Predicting Hospital Readmissions from Home Healthcare in Medicare Beneficiaries

Predicting 30-day unplanned hospital readmission after revision total knee arthroplasty: machine learning model analysis of a national patient cohort

Revision total knee arthroplasty (TKA) is associated with a higher risk of readmission than primary TKA. Identifying individual patients predisposed to readmission can facilitate proactive optimization and increase care efficiency. This study developed machine learning (ML) models to predict unplanned readmission following revision TKA using a national-scale patient dataset. A total of 17,443 revision TKA cases (2013-2020) were acquired from the ACS NSQIP database. Four ML models (artificial neural networks, random forest, histogram-based gradient boosting, and k-nearest neighbor) were developed on relevant patient variables to predict readmission following revision TKA. The length of stay, operation time, body mass index (BMI), and laboratory test results were the strongest predictors of readmission. Histogram-based gradient boosting was the best performer in distinguishing readmission (AUC: 0.95) and estimating the readmission probability for individual patients (calibration slope: 1.13; calibration intercept: -0.00; Brier score: 0.064). All models produced higher net benefit than the default strategies of treating all or no patients, supporting the clinical utility of the models. ML demonstrated excellent performance for the prediction of readmission following revision TKA. Optimization of important predictors highlighted by our model may decrease preventable hospital readmission following surgery, thereby leading to reduced financial burden and improved patient satisfaction.

Factors Associated With Mortality and Hospice Use Among Medicare Beneficiaries With Heart Failure Who Received Home Health Services

BACKGROUND

Although many Medicare beneficiaries with heart failure (HF) are discharged with home health services, little is known about mortality rates and hospice use in this group.

OBJECTIVES

To identify risk factors for 6-month mortality and hospice use among patients hospitalized due to HF who receive home health care, which could inform efforts to improve palliative and hospice use for these patients.

METHODS

A retrospective cohort analysis was conducted in a 100% national sample of Medicare fee-for-service beneficiaries with HF who were discharged to home health care between 2017 and 2018. Multivariable Cox regression models examined factors associated with 6-month mortality, and multivariable logistic regression models examined factors associated with hospice use at the time of death.

RESULTS

A total of 285,359 Medicare beneficiaries were hospitalized with HF and discharged with home health care; 15.5% (44,174) died within 6 months. Variables most strongly associated with mortality included: age > 85 years (hazard ratio [HR] 1.66, 95% CI 1.61-1.71), urgent/emergency hospital admission (HR 1.68, 1.61-1.76), and "serious" condition compared to "stable" condition (HR 1.64, CI 1.52-1.78). Among 44,174 decedents, 48.2% (21,284) received hospice care at the time of death. Those with lower odds of hospice use at death included patients who were: < 65 years (odds ratio [OR] 0.65, CI 0.59-0.72); of Black (OR 0.64, CI 0.59-0.68) or Hispanic race/ethnicity (OR 0.79, CI 0.72-0.88); and Medicaid-eligible (OR 0.80, CI 0.76-0.85).

CONCLUSIONS

Although many patients hospitalized for HF are at risk of 6-month mortality and may benefit from palliative and/or hospice services, our findings indicate under-use of hospice care and important disparities in hospice use by race/ethnicity and socioeconomic status.

Predicting emergency department visits and hospitalizations for patients with heart failure in home healthcare using a time series risk model

OBJECTIVES

Little is known about proactive risk assessment concerning emergency department (ED) visits and hospitalizations in patients with heart failure (HF) who receive home healthcare (HHC) services. This study developed a time series risk model for predicting ED visits and hospitalizations in patients with HF using longitudinal electronic health record data. We also explored which data sources yield the best-performing models over various time windows.

MATERIALS AND METHODS

We used data collected from 9362 patients from a large HHC agency. We iteratively developed risk models using both structured (eg, standard assessment tools, vital signs, visit characteristics) and unstructured data (eg, clinical notes). Seven specific sets of variables included: (1) the Outcome and Assessment Information Set, (2) vital signs, (3) visit characteristics, (4) rule-based natural language processing-derived variables, (5) term frequency-inverse document frequency variables, (6) Bio-Clinical Bidirectional Encoder Representations from Transformers variables, and (7) topic modeling. Risk models were developed for 18 time windows (1-15, 30, 45, and 60 days) before an ED visit or hospitalization. Risk prediction performances were compared using recall, precision, accuracy, F1, and area under the receiver operating curve (AUC).

