Predictors of Unplanned Hospital Readmissions and Emergency Department Revisits in Patients with Acute Ischemic Stroke

 Where is the research that will prevent these readmissions? Complete failure of stroke leadership to address this problem!

Predictors of Unplanned Hospital Readmissions and Emergency Department Revisits in Patients with Acute Ischemic Stroke

Ammar Jumah^1*Jennifer Ro¹Tianwen Ma²Javacia Owens¹Yunong Wu²Danielle Starnes¹Hannah Christopher¹Debra Blanke¹Laura Henriquez¹Samir R. Belagaje¹Nino Kvantaliani¹Jacquelyn Cabral¹Teri Walczak¹Judy Collier¹Kourtni Mack¹Fadi Nahab¹

• ¹Department of Neurology, Emory University, Atlanta, GA, United States
• ²Department of Biostatistics and Bioinformatics, Rollins School of Public Health, Emory University, Atlanta, GA, United States

Objectives: We aim to identify factors associated with emergency department (ED) revisits and hospital readmissions after acute ischemic stroke (AIS) diagnosis and to determine if early outpatient follow-up can reduce readmissions.

Methods: We retrospectively identified all AIS patients discharged from a hospital network, from October 1, 2022 to March 31, 2024. Baseline characteristics, inpatient metrics and post-discharge outpatient follow-up were assessed to identify factors associated with ED revisits and readmissions to the healthcare system within 90-days.

Results: Of 1,973 patients, 464 (23.5%) had ED visits within 90 days and 263 (13.3%) had hospital readmission within 90 days. The median age was 68 [IQR 58, 77]. In multiple logistic regression analyses, factors independently associated with 90-day ED visit were history of heart failure (HF) (OR 1.46, 95% CI 1.11–1.93; p = 0.007), diabetes mellitus (DM) (OR 1.41, 95% CI 1.12–1.77; p = 0.003), atrial fibrillation (AF) (OR 1.47, 95% CI 1.13–1.92, p = 0.004) and an increasing Charlson comorbidity index (CCI) score (OR 1.10, 95% CI 1.03–1.18), p = 0.003). Factors associated with 90-day readmission were HF (OR 1.51, 95% CI 1.08–2.11, p = 0.015), DM (OR 1.50, 95% CI 1.13–2.01, p = 0.006), AF (OR 1.40, 95% CI 1.00–1.94, p = 0.047) and increasing CCI score (OR 1.12, 95% CI 1.03–1.21, p = 0.006). Discharge to inpatient rehabilitation or skilled nursing facility (vs. home or home health) were associated with 90-day ED revisits and hospital readmissions. Patients who completed early (<30 days) outpatient stroke clinic follow-up had a lower likelihood of 90-day readmission (OR 0.68, 95% CI 0.52–0.90; p = 0.006).

Conclusion: Patients with certain comorbidities including HF, DM, AF and those with a higher CCI score have a higher likelihood of a 90-day ED revisit and hospital readmission. Unplanned hospital readmissions may be preventable with early outpatient visits in a dedicated stroke clinic after discharge for AIS patients.

Transitions of care clinic improves 30-day stroke outcomes

Better but NOT GOOD ENOUGH!, should be trying for 100% recovery. Since they are not measuring that, they have given up on trying to get you 100% recovered.

Transitions of care clinic improves 30-day stroke outcomes

By Joe Gramigna, MA

Source/Disclosures

ADD TOPIC TO EMAIL ALERTS

Use of a transitions of care clinic may reduce 30-day stroke readmissions, according to findings presented at the American Neurological Association annual meeting.

“Between 8% and 14.5% of stroke patients are readmitted within 30 days of discharge, on average, which obviously leads to worse patient outcomes,” Evan Kolesnick, MS, of Philadelphia College of Osteopathic Medicine, said during a virtual presentation. “Transitions of care clinic is an intervention which is designed for the patient to be seen shortly after hospital discharge in order to help them better manage their condition. It has been shown to reduce readmissions for various conditions by clarifying diagnosis and treatment, improving medication adherence and assuaging some patient concerns.”

