Mortality in Residents in Long-Term Care Homes

The most adverse outcomes of the COVID-19 pandemic include high post-infection mortality among long-term care (LTC) home residents. Research about mortality over a longer period after contracting COVID-19 and in different pandemic years is limited. In the current study, we examined outcomes for 1,596 LTC residents from the day of a positive COVID-19 test until January 31, 2023. We reported all-cause mortality 30 days after contracting COVID-19 and monthly throughout the follow-up, up to 35 months after the pandemic start. We also examined mortality among 2,724 residents residing in the same LTC homes, with no history of COVID-19 during the same period. The results underscored a large number of deaths in the first month post-infection, with 30-day mortality substantially decreasing over the years-from 28% (95% CI [24.3, 31.8]) among residents contracting COVID-19 in 2020, to 8.3% (95% CI [7.4, 9.2]) in the 2022 cohort. Observed over longer periods, monthly mortality among residents with a COVID-19 history was similar to mortality in the No-COVID residents, and no evidence was found of increased mortality risk in the COVID group beyond the first post-infection month. We discuss mortality in LTC during the pandemic and a continuing need to reduce mortality in the acute phase of COVID-19.

Inter-provincial variation in older home care clients and their pathways: a population-based retrospective cohort study in Canada

BACKGROUND

In Canada, publicly-funded home care programs enable older adults to remain and be cared for in their home for as long as possible but they often differ in types of services offered, and the way services are delivered. This paper examines whether these differing approaches to care shape the pathway that home care clients will take. Older adult client pathways refer to trajectories within, and out of, the home care system (e.g., improvement, long term care (LTC) placement, death).

METHODS

A retrospective analysis of home care assessment data (RAI-HC was linked with health administrative data, long-term care admissions and vital statistics in Nova Scotia Health (NSH) and Winnipeg Regional Health Authority (WRHA). The study cohort consists of clients age 60 + years, admitted to home care between January 1, 2011 to December 31, 2013 and up to four years from baseline. Differences in home care service use, client characteristics and their pathways were tested across the two jurisdictions overall, and among the four discharge streams within jurisdictions using t-tests and chi-square tests of significance.

RESULTS

NS and WHRA clients were similar in age, sex, and marital status. NS clients had higher levels of need (ADL, cognitive impairment, CHESS) at base line and were more likely discharged to LTC (43% compared to 38% in WRHA). Caregiver distress was a factor correlated with being discharged to LTC. While a third remained as home care clients after 4 years; more than half were no longer in the community - either discharged to LTC placement or death. Such discharges occurred on average at around two years, a relatively short time period.

CONCLUSIONS

By following older clients over 4 years, we provide enhanced evidence of client pathways, the characteristics that influence these paths, as well as the length of time to the outcomes. This evidence is central to identification of clients at risk in the community and aids in planning for future home care servicing needs that will allow more older adults to remain living in the community.

Multistate Competing Risk Analysis of Transition Back to the Community Among Long-Term Care Home (LTC) Destined Patients: A Brief Report

OBJECTIVE

The demand for long-term care in community and facilitybased settings in Canada is expected to increase with population growth. The Toronto Grace Health Center piloted an intervention program that aims to support return to the community of acute hospital patients designated for LTC placement. We investigated whether this program was effective in transitioning the program patients back to their homes in the community and the factors associated with transitioning patients to different destinations.

METHOD

We performed a competing risk multi-state analysis of 111 patients enrolled into the Harbour Light (HL) transitional unit program between January 2020 and June 2023.

RESULTS

At the time of the study, 92 enrolled patients had been discharged and of those these, 48.9% (45) were successfully transitioned back to their private home in the community. The remaining 51.1% (46) were discharged to other destinations. Being a female was the only positive predictor of transitioning back home. Higher CPS scores (HR 0.53, 95% CI 0.31-0.88), PADDRS scale of 1+, and higher ADL Hierarchy scale, strongly predicted lower odds of transitioning back to the community.

CONCLUSION

Within the context of rising LTC bed demand and lengthy waiting time in Canada, with appropriate measures, this program successfully transitioned LTC home bound persons back to their homes. If replicable on a large scale, this could provide short and long-term solution to LTC bed demand in Canada.

Changes in Urinary Continence After Admission to a Complex Care Setting: A Multistate Transition Model

OBJECTIVES

To examine changes in urinary continence for post-acute, Complex Continuing Care hospital patients from time of admission to short-term follow-up, either in hospital or after discharge to long-term care or home with services.