RESULTS

The best-performing model was built using a combination of all 7 sets of variables and the time window of 4 days before an ED visit or hospitalization (AUC = 0.89 and F1 = 0.69).

DISCUSSION AND CONCLUSION

This prediction model suggests that HHC clinicians can identify patients with HF at risk for visiting the ED or hospitalization within 4 days before the event, allowing for earlier targeted interventions.

Machine learning applied to electronic health record data in home healthcare: A scoping review

OBJECTIVE

Despite recent calls for home healthcare (HHC) to integrate informatics, the application of machine learning in HHC is relatively unknown. Thus, this study aimed to synthesize and appraise the literature describing the application of machine learning to predict adverse outcomes (e.g., hospitalization, mortality) using electronic health record (EHR) data in the HHC setting. Our secondary aim was to evaluate the comprehensiveness of predictors used in the machine learning algorithms guided by the Biopsychosocial Model.

METHODS

During March 2022 we conducted a literature search in four databases: PubMed, Embase, CINAHL, and Scopus. Inclusion criteria were 1) describing services provided in the HHC setting, 2) applying machine learning algorithms to predict adverse outcomes, defined as outcomes related to patient deterioration, 3) using EHR data and 4) focusing on the adult population. Predictors were mapped to the Biopsychosocial Model. A risk of bias analysis was conducted using the Prediction Model Risk Of Bias Assessment Tool.

RESULTS

The final sample included 20 studies. Eighteen studies used predictors from standardized assessments integrated in the EHR. The most common outcome of interest was hospitalization (55%), followed by mortality (25%). Psychological predictors were frequently excluded (35%). Tree based algorithms were most frequently applied (75%). Most studies demonstrated high or unclear risk of bias (75%).

CONCLUSION

Future studies in HHC should consider incorporating machine learning algorithms into clinical decision support systems to identify patients at risk. Based on the Biopsychosocial model, psychological and interpersonal characteristics should be used along with biological characteristics to enhance risk prediction. To facilitate the widespread adoption of machine learning, stakeholders should encourage standardization in the HHC setting.

Situation of the Elderly Living Alone: Morbidity and Services Provided from the Field of Primary Health Care of Gran Canaria

The elderly suffer a greater number of health problems and have greater need for assistance and care. (1) Background: to determine the profile of the elderly who live alone, identified according to the Primary Care Health Record of Gran Canaria, and to analyze the sociodemographic data of the target population and determine the characteristics related to morbidity. (2) Methods: descriptive, prospective, cross-sectional study carried out in the Primary Health Care Management of Gran Canaria. The study population was all adults over 65 years of age living alone. The instrument used was the Drago-Electronic Health Record. Data analysis was carried out using RStudio version 1.1.447 software, and descriptive analysis and inferential analysis were carried out using the Chi-square values, T-test for independent samples, and ANOVA. (3) Results: The sample amounted to 8679 subjects, predominantly female sex (86.14%) and with a mean age of 79.4 years. Of the sample, 6.4% lived alone. Based on the classification by Adjusted Morbidity Groups (AMG), subjects with “moderate complications” predominated at 45.5%. (4) Conclusions: It is necessary to implement this type of stratification tool, which allows interventions to be carried out in elderly people at risk.

“The Crossover to Hospice”: Perspectives of Home Healthcare Nurses and Social Workers

Background:

Although home healthcare(HHC) clinicians increasingly provide care to a homebound population with advanced illness and high symptom burden,  we know little about how HHC clinicians navigate discussions about hospice with patients and families in this setting.

Objective

We sought to explore perspectives on transition from HHC to hospice among HHC nurses and social workers.

Design

PQualitative study using semi-structured interviews and thematic analysis.

Results:

Fifteen nurses and 3 Social workers participated in the study. Four main themes emerged from the interviews: (1) Regulatory Forces of Hospice and HHC; (2) Structure of HHC; (3) Individual beliefs—Hospice means giving up; and (4) Dynamics of Communication in HHC to Facilitate Transitions to Hospice.

Conclusion

Introducing the option of hospice to patients and families nearing end-of-life in the HHC setting is complex and challenging.  Facilitators of hospice discussions in the HHC setting  include interdisciplinary team-based clinical review, clinical decision support tools to identify patients who are hospice-eligible, and staff training.  These factors provide targets for future interventions.