Kolesnick and colleagues sought to determine whether a transitions of care (TOC) clinic reduced stroke readmission at Thomas Jefferson University Hospital (TJUH) in Pennsylvania. Specifically, they aimed to examine potential differences between 30-day stroke readmission with the TOC clinic and the top quartile of stroke-performing hospitals, as well as with TJUH’s general service. Other study aims included comparing in-person and telehealth TOC clinic results and identifying improvement areas via self-report measures of self-efficacy and perceived TOC value.

Researchers used ICD-10 codes that corresponded with ischemic stroke, hemorrhagic stroke and transient ischemic attack to identify 194 patients (92 in-person patients; 102 telehealth patients) for inclusion in quantitative analysis. In January 2019, patients received automatic scheduling for in-person TOC appointment 1 to 3 weeks after discharge. These transitioned to telehealth visits in February 2020. Kolesnick and colleagues collected readmission data via the electronic health record and delivered questionnaires via telephone within 6 months to assess self-efficacy and perceived value.

Readmissions occurred among seven (3.61%) patients, of whom one (1.09%) was seen in person and six (5.88%) via telehealth. Results of Chi-square analysis revealed no difference between in-person and telehealth performance. Chi-square comparisons showed no distinct difference between TOC and TJUH’s general service or the top quartile of stroke-performing hospitals in terms of 30-day readmissions. However, Kolesnick and colleagues noted the TOC clinic improved by 1.04% upon TJUH’s figure of 4.65% and 2.67% upon the top-quartile 2-year readmission figure of 6.28%.

The TOC clinic combined figure of 3.61% outperformed most recent CMS data (11.8%) between 2015 and 2016, according to the study abstract. The readmission rate was higher in 2020 compared with 2019, with possible causes related to effects of the COVID-19 pandemic on care. However, the pandemic did not appear to affect the number of monthly stroke discharges to home. Both in-person and telehealth groups reported satisfaction with TOC.

Read next

Rivaroxaban superior to warfarin for stroke protection in real-world nonvalvular AF population
Mortality, stroke rates after TAVR do not vary by age at 2 years

“Future directions include translating our findings into practical institution policy,” Kolesnick said during the presentation. “We also want to perform qualitative study of the barriers to telemedicine stroke clinic and make the platform more accessible to patients in the future. We intend to continue our data collection in the post-COVID era to determine if the pandemic confounded some of our findings.”

 

Prediction of 30-Day Readmission After Stroke Using Machine Learning and Natural Language Processing

All you have to do is change one word in your research and you would have done something useful. Prediction to Prevention.

Prediction of 30-Day Readmission After Stroke Using Machine Learning and Natural Language Processing

Christina M. Lineback¹, Ravi Garg², Elissa Oh², Andrew M. Naidech^1,2, Jane L. Holl² and Shyam Prabhakaran^3*

• ¹Department of Neurology, Feinberg School of Medicine, Northwestern University, Chicago, IL, United States
• ²Department of Neurology, Biological Sciences, Division and Center for Healthcare Delivery Science and Innovation, University of Chicago, Chicago, IL, United States
• ³Department of Neurology, University of Chicago, Chicago, IL, United States

Background and Purpose: This study aims to determine whether machine learning (ML) and natural language processing (NLP) from electronic health records (EHR) improve the prediction of 30-day readmission after stroke.

Methods: Among index stroke admissions between 2011 and 2016 at an academic medical center, we abstracted discrete data from the EHR on demographics, risk factors, medications, hospital complications, and discharge destination and unstructured textual data from clinician notes. Readmission was defined as any unplanned hospital admission within 30 days of discharge. We developed models to predict two separate outcomes, as follows: (1) 30-day all-cause readmission and (2) 30-day stroke readmission. We compared the performance of logistic regression with advanced ML algorithms. We used several NLP methods to generate additional features from unstructured textual reports. We evaluated the performance of prediction models using a five-fold validation and tested the best model in a held-out test dataset. Areas under the curve (AUCs) were used to compare discrimination of each model.

Results: In a held-out test dataset, advanced ML methods along with NLP features out performed logistic regression for all-cause readmission (AUC, 0.64 vs. 0.58; p < 0.001) and stroke readmission prediction (AUC, 0.62 vs. 0.52; p < 0.001).

Conclusion: NLP-enhanced machine learning models potentially advance our ability to predict readmission after stroke. However, further improvement is necessary before being implemented in clinical practice given the weak discrimination.