DESIGN

Retrospective cohort study of patients in Complex Continuing Care hospitals using clinical data collected with interRAI Minimum Data Set 2.0 and interRAI Resident Assessment Instrument Home Care.

SETTING AND PARTICIPANTS

Adults aged 18 years and older, admitted to Complex Continuing Care hospitals in Ontario, Canada, between 2009 and 2015 (n = 78,913).

METHODS

A multistate transition model was used to characterize the association between patient characteristics measured at admission and changes in urinary continence state transitions (continent, sometimes continent, and incontinent) between admission and follow-up.

RESULTS

The cohort included 27,896 patients. At admission, 9583 (34.3%) patients belonged to the continent state, 6441 (23.09%) patients belonged to the sometimes incontinent state, and the remaining 11,872 (42.6%) patients belonged to the incontinent state. For patients who were continent at admission, the majority (62.7%) remained continent at follow-up. However, nearly a quarter (23.9%) transitioned to the sometimes continent state, and an additional 13.4% became incontinent at follow-up. Several factors were associated with continence state transitions, including cognitive impairment, rehabilitation potential, stroke, Parkinson's disease, Alzheimer's disease and related dementias, and hip fracture.

CONCLUSIONS AND IMPLICATIONS

This study suggests that urinary incontinence is a prevalent problem for Complex Continuing Care hospital patients and multiple factors are associated with continence state transitions. Standardized assessment of urinary incontinence is helpful in this setting to identify patients in need of further assessment and patient-centered intervention and as a quality improvement metric to examine changes in continence from admission to discharge.

Factors associated with increased Emergency Department transfer in older long-term care residents: a systematic review

The proportion of adults older than 65 years is rapidly increasing. Care home residents in this age group have disproportionate rates of transfer to the Emergency Department (ED) and around 40% of attendances might be avoidable. We did a systematic review to identify factors that predict ED transfer from care homes. Six electronic databases were searched. Observational studies that provided estimates of association between ED attendance and variables at a resident or care home level were included. 26 primary studies met the inclusion criteria. Seven common domains of factors assessed for association with ED transfer were identified and within these domains, male sex, age, presence of specific comorbidities, polypharmacy, rural location, and care home quality rating were associated with likelihood of ED transfer. The identification of these factors provides useful information for policy makers and researchers intending to either develop interventions to reduce hospitalisations or use adjusted rates of hospitalisation as a care home quality indicator.

Outcomes of advance care directives after admission to a long-term care home: DNR the DNH?

BACKGROUND

Residents of long-term care homes (LTCH) often experience unnecessary and non-beneficial hospitalizations and interventions near the end-of-life. Advance care directives aim to ensure that end-of-life care respects resident needs and wishes.

METHODS

In this retrospective cohort study, we used multistate models to examine the health trajectories associated with Do-Not-Resuscitate (DNR) and Do-Not-Hospitalize (DNH) directives of residents admitted to LTCH in Ontario, Alberta, and British Columbia, Canada. We adjusted for baseline frailty-related health instability. We considered three possible end states: change in health, hospitalization, or death. For measurements, we used standardized RAI-MDS 2.0 LTCH assessments linked to hospital records from 2010 to 2015.

RESULTS

We report on 123,003 LTCH residents. The prevalence of DNR and DNH directives was 71 and 26% respectively. Both directives were associated with increased odds of transitioning to a state of greater health instability and death, and decreased odds of hospitalization. The odds of hospitalization in the presence of a DNH directive were lowered, but not eliminated, with odds of 0.67 (95% confidence interval 0.65-0.69), 0.63 (0.61-0.65), and 0.47 (0.43-0.52) for residents with low, moderate and high health instability, respectively.

CONCLUSION

Even though both DNR and DNH orders are associated with serious health outcomes, DNH directives were not frequently used and often overturned. We suggest that policies recommending DNH directives be re-evaluated, with greater emphasis on advance care planning that better reflects resident values and wishes.

Predictors of hospitalisations and emergency department presentations shortly after entering a residential aged care facility in Australia: a retrospective cohort study

OBJECTIVES

To: (1) examine the 90-day incidence of unplanned hospitalisation and emergency department (ED) presentations after residential aged care facility (RACF) entry, (2) examine individual-related, facility-related, medication-related, system-related and healthcare-related predictors of these outcomes and (3) create individual risk profiles.

DESIGN

Retrospective cohort study using the Registry of Senior Australians. Fine-Gray models estimated subdistribution HRs and 95% CIs. Harrell's C-index assessed risk models' predictive ability.