Documentation of hospitalization risk factors in electronic health records (EHRs): a qualitative study with home healthcare clinicians

OBJECTIVE

To identify the risk factors home healthcare (HHC) clinicians associate with patient deterioration and understand how clinicians respond to and document these risk factors.

METHODS

We interviewed multidisciplinary HHC clinicians from January to March of 2021. Risk factors were mapped to standardized terminologies (eg, Omaha System). We used directed content analysis to identify risk factors for deterioration. We used inductive thematic analysis to understand HHC clinicians' response to risk factors and documentation of risk factors.

RESULTS

Fifteen HHC clinicians identified a total of 79 risk factors that were mapped to standardized terminologies. HHC clinicians most frequently responded to risk factors by communicating with the prescribing provider (86.7% of clinicians) or following up with patients and caregivers (86.7%). HHC clinicians stated that a majority of risk factors can be found in clinical notes (ie, care coordination (53.3%) or visit (46.7%)).

DISCUSSION

Clinicians acknowledged that social factors play a role in deterioration risk; but these factors are infrequently studied in HHC. While a majority of risk factors were represented in the Omaha System, additional terminologies are needed to comprehensively capture risk. Since most risk factors are documented in clinical notes, methods such as natural language processing are needed to extract them.

CONCLUSION

This study engaged clinicians to understand risk for deterioration during HHC. The results of our study support the development of an early warning system by providing a comprehensive list of risk factors grounded in clinician expertize and mapped to standardized terminologies.

Symptom Burden and Activity of Daily Living (ADL) Dependency Among Home Health care Patients Discharged to Home Hospice

Background: We sought to examine sociodemographic and clinical characteristics present on admission to HHC associated with discharge to hospice. Methods: We used a 5% random sample of 2017 national Outcome and Assessment Information Set (OASIS) data. A Cox proportional hazards regression model was estimated for the primary outcome (discharge to hospice) to examine the associations with sociodemographic and clinical characteristics of HHC patients. Results: Among 489, 230 HHC patients, 4268 were discharged to hospice. The median (interquartile range) length of HHC stay for patients discharged to hospice care was 33 (14-78) days. Compared to White patients, Black, Hispanic, and other race, (hazard ratio [HR] = .50 [95% confidence interval, CI = .44-.57]), (HR = .53 [95% CI = .46-.62]), and (HR = .49 [95% CI = .40-.61], respectively) was associated with shorter time to discharge to hospice care. Clinical characteristics including severe dependence in activities of daily (ADL) (HR = 1.68 [95% CI = 1.01-2.78]), cognitive impairment (HR = 1.10 [95% CI = 1.01-1.20]), disruptive behavior daily (HR = 1.11 [95% CI = 1.02-1.22]), and inability to feed oneself (HR = 4.78, 95% CI = 4.30, 5.31) was associated with shorter time to discharge to hospice. Symptoms of anxiety daily (HR = 1.55 [95% CI = 1.43-1.68]), and pain daily or all the time (HR = 1.54 [95% CI = 1.43-1.64]) were associated with shorter time to discharge to hospice. Conclusions: High symptom burden, ADL dependency, and cognitive impairment on admission to HHC services was associated with greater likelihood of discharge to hospice.

Association between home health agency ownership status and discharge to community among Medicare beneficiaries

To investigate the association of ownership status, discharge rate and length of stay (LOS) of home health care (HH) services under the prospective payment system (PPS). We used 2016-2018 Outcome Assessment and Information Set (OASIS) data sets for Medicare beneficiaries. Two outcome variables were investigated: rate of discharge from an HH agency and LOS. Our main independent variable was ownership status: for-profit (FP) versus not-for-profit (NFP). FP agencies were 4.2% (p <.01) less likely to discharge patients to the community but more likely (7.3%; p <.001) to have longer LOS (>99 days) compared to NFPs. Findings that FP agencies were less likely to discharge patients to the community and more likely to have a longer length of stay than NFP agencies have implications for quality of care initiatives by the Medicare Post-Acute Transformation Act 2014.

