Dataset Information

Developing a Machine Learning Risk-adjustment Method for Hospitalizations and Emergency Department Visits of Nursing Home Residents With Dementia.

ABSTRACT:

Background

Long-stay nursing home (NH) residents with Alzheimer disease and related dementias (ADRD) are at high risk of hospital transfers. Machine learning might improve risk-adjustment methods for NHs.

Objectives

The objective of this study was to develop and compare NH risk-adjusted rates of hospitalizations and emergency department (ED) visits among long-stay residents with ADRD using Extreme Gradient Boosting (XGBoost) and logistic regression.

Research design

Secondary analysis of national Medicare claims and NH assessment data in 2012 Q3. Data were equally split into the training and test sets. Both XGBoost and logistic regression predicted any hospitalization and ED visit using 58 predictors. NH-level risk-adjusted rates from XGBoost and logistic regression were constructed and compared. Multivariate regressions examined NH and market factors associated with rates of hospitalization and ED visits.

Subjects

Long-stay Medicare residents with ADRD (N=413,557) from 14,057 NHs.

Results

A total of 8.1% and 8.9% residents experienced any hospitalization and ED visit in a quarter, respectively. XGBoost slightly outperformed logistic regression in area under the curve (0.88 vs. 0.86 for hospitalization; 0.85 vs. 0.83 for ED visit). NH-level risk-adjusted rates from XGBoost were slightly lower than logistic regression (hospitalization=8.3% and 8.4%; ED=8.9% and 9.0%, respectively), but were highly correlated. Facility and market factors associated with the XGBoost and logistic regression-adjusted hospitalization and ED rates were similar. NHs serving more residents with ADRD and having a higher registered nurse-to-total nursing staff ratio had lower rates.

Conclusions

XGBoost and logistic regression provide comparable estimates of risk-adjusted hospitalization and ED rates.

SUBMITTER: Xu H

PROVIDER: S-EPMC10526959 | biostudies-literature | 2023 Sep

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Developing a Machine Learning Risk-adjustment Method for Hospitalizations and Emergency Department Visits of Nursing Home Residents With Dementia.

Xu Huiwen H Bowblis John R JR Becerra Adan Z AZ Intrator Orna O

Medical care 20230713 9

<h4>Background</h4>Long-stay nursing home (NH) residents with Alzheimer disease and related dementias (ADRD) are at high risk of hospital transfers. Machine learning might improve risk-adjustment methods for NHs.<h4>Objectives</h4>The objective of this study was to develop and compare NH risk-adjusted rates of hospitalizations and emergency department (ED) visits among long-stay residents with ADRD using Extreme Gradient Boosting (XGBoost) and logistic regression.<h4>Research design</h4>Secondar ...[more]

PMID: 37440719

Similar Datasets

Project description:BackgroundNursing home (NH) residents are frequent users of emergency departments (ED) and while prior research suggests that repeat visits are common, there is little data describing this phenomenon. Our objectives were to describe repeat ED visits over one year, identify risk factors for repeat use, and characterize "frequent" ED visitors.MethodsUsing provincial administrative data from Ontario, Canada, we identified all NH residents 65 years or older who visited an ED at least once between January 1 and March 31, 2010 and then followed them for one year to capture all additional ED visits. Frequent ED visitors were defined as those who had 3 or more repeat ED visits. We used logistic regression to estimate risk factors for any repeat ED visit and for being a frequent visitor and Andersen-Gill regression to estimate risk factors for the rate of repeat ED visits.ResultsIn a cohort of 25,653 residents (mean age 84.5 (SD?=?7.5) years, 68.2% female), 48.8% had at least one repeat ED visit. Residents who experienced a repeat ED visit were generally similar to others but they tended to be slightly younger, have a higher proportion male, and a higher proportion with minimal cognitive or physical impairment. Risk factors for a repeat ED visit included: being male (adjusted odds ratio 1.27, (95% confidence interval 1.19-1.36)), diagnoses such as diabetes (AOR 1.28 (1.19-1.37)) and congestive heart failure (1.26 (1.16-1.37)), while severe cognitive impairment (AOR 0.92 (0.84-0.99)) and 5 or more chronic conditions (AOR 0.82 (0.71-0.95)) appeared protective. Eleven percent of residents were identified as frequent ED visitors, and they were more often younger then 75 years, male, and less likely to have Alzheimer's disease or other dementias than non-frequent visitors.ConclusionsRepeat ED visits were common among NH residents but a relatively small group accounted for the largest number of visits. Although there were few clear defining characteristics, our findings suggest that medically complex residents and younger residents without cognitive impairments are at risk for such outcomes.