SETTING AND PARTICIPANTS

Individuals aged ≥65 years old entering a RACF as permanent residents in three Australian states between 1 January 2013 and 31 December 2016 (N=116 192 individuals in 1967 RACFs).

PREDICTORS EXAMINED

Individual-related, facility-related, medication-related, system and healthcare-related predictors ascertained at assessments or within 90 days, 6 months or 1 year prior to RACF entry.

OUTCOME MEASURES

90-day unplanned hospitalisation and ED presentation post-RACF entry.

RESULTS

The cohort median age was 85 years old (IQR 80-89), 62% (N=71 861) were women, and 50.5% (N=58 714) had dementia. The 90-day incidence of unplanned hospitalisations was 18.0% (N=20 919) and 22.6% (N=26 242) had ED presentations. There were 34 predictors of unplanned hospitalisations and 34 predictors of ED presentations identified, 27 common to both outcomes and 7 were unique to each. The hospitalisation and ED presentation models out-of-sample Harrell's C-index was 0.664 (95% CI 0.657 to 0.672) and 0.655 (95% CI 0.648 to 0.662), respectively. Some common predictors of high risk of unplanned hospitalisation and ED presentations included: being a man, age, delirium history, higher activity of daily living, behavioural and complex care needs, as well as history, number and recency of healthcare use (including hospital, general practitioners attendances), experience of a high sedative load and several medications.

CONCLUSIONS

Within 90 days of RACF entry, 18.0% of individuals had unplanned hospitalisations and 22.6% had ED presentations. Several predictors, including modifiable factors, were identified at the time of care entry. This is an actionable period for targeting individuals at risk of hospitalisations.

Identifying Unexpected Deaths in Long-Term Care Homes

OBJECTIVES

Predicting unexpected deaths among long-term care (LTC) residents can provide valuable information to clinicians and policy makers. We study multiple methods to predict unexpected death, adjusting for individual and home-level factors, and to use as a step to compare mortality differences at the facility level in the future work.

DESIGN

We conducted a retrospective cohort study using Resident Assessment Instrument Minimum Data Set assessment data for all LTC residents in Ontario, Canada, from April 2017 to March 2018.

SETTING AND PARTICIPANTS

All residents in Ontario long-term homes. We used data routinely collected as part of administrative reporting by health care providers to the funder: Ontario Ministry of Health and Long-Term Care. This project is a component of routine policy development to ensure safety of the LTC system residents.

METHODS

Logistic regression (LR), mixed-effect LR (mixLR), and a machine learning algorithm (XGBoost) were used to predict individual mortality over 5 to 95 days after the last available RAI assessment.

RESULTS

We identified 22,419 deaths in the cohort of 106,366 cases (mean age: 83.1 years; female: 67.7%; dementia: 68.8%; functional decline: 16.6%). XGBoost had superior calibration and discrimination (C-statistic 0.837) over both mixLR (0.819) and LR (0.813). The models had high correlation in predicting death (LR-mixLR: 0.979, LR-XGBoost: 0.885, mixLR-XGBoost: 0.882). The inter-rater reliability between the models LR-mixLR and LR-XGBoost was 0.56 and 0.84, respectively. Using results in which all 3 models predicted probability of actual death of a resident at <5% yielded 210 unexpected deaths or 0.9% of the observed deaths.

CONCLUSIONS AND IMPLICATIONS

XGBoost outperformed other models, but the combination of 3 models provides a method to detect facilities with potentially higher rates of unexpected deaths while minimizing the possibility of false positives and could be useful for ongoing surveillance and quality assurance at the facility, regional, and national levels.

Assessments of heart failure and frailty-related health instability provide complementary and useful information for home care planning and prognosis

BACKGROUND

Health instability, measured with the Changes in Health and End-stage disease Signs and Symptoms (CHESS) scale, predicts hospitalizations and mortality in home care clients. Heart failure (HF) is also common among home care clients. We seek to understand how HF contributes to the odds of death, hospitalization or worsening health among new home care clients depending on admission health instability.

METHODS

We undertook a retrospective cohort study of home care clients aged 65 years and older between January 1st 2010 and March 31st 2015 from Alberta, British Columbia, Ontario, and the Yukon, Canada. We used multistate Markov models to derive adjusted odds ratios (OR) for transitions to different health instability states, hospitalization, and death. We examined the role of HF and CHESS at 6 months after home care admission.

RESULTS

The sample included 286,232 clients. Those with HF had greater odds of worsening health instability than those without HF. At low-moderate admission health instability (CHESS 0-2), clients with HF had greater odds of hospitalization and death than those without HF. Clients with HF and high health instability (CHESS≥3) had slightly greater odds of hospitalization (OR 1.08, 95% Confidence Interval 1.02-1.13) but similar odds of death (OR 1.024, 95% CI 0.937-1.120) compared to clients without HF.

CONCLUSIONS

Among new home care clients, a HF diagnosis predicts death, hospitalization and worsening health, predominantly among those with low-moderate admission health instability. A HF diagnosis and admission CHESS score provide complementary information to support care planning in this population.

Bringing Advances in Elder Abuse Research Methodology and Theory to Evaluation of Interventions

The elder abuse field needs high-quality intervention research to assess the best strategies to combat the problem, but few such studies exist. Significant advances have been made in elder abuse research methodology, measurement, justice theory, and other relevant areas, which may remedy this gap. Particular advances include the use of elder abuse severity measures and goal attainment scaling, and the application of restorative justice theory to intervention. Elder abuse research also may benefit from advances in biomedical clinical trial methodology, including characterizing and following up with subjects excluded from trials or those who drop out, estimating numbers needed to treat to assess cost-effectiveness, and a priori stopping rules for when an intervention proves quickly beneficial or harmful. In this article, we argue these advances can and should inform elder abuse intervention research, propose a theoretical framework to guide such efforts, and demonstrate how this framework can inform practice, using elder abuse multidisciplinary teams and nursing home shelters as examples.

Applying the Knowledge-to-Action Framework to Engage Stakeholders and Solve Shared Challenges with Person-Centered Advance Care Planning in Long-Term Care Homes

As they near the end of life, long term care (LTC) residents often experience unmet needs and unnecessary hospital transfers, a reflection of suboptimal advance care planning (ACP). We applied the knowledge-to-action framework to identify shared barriers and solutions to ultimately improve the process of ACP and improve end-of-life care for LTC residents. We held a 1-day workshop for LTC residents, families, directors/administrators, ethicists, and clinicians from Manitoba, Alberta, and Ontario. The workshop aimed to identify: (1) shared understandings of ACP, (2) barriers to respecting resident wishes, and (3) solutions to better respect resident wishes. Plenary and group sessions were recorded and thematic analysis was performed. We identified four themes: (1) differing provincial frameworks, (2) shared challenges, (3) knowledge products, and 4) ongoing ACP. Theme 2 had four subthemes: (i) lacking clarity on substitute decision maker (SDM) identity, (ii) lacking clarity on the SDM role, (iii) failing to share sufficient information when residents formulate care wishes, and (iv) failing to communicate during a health crisis. These results have informed the development of a standardized ACP intervention currently being evaluated in a randomized trial in three Canadian provinces.

Using a Semi-Markov Model to Estimate Medicaid Cost Savings due to Minnesota's Return to Community Initiative

OBJECTIVES

To provide an estimate and level of uncertainty for Medicaid cost savings due to Minnesota's Return to Community Initiative (RTCI).

DESIGN

Medicaid cost savings are estimated using a semi-Markov model and simulation approach.

SETTING AND PARTICIPANTS

RTCI is a statewide program that assists private paying nursing home residents with discharge to the community. When originally proposed, it was expected that the program would reduce state Medicaid expenditures, primarily through the shifting of residents from nursing homes to a less costly community setting. In prior analysis, we estimated that approximately 1 in 9 residents targeted for transition by the program would not have returned to the community without the RTCI. Accurate cost savings estimates require consideration of complex resident care trajectories, that is, nursing home readmissions, use of assisted living and community-based services, and mortality.

MEASURES

Data were from 30,234 private pay nursing home residents admitted during 2011, primarily for post-acute stays, to 378 facilities in Minnesota, and followed for 4 years postadmission for outcomes and time to event. Resident characteristics were taken from the Minimum Data Set (MDS) admission assessment. We modeled variability in care trajectories with a semi-Markov formulation. Transition probabilities were estimated using Multinomial regression. Time to event was modeled using the best-fitting, positive, right-skewed distribution for each path. The simulation was run (1000 times) with and without the RTCI impact to estimate change in Medicaid days in various settings.

RESULTS

Program savings was estimated at $4.1 million per year of effort over a 4-year accumulation period.

CONCLUSIONS AND IMPLICATIONS

The RTCI produced a modest Medicaid cost savings in excess of the annual program budget of $3.5 million. Findings from the semi-Markov model and simulation increase our understanding of care transitions between nursing home, community, Medicaid status, and mortality.