Artificial Learning and Machine Learning Decision Guidance Applications in Total Hip and Knee Arthroplasty: A Systematic Review

Background

Artificial intelligence (AI) and machine learning (ML) modeling in hip and knee arthroplasty (total joint arthroplasty [TJA]) is becoming more commonplace. This systematic review aims to quantify the accuracy of current AI- and ML-based application for cognitive support and decision-making in TJA.

Methods

A comprehensive search of publications was conducted through the EMBASE, Medline, and PubMed databases using relevant keywords to maximize the sensitivity of the search. No limits were placed on level of evidence or timing of the study. Findings were reported according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. Analysis of variance testing with post-hoc Tukey test was applied to compare the area under the curve (AUC) of the models.

Results

After application of inclusion and exclusion criteria, 49 studies were included in this review. The application of AI/ML-based models and average AUC is as follows: cost prediction-0.77, LOS and discharges-0.78, readmissions and reoperations-0.66, preoperative patient selection/planning-0.79, adverse events and other postoperative complications-0.84, postoperative pain-0.83, postoperative patient-reported outcomes measures and functional outcome-0.81. Significant variability in model AUC across the different decision support applications was found (P < .001) with the AUC for readmission and reoperation models being significantly lower than that of the other decision support categories.

Conclusions

AI/ML-based applications in TJA continue to expand and have the potential to optimize patient selection and accurately predict postoperative outcomes, complications, and associated costs. On average, the AI/ML models performed best in predicting postoperative complications, pain, and patient-reported outcomes and were less accurate in predicting hospital readmissions and reoperations.

Postacute care outcomes in home health or skilled nursing facilities in patients with a diagnosis of dementia

OBJECTIVE

To compare the outcomes of postacute care between home health (HH) and skilled nursing facilities (SNFs) following hospitalization among Medicare beneficiaries with a diagnosis of dementia.

DATA SOURCES

100% MedPAR data, Minimum Data Set, and Outcome and Assessment Information Set assessment data from January 1, 2015 to December 31, 2016.

STUDY DESIGN

Retrospective cohort analysis using an instrumental variable design to compare outcomes (30-day readmission and mortality, 100-day mortality) of HH versus SNF following acute hospitalization. We used the differential distance between patients' home and the closest HH agency and SNF to instrument for nonrandom allocation of patients.

DATA COLLECTION/EXTRACTION METHODS

We identified hospital discharges followed by SNF and HH stays for Medicare fee-for-service beneficiaries with dementia. We excluded beneficiaries younger than age 65, admitted to the hospital from a nursing home, or enrolled in hospice. We identified dementia using validated diagnostic codes with a 3-year look-back.

PRINCIPAL FINDINGS

Our sample included 977,946 beneficiaries with a diagnosis of dementia; 297,732 (30.4%) received HH, while 680,214 (69.6%) went to SNF. Overall, 16.8% were readmitted to the hospital and 6.1% died within 30 days, while 15.4% died within 100 days of hospital discharge. In the instrumental variable analysis, there were no differences in any outcome between the two postacute care settings.

CONCLUSIONS

Medicare beneficiaries with a diagnosis of dementia receiving postacute care in HH or SNF experienced similar rates of readmission and mortality across settings. This finding raises important questions about current postacute care referral patterns, given 7 in 10 patients with a diagnosis of dementia in our sample were discharged to SNF.

Application of machine learning in predicting hospital readmissions: a scoping review of the literature

Background

Advances in machine learning (ML) provide great opportunities in the prediction of hospital readmission. This review synthesizes the literature on ML methods and their performance for predicting hospital readmission in the US.

Methods

This review was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis Extension for Scoping Reviews (PRISMA-ScR) Statement. The extraction of items was also guided by the Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies (CHARMS). Electronic databases PUBMED, MEDLINE, and EMBASE were systematically searched from January 1, 2015, through December 10, 2019. The articles were imported into COVIDENCE online software for title/abstract screening and full-text eligibility. Observational studies using ML techniques for hospital readmissions among US patients were eligible for inclusion. Articles without a full text available in the English language were excluded. A qualitative synthesis included study characteristics, ML algorithms utilized, and model validation, and quantitative analysis assessed model performance. Model performances in terms of Area Under the Curve (AUC) were analyzed using R software. Quality in Prognosis Studies (QUIPS) tool was used to assess the quality of the reviewed studies.

Results

Of 522 citations reviewed, 43 studies met the inclusion criteria. A majority of the studies used electronic health records (24, 56%), followed by population-based data sources (15, 35%) and administrative claims data (4, 9%). The most common algorithms were tree-based methods (23, 53%), neural network (NN) (14, 33%), regularized logistic regression (12, 28%), and support vector machine (SVM) (10, 23%). Most of these studies (37, 85%) were of high quality. A majority of these studies (28, 65%) reported ML algorithms with an AUC above 0.70. There was a range of variability within AUC reported by these studies with a median of 0.68 (IQR: 0.64–0.76; range: 0.50–0.90).

Conclusions

The ML algorithms involving tree-based methods, NN, regularized logistic regression, and SVM are commonly used to predict hospital readmission in the US. Further research is needed to compare the performance of ML algorithms for hospital readmission prediction.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12874-021-01284-z.

Unmet family caregiver training needs associated with acute care utilization during home health care

BACKGROUND/OBJECTIVES

Medicare-certified home health agencies are required to offer family caregiver training, but little is known regarding the potential impact of this training on outcomes during home health care. We estimate the proportion of family caregivers assisting Medicare home health patients who have unmet training needs and examine whether these unmet training needs are associated with older adults' risk of acute care utilization during home health care.

DESIGN

Observational, nationally representative cohort study.

SETTING

Linked National Health and Aging Trends Study, Outcome and Assessment Information Set (OASIS), Medicare Provider of Services file, and Medicare claims data from 2011 to 2016.

PARTICIPANTS

Thousand two hundred seventeen (weighted n = 5,870,905) community-living Medicare beneficiaries who received home health care between 2011 and 2016.

MEASUREMENTS

Family caregivers' unmet training needs measured from OASIS and Medicare claims; home health patients' acute care utilization (including emergency department use and hospitalization) measured from OASIS.

RESULTS

Rates of unmet need for training varied by activity, from 8.2% of family caregivers assisting with household chores to 16.0% assisting with self-care tasks. After controlling for older adult and home health provider characteristics, older adults whose family caregivers had an unmet need for training with any caregiving activity were twice as likely to incur acute care utilization during their home health episode (adjusted odds ratio [aOR]: 2.01, 95% confidence interval [CI]: 1.20-3.38). This relationship held across specific caregiving activities including household chores (aOR: 1.98; 95% CI: 1.13-3.46), medication management (aOR: 2.50; 95% CI: 1.46-4.26), patient supervision (aOR: 2.92; 95% CI: 1.36-6.24), and self-care tasks (aOR: 3.11; 95% CI: 1.62-6.00).

CONCLUSIONS AND RELEVANCE

Unmet training needs among family caregivers are associated with greater likelihood of acute care utilization among Medicare beneficiaries receiving home health care. Identifying and addressing family caregivers' training needs may reduce older adults' risk of acute care utilization during home health care.

LACE Score-Based Risk Management Tool for Long-Term Home Care Patients: A Proof-of-Concept Study in Taiwan

Background: Effectively predicting and reducing readmission in long-term home care (LTHC) is challenging. We proposed, validated, and evaluated a risk management tool that stratifies LTHC patients by LACE predictive score for readmission risk, which can further help home care providers intervene with individualized preventive plans. Method: A before-and-after study was conducted by a LTHC unit in Taiwan. Patients with acute hospitalization within 30 days after discharge in the unit were enrolled as two cohorts (Pre-Implement cohort in 2017 and Post-Implement cohort in 2019). LACE score performance was evaluated by calibration and discrimination (AUC, area under receiver operator characteristic (ROC) curve). The clinical utility was evaluated by negative predictive value (NPV). Results: There were 48 patients with 87 acute hospitalizations in Pre-Implement cohort, and 132 patients with 179 hospitalizations in Post-Implement cohort. These LTHC patients were of older age, mostly intubated, and had more comorbidities. There was a significant reduction in readmission rate by 44.7% (readmission rate 25.3% vs. 14.0% in both cohorts). Although LACE score predictive model still has room for improvement (AUC = 0.598), it showed the potential as a useful screening tool (NPV, 87.9%; 95% C.I., 74.2–94.8). The reduction effect is more pronounced in infection-related readmission. Conclusion: As real-world evidence, LACE score-based risk management tool significantly reduced readmission by 44.7% in this LTHC unit. Larger scale studies involving multiple homecare units are needed to assess the generalizability of this study.