Project description:BackgroundNursing home residents (NHRs) are often transferred to emergency departments (EDs). A great proportion of ED visits is considered inappropriate. There is evidence that male NHRs are more often hospitalised, but this is less clear for ED visits. It is unclear, which influence age has on ED visits. We aimed to study the epidemiology of ED visits in NHRs focusing on age- and sex-specific differences.MethodsA systematic review was carried out based on articles found in MEDLINE (via PubMed), CINAHL and Scopus. Articles published on or before Aug 31, 2017 were eligible. Two reviewers independently identified articles for inclusion. The quality of studies was assessed by the Joanna Briggs Institute critical appraisal tool for prevalence studies.ResultsOut of 1192 references, we found seven studies meeting our inclusion criteria. Six studies were conducted in the USA or Canada. Overall, 29-62% of NHRs had at least one ED visit over the course of 1 year. Most studies assessing the influence of sex found that male residents visited EDs more frequently. All but one of the five studies with multivariable analyses reported a statistically significant positive association (with odds or rate ratios of 1.05-1.38). All studies assessed the influence of age. There was no clear pattern with some studies showing no association between ED visits and age and other studies reporting decreasing ED visits with increasing age or increasing proportions followed by a decrease in the highest age group. Studies used 85+ or 86+ years as the highest age category. Hospital admission rate ranged from 36.4 to 48.7%. There was no study reporting stratified analyses by age and sex. Only one study reported main diagnoses leading to ED visits stratified by sex.ConclusionMale NHRs visit EDs more often than females, but there is no evidence on reasons. The association with age is unclear. Any future study on acute care of NHRs should assess the influence of age and sex. These studies should include large sample sizes to provide a more differentiated age categorisation.Trial registrationPROSPERO CRD42017074845 .

Project description:ObjectiveExplore within and across nursing home (NH) racial disparities in end-of-life (EOL) hospitalizations for residents with Alzheimer's disease or related dementia (ADRD), and examine whether severe cognitive impairment influences these relationships.DesignObservational study merging, at the individual level, C2014-2017 national-level Minimum Data Set (MDS), Medicare Beneficiary Summary Files (MBSF), and Medicare Provider Analysis and Review (MedPAR). Nursing Home Compare (NHC) was also used.SettingLong-stay residents who died in a NH or a hospital within 8 days of discharge.ParticipantsAnalytical sample included 665,033 decedent residents with ADRD in 14,595 facilities.Main outcomes and measuresThe outcome was hospitalization within 30 days of death. Key independent variables were race, severe cognitive impairment, and NH-level proportion of black residents. Other covariates included socio-demographics, dual eligibility, hospice enrollment, and chronic conditions. Facility-level characteristics were also included (e.g. profit status, staffing hours, etc.). We fit linear probability models with robust standard errors, fixed and random effects.ResultsCompared to whites, black decedents had a significantly (p < 0.01) higher risk of EOL hospitalizations (7.88%). Among those with severe cognitive impairment, whites showed a lower risk of hospitalizations (6.04%). But EOL hospitalization risk among blacks with severe cognitive impairment was still significantly elevated (β = 0.0494; p < 0.01). A comparison of the base model with the fixed and random-effects models showed statistically significant hospitalization risk by decedent's race both within and across facilities.Conclusions and relevanceWe found disparities between black and white residents with ADRD both within and across facilities. The within-facility disparities may be due to residents' preferences and/or NH practices that contribute to differential treatment. The across facility differences point to the overall quality of care disparities in homes with a higher prevalence of black residents. Persistence of such systemic disparities among the most vulnerable individuals is extremely troubling.

Project description:BackgroundEmergency department (ED) visits for nursing home residents lead to higher morbidity and mortality. Therefore, inappropriate visits (for conditions treatable elsewhere) or potentially avoidable visits (those avoidable through adequate chronic care management) must be minimized. This study aimed to investigate factors and resource consumption patterns associated with inappropriate and potentially avoidable visits in a Swiss tertiary hospital.MethodsThis is a single-center retrospective chart review in an urban Swiss university hospital ED. A consecutive sample of 1276 visits by nursing home residents (≥ 65 years old), recorded between January 1, 2015 and December 31, 2017 (three calendar years) were included. Case characteristics were extracted from ED electronic documentation. Appropriateness was assessed via a structured Appropriateness Evaluation Protocol; potentially avoidable visits-measured as ambulatory-care sensitive conditions (ACSCs)-were analyzed separately. Inter-group differences concerning ED resource use were tested respectively with chi-square or Wilcoxon rank sum tests. To identify predictors of inappropriate or potentially-avoidable visits, we used multivariable logistic regression analysis.ResultsSix percent of visits were rated as inappropriate: they had lower triage levels (OR 0.55 [95%-CI 0.33-0.92], p=0.024) and, compared to ambulance calls, they had higher odds of initiation via either patient-initiated walk-in (OR 3.42 [95%-CI 1.79-6.55], p≤0.001) or GP referrals (OR 2.13 [95%-CI 1.16-3.90], p=0.015). For inappropriate visits, overall ED resource use was significantly lower (median 568 vs. 1403 tax points, p≤0.001). Of all visits included, 29% were due to (often potentially-avoidable) ACSCs. In those cases, compared to ambulance initiation, odds of being potentially-avoidable were considerably lower for walk-in patients (OR 0.46 [95%-CI 0.27-0.77], p=0.004) but higher for GP referrals (OR 1.40 [95%-CI 1.00-1.94], p=0.048). Nurse work (93 tax points vs. 64, p≤0.001) and laboratory resource use (334 tax points vs. 214, p≤0.001) were higher for potentially-avoidable ED visits.ConclusionsWe revealed substantial differences between the investigated groups. While nearly one third of ED visits from nursing homes were potentially avoidable, inappropriate visits were lower in numbers and not resource-intensive. Further research is required to differentiate potentially avoidable visits from inappropriate ones and to determine these findings' public health implications.