Introduction

Nearly 800,000 patients experience a stroke each year in the USA (1). The cost of initial admissions for stroke averages US$20,000 while readmissions cost on average US$10,000 (1–3). Reduction in readmission is, thus, an important target to reduce healthcare costs and improve patient care. However, several studies have demonstrated that available prediction models for readmission perform modestly (4, 5). A better understanding of the causes leading to readmission and better prediction tools may allow hospital systems to better allocate resources to the patients who are most at risk for readmission (6, 7).

Prior efforts to stratify risk of readmission have utilized basic statistical models, such as logistic regression, with modest results (AUC range: 0.53–0.67) (5, 7, 8). However, these studies do not report results on a separate held out dataset thereby not addressing the generalizability of these results. Also, since these methods are trained and validated on the same datasets, the results are highly prone to be inflated due to overfitting. Furthermore, logistic regression base models are incapable of properly weighing the interactions between the complex variables in additive analyses (4, 9).

Machine learning (10) (ML) has emerged as a new statistical approach to overcome the limitation of non-linearity and improve predictive analysis in healthcare. Advanced ML methods have shown to be superior for predicting readmission in patients with heart failure (11). Furthermore, natural language processing (NLP) methods can be utilized to automatically extract much of the rich but difficult-to-access medical information that is often buried in unstructured text notes within electronic health records (EHR). There has been widespread interest to use ML in conjunction with NLP to build clinical tools for cohort construction, clinical trials, and clinical decision support (9, 12). There has been, however, no study to use NLP of clinical notes and ML to predict readmissions after stroke. We, therefore, sought to evaluate advanced ML algorithms that incorporate NLP features of textual data in the EHR to improve prediction of 30-day readmission after stroke. We also seek to evaluate our models on a separate held out dataset in order to test the generalizability of our results.

More at link.

 

 

Protocol of a randomized controlled trial investigating the effectiveness of Recovery-focused Community support to Avoid readmissions and improve Participation after Stroke (ReCAPS)

I see nothing here that suggests that your doctors have a protocol for their tasks to prevent readmissions. That is totally evading their responsibility to get you 100% recovered.

Protocol of a randomized controlled trial investigating the effectiveness of Recovery-focused Community support to Avoid readmissions and improve Participation after Stroke (ReCAPS) 

Show all authors

Dominique A Cadilhac, Jan Cameron, Monique F Kilkenny, ...

First Published June 14, 2021 Product Review Find in PubMed 

https://doi.org/10.1177/17474930211022678

Article information

Abstract

Rationale

To address unmet needs, electronic messages to support person-centered goal attainment and secondary prevention may avoid hospital presentations/readmissions after stroke, but evidence is limited.

Hypothesis

Compared to control participants, there will be a 10% lower proportion of intervention participants who represent to hospital (emergency/admission) within 90 days of randomization.

Methods and design

Multicenter, double-blind, randomized controlled trial with intention-to-treat analysis. The intervention group receives 12 weeks of personalized, goal-centered, and administrative electronic messages, while the control group only receive administrative messages. The trial includes a process evaluation, assessment of treatment fidelity, and an economic evaluation. Participants: Confirmed stroke (modified Rankin Score: 0-4), aged ≥18 years with internet/mobile phone access, discharged directly home from hospital. Randomization: 1:1 computer-generated, stratified by age and baseline disability. Outcomes assessments: Collected at 90 days and 12 months following randomization.

Outcomes

Primary outcomes include hospital emergency presentations/admissions within 90 days of randomization. Secondary outcomes include goal attainment, self-efficacy, mood, unmet needs, disability, quality-of-life, recurrent stroke/cardiovascular events/deaths at 90 days and 12 months, and death and cost-effectiveness at 12 months. Sample size: To test our primary hypothesis, we estimated a sample size of 890 participants (445 per group) with 80% power and two-tailed significance threshold of α = 0.05. Given uncertainty for the effect size of this novel intervention, the sample size will be adaptively re-estimated when outcomes for n = 668 are obtained, with maximum sample capped at 1100.

Discussion

We will provide new evidence on the potential effectiveness, implementation, and cost-effectiveness of a tailored eHealth intervention for survivors of stroke.

Keywords

Stroke, eHealth, self-management, healthcare technology, clinical trial protocol

Access Options