Using Machine Learning to Predict Rehabilitation Outcomes in Postacute Hip Fracture Patients

OBJECTIVE

To use machine learning-based methods in designing a predictive model of rehabilitation outcomes for postacute hip fracture patients.

DESIGN

A retrospective analysis using linear models, AdaBoost, CatBoost, ExtraTrees, K-Nearest Neighbors, RandomForest, Support vector machine, XGBoost, and voting of all models to develop and validate a predictive model.

SETTING

A university-affiliated 300-bed major postacute geriatric rehabilitation center.

PARTICIPANTS

Consecutive hip fracture patients (N=1625) admitted to an postacute rehabilitation department.

MAIN OUTCOME MEASURES

The FIM instrument, motor FIM (mFIM), and the relative functional gain on mFIM (mFIM effectiveness) as a continuous and binary variable. Ten predictive models were created: base models (linear/logistic regression), and 8 machine learning models (AdaBoost, CatBoost, ExtraTrees, K-Nearest Neighbors, RandomForest, Support vector machine, XGBoost, and a voting ensemble). R² was used to evaluate their performance in predicting a continuous outcome variable, and the area under the receiver operating characteristic curve was used to evaluate the binary outcome. A paired 2-tailed t test compared the results of the different models.

RESULTS

Machine learning-based models yielded better results than the linear and logistic regression models in predicting rehabilitation outcomes. The 3 most important predictors of the mFIM effectiveness score were the Mini Mental State Examination (MMSE), prefracture mFIM scores, and age. The 3 most important predictors of the discharge mFIM score were the admission mFIM, MMSE, and prefracture mFIM scores. The most contributing factors for favorable outcomes (mFIM effectiveness > median) with higher prediction confidence level were high MMSE (25.7±2.8), high prefacture mFIM (81.5±7.8), and high admission mFIM (48.6±8) scores. We present a simple prediction instrument for estimating the expected performance of postacute hip fracture patients.

CONCLUSIONS

The use of machine learning models to predict rehabilitation outcomes of postacute hip fracture patients is superior to linear and logistic regression models. The higher the MMSE, prefracture mFIM, and admission mFIM scores are, the higher the confidence levels of the predicted parameters.

Skilled Nursing Facility Patients Discharged to Home Health Agency Services Spend More Days at Home

OBJECTIVES

To investigate the association of the utilization of Medicare-certified home health agency (CHHA) services with post-acute skilled nursing facility (SNF) discharge outcomes that included home time, rehospitalization, SNF readmission, and mortality.

DESIGN

Retrospective cohort study.

SETTING

New York State fee-for-service Medicare beneficiaries aged 65 years and older admitted to SNFs for post-acute care and discharged to the community in 2014.

PARTICIPANTS

A total of 25,357 older adults.

MEASUREMENTS

The outcomes included days spent alive in the community ("home time"), rehospitalization, SNF readmission, and mortality within 30- and 90-day post-SNF discharge periods. The primary independent variables were SNF five-star overall quality rating and receipt of CHHA services within 7 days of SNF discharge. Zero-inflated negative binomial regression and logistic regression models characterized the association of CHHA linkage with home time and other outcomes, respectively.

RESULTS

Following SNF discharge, 17,657 (69.6%) patients received CHHA services. In analyses that adjusted for patient-, market-, and other SNF-level factors, older adults discharged from higher quality SNFs were more likely to receive CHHA services. In analyses that adjusted for patient- and market-level factors, receipt of post-SNF CHHA services was associated with 2.03 and 4.17 (P < .001) more days in the community over 30- and 90-day periods. Receiving CHHA services was also associated with decreased odds for rehospitalization (odds ratio [OR] = .68; P < .001; OR = .91; P = .008), SNF readmission (OR = .36; P < .001; OR = .62; P < .001), and death (OR = .34; P < .001; OR = .63; P < .001) over 30- and 90-day periods, respectively.

CONCLUSION

Among older adults discharged from a post-acute SNF stay, those who received CHHA services had better discharge outcomes. They were less likely to experience admissions to institutional care settings and had a lower mortality risk. Future efforts that examine how the type and intensity of CHHA services affect outcomes would build on this work. J Am Geriatr Soc 68:1573-1578, 2020.