Project description:Background: Antipsychotics are commonly prescribed to people living with dementia in nursing home settings, despite strong guideline recommendations against their use except in limited circumstances. We aimed to transparently describe the development process for a complex intervention targeting appropriate requesting and prescribing of antipsychotics to nursing home residents with dementia in Ireland, by nurses and general practitioners (GPs) respectively. Methods: We report the development process for the 'Rationalising Antipsychotic Prescribing in Dementia' (RAPID) complex intervention, in accordance with the 'Guidance for reporting intervention development studies in health research' (GUIDED) checklist. The UK Medical Research Council framework for developing and evaluating complex interventions guided our overall approach, incorporating evidence and theory into the intervention development process. To unpack the intervention development process in greater detail, we followed the Behaviour Change Wheel approach. Guided by our stakeholders, we conducted three sequential studies (systematic review and qualitative evidence synthesis, primary qualitative study and expert consensus study), to inform the intervention development. Results: The RAPID complex intervention was developed in collaboration with a broad range of stakeholders, including people living with dementia and family carers, between 2015 and 2017. The finalised RAPID complex intervention was comprised of the following three components; 1) Education and training sessions with nursing home staff; 2) Academic detailing with GPs; 3) Introduction of an assessment tool to the nursing home. Conclusions: This paper describes the steps used by the researchers to develop a complex intervention targeting antipsychotic prescribing to nursing home residents with dementia in Ireland, according to the GUIDED checklist. We found that the GUIDED checklist provided a useful way of reporting all elements in a cohesive manner and complemented the other tools and frameworks used. Transparency in the intervention development processes can help in the translation of evidence into practice.

Project description:ImportanceSevere maternal morbidity (SMM) can have long-term health consequences for the affected mother. The association between SMM and future maternal mental health conditions has not been well studied.ObjectiveTo assess the association between SMM in the first recorded birth and the risk of hospitalization or emergency department (ED) visits for a mental health condition over a 13-year period.Design, setting, and participantsThis population-based retrospective cohort study used data from postpartum individuals aged 18 to 55 years with a first hospital delivery between 2008 and 2021 in 11 provinces and territories in Canada, except Québec. Data were analyzed from January to June 2023.ExposureSMM, defined as a composite of conditions, such as septic shock, severe preeclampsia or eclampsia, severe hemorrhage with intervention, or other complications, occurring after 20 weeks' gestation and up to 42 days after a first delivery.Main outcomes and measuresThe main outcome was a hospitalization or ED visit for a mental health condition, including mood and anxiety disorders, substance use, schizophrenia, and other psychotic disorder, or suicidality or self-harm event, arising at least 43 days after the first birth hospitalization. Cox regression models generated hazard ratios with 95% CIs, adjusted for baseline maternal comorbidities, maternal age at delivery, income quintile, type of residence, hospital type, and delivery year.ResultsOf 2 026 594 individuals with a first hospital delivery, 1 579 392 individuals (mean [SD] age, 30.0 [5.4] years) had complete ED and hospital records and were included in analyses; among these, 35 825 individuals (2.3%) had SMM. Compared with individuals without SMM, those with SMM were older (mean [SD] age, 29.9 [5.4] years vs 30.7 [6.0] years), were more likely to deliver in a teaching tertiary care hospital (40.8% vs 51.1%), and to have preexisting conditions (eg, ≥2 conditions: 1.2% vs 5.3%), gestational diabetes (8.2% vs 11.7%), stillbirth (0.5% vs 1.6%), preterm birth (7.7% vs 25.0%), or cesarean delivery (31.0% vs 54.3%). After a median (IQR) duration of 2.6 (1.3-6.4) years, 1287 (96.1 per 10 000) individuals with SMM had a mental health hospitalization or ED visit, compared with 41 779 (73.2 per 10 000) individuals without SMM (adjusted hazard ratio, 1.26 [95% CI, 1.19-1.34]).Conclusions and relevanceIn this cohort study of postpartum individuals with and without SMM in pregnancy and delivery, there was an increased risk of mental health hospitalizations or ED visits up to 13 years after a delivery complicated by SMM. Enhanced surveillance and provision of postpartum mental health resources may be especially important after SMM.

Dataset Information

Developing a Machine Learning Risk-adjustment Method for Hospitalizations and Emergency Department Visits of Nursing Home Residents With Dementia.

Background

Objectives

Research design

Subjects

Results

Conclusions

Publications

Developing a Machine Learning Risk-adjustment Method for Hospitalizations and Emergency Department Visits of Nursing Home Residents With Dementia.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets