Hospital readmission in general medicine patients: a prediction model

Do Postsurgical Follow-Up Calls Reduce Unplanned 30-Day Readmissions in Neurosurgery Patients? A Quality Improvement Project in a University Hospital

BACKGROUND

Unplanned 30-day readmissions after surgery are a source of patient dissatisfaction, monitored by the Centers for Medicare and Medicaid Services, have financial penalties for hospitals, and are publicly reported. Neurosurgical operations have a higher 30-day unplanned readmission rate after the index discharge than other specialties. After a simple initiative for a 48-72-hour postdischarge telephone call, there was an observed significant decrease in readmission rates from 17% to 8% in 7 months at Thomas Jefferson University. To better understand the role of postoperative telephone calls in this reduction, a retrospective evaluation over a longer period was performed.

METHODS

A quality improvement initiative was assessed using patient records between August 2018 and May 2023. The primary observed subject is the 30-day unplanned readmission rate and secondarily a change in Physician Communication Score. Thirty-day unplanned readmission rate and Physician Communication Scores before and after the telephone call initiative were compared, checking for difference, variance, and correlation.

RESULTS

874 readmissions (average, 28/month; 95% confidence interval [CI], 25.3-29.3), 12.9% (95% CI, 11.9-13.9) were reported before the telephone call; of 673 readmissions (average, 26/month; 95% CI, 23-28.8), 12.9% (95% CI, 11.6-14.1) were reported after the telephone call. No significant difference, variance of scores or rates, or correlation of rate with communication score were noted before and after the initiative.

CONCLUSIONS

Telephone calls and peridischarge efficient communication are needed after neurologic surgery. This approach decreased unplanned readmissions in certain instances without having a significant impact on neurosurgical patients.

A simple prediction model for the risk of boron neutron capture therapy-induced nausea and vomiting in head and neck cancer

BACKGROUND AND PURPOSE

Head and neck cancer patients undergoing boron neutron capture therapy (BNCT) often experience BNCT-induced nausea and vomiting (BINV). This study aimed to construct a BINV risk prediction model.

MATERIALS AND METHODS

In this retrospective study, 237 patients were randomly divided into a training and test cohort. In the training cohort, a univariate analysis was performed to identify factors associated with BINV. Multivariate analysis was used to identify factors and calculate coefficients for the model. The Hosmer-Lemeshow test was used to assess the goodness of fit, and receiver operating characteristic curves were plotted to evaluate the accuracy of the model. For both the training and test sets, the predictive model was used to generate the scores and calculate the sensitivity and specificity.

RESULTS

The incidence of nausea and vomiting was 50 % and 18 %, respectively. Female gender, younger age, non-squamous cell carcinoma, no prior chemotherapy, and beam entry from the face/lateral region were associated with the occurrence of BINV. The prediction model showed a good fit (P = 0.96) and performance (area under the curve = 0.75). The sensitivity and specificity were 83 % and 45 % for the training cohort (n = 193) and 86 % and 59 % for the test cohort (n = 44), respectively.

CONCLUSION

We developed a simple model that predicts BINV. This will enable appropriate care to be implemented based on increased risk to prevent its occurrence.

Can prediction models for hospital readmission be improved by incorporating patient-reported outcome measures? A systematic review and narrative synthesis

PURPOSE

To investigate the roles, challenges, and implications of using patient-reported outcome measures (PROMs) in predicting the risk of hospital readmissions.

METHODS

We systematically searched four bibliometric databases for peer-reviewed studies published in English between 1 January 2000 and 15 June 2023 and used validated PROMs to predict readmission risks for adult populations. Reported studies were analysed and narratively synthesised in accordance with the CHARMS and PRISMA guidelines.

RESULTS

Of the 2858 abstracts reviewed, 23 studies met predefined eligibility criteria, representing diverse geographic regions and medical specialties. Among those, 19 identified the positive contributions of PROMs in predicting readmission risks. Seven studies utilised generic PROMs exclusively, eleven used generic and condition-specific PROMs, while 5 focussed solely on condition-specific PROMs. Logistic regression was the most used modelling approach, with 13 studies aiming at predicting 30-day all-cause readmission risks. The c-statistic, ranging from 0.54 to 0.84, was reported in 22/23 studies as a measure of model discrimination. Nine studies reported model calibration in addition to c-statistic. Thirteen studies detailed their approaches to dealing with missing data.

CONCLUSION

Our study highlights the potential of PROMs to enhance predictive accuracy in readmission models, while acknowledging the diversity in data collection methods, readmission definitions, and model evaluation approaches. Recognizing that PROMs serve various purposes beyond readmission reduction, our study supports routine data collection and strategic integration of PROMs in healthcare practices to improve patient outcomes. To facilitate comparative analysis and broaden the use of PROMs in the prediction framework, it is imperative to consider the methodological aspects involved.

Using Machine Learning Models to Identify Factors Associated With 30-Day Readmissions After Posterior Cervical Fusions: A Longitudinal Cohort Study

OBJECTIVE

Readmission rates after posterior cervical fusion (PCF) significantly impact patients and healthcare, with complication rates at 15%-25% and up to 12% 90-day readmission rates. In this study, we aim to test whether machine learning (ML) models that capture interfactorial interactions outperform traditional logistic regression (LR) in identifying readmission-associated factors.

METHODS

The Optum Clinformatics Data Mart database was used to identify patients who underwent PCF between 2004-2017. To determine factors associated with 30-day readmissions, 5 ML models were generated and evaluated, including a multivariate LR (MLR) model. Then, the best-performing model, Gradient Boosting Machine (GBM), was compared to the LACE (Length patient stay in the hospital, Acuity of admission of patient in the hospital, Comorbidity, and Emergency visit) index regarding potential cost savings from algorithm implementation.

RESULTS

This study included 4,130 patients, 874 of which were readmitted within 30 days. When analyzed and scaled, we found that patient discharge status, comorbidities, and number of procedure codes were factors that influenced MLR, while patient discharge status, billed admission charge, and length of stay influenced the GBM model. The GBM model significantly outperformed MLR in predicting unplanned readmissions (mean area under the receiver operating characteristic curve, 0.846 vs. 0.829; p < 0.001), while also projecting an average cost savings of 50% more than the LACE index.

CONCLUSION

Five models (GBM, XGBoost [extreme gradient boosting], RF [random forest], LASSO [least absolute shrinkage and selection operator], and MLR) were evaluated, among which, the GBM model exhibited superior predictive performance, robustness, and accuracy. Factors associated with readmissions impact LR and GBM models differently, suggesting that these models can be used complementarily. When analyzing PCF procedures, the GBM model resulted in greater predictive performance and was associated with higher theoretical cost savings for readmissions associated with PCF complications.

Automated Text Message-Based Program and Use of Acute Health Care Resources After Hospital Discharge: A Randomized Clinical Trial

Importance

Postdischarge outreach from the primary care practice is an important component of transitional care support. The most common method of contact is via telephone call, but calls are labor intensive and therefore limited in scope.

Objective

To test whether a 30-day automated texting program to support primary care patients after hospital discharge reduces acute care revisits.

Design, Setting, and Participants

A 2-arm randomized clinical trial was conducted from March 29, 2022, through January 5, 2023, at 30 primary care practices within a single academic health system in Philadelphia, Pennsylvania. Patients were followed up for 60 days after discharge. Investigators were blinded to assignment, but patients and practice staff were not. Participants included established patients of the study practices who were aged 18 years or older, discharged from an acute care hospitalization, and considered medium to high risk for adverse health events by a health system risk score. All analyses were conducted using an intention-to-treat approach.

Intervention

Patients in the intervention group received automated check-in text messages from their primary care practice on a tapering schedule for 30 days following discharge. Any needs identified by the automated messaging platform were escalated to practice staff for follow-up via an electronic medical record inbox. Patients in the control group received a standard transitional care management telephone call from their practice within 2 business days of discharge.

Main Outcomes and Measures

The primary study outcome was any acute care revisit (readmission or emergency department visit) within 30 days of discharge.

Results

Of the 4736 participants, 2824 (59.6%) were female; the mean (SD) age was 65.4 (16.5) years. The mean (SD) length of index hospital stay was 5.5 (7.9) days. A total of 2352 patients were randomized to the intervention arm and 2384 were randomized to the control arm. There were 557 (23.4%) acute care revisits in the control group and 561 (23.9%) in the intervention group within 30 days of discharge (risk ratio, 1.02; 95% CI, 0.92-1.13). Among the patients in the intervention arm, 79.5% answered at least 1 message and 41.9% had at least 1 need identified.

Conclusions and Relevance

In this randomized clinical trial of a 30-day postdischarge automated texting program, there was no significant reduction in acute care revisits.

Trial Registration

ClinicalTrials.gov Identifier: NCT05245773.

Clinical outcomes and predictors for 30-day readmissions of endoscopic retrograde cholangiopancreatography in the United States

BACKGROUND/OBJECTIVES

We aimed to assess 30-day readmissions of endoscopic retrograde cholangiopancreatography (ERCP) in the United States.

METHODS

The National Readmission Database was utilized from 2016 to 2020 to identify 30-day readmissions of ERCP. Hospitalization characteristics and outcomes were compared between index hospitalizations and readmissions. Predictors of 30-day readmission and mortality were also identified.

RESULTS

Between 2016 and 2020, 885 416 index hospitalizations underwent ERCP. Of these, 88 380 (10.15%) were readmitted within 30 days. Compared to index hospitalizations, 30-day readmissions had higher mean age (63.76 vs 60.8 years, P < 0.001) and proportion of patients with Charlson Comorbidity Index (CCI) score ≥3 (48.26% vs 29.91%, P < 0.001). Sepsis was the most common readmission diagnosis. Increasing age, male gender, higher CCI scores, admissions at large metropolitan teaching hospitals, cholecystectomy on index hospitalization, biliary stenting, increasing length of stay (LOS) at index admission, post-ERCP pancreatitis, post-ERCP hemorrhage, and gastrointestinal tract perforation were independent predictors of 30-day readmissions. Furthermore, 30-day readmissions had higher odds of inpatient mortality (4.42% vs 1.66%, aOR 1.9, 95% CI: 1.79-2.01, P < 0.001) compared to index hospitalizations. However, we noted a shorter LOS (5.78 vs 6.22 days, mean difference 1.2, 95% CI: 1.12-1.28, P < 0.001) and lower total hospital charge ($71 076 vs $93 418, mean difference $31 452, 95% CI: 29 835-33 069, P < 0.001) for 30-day readmissions compared to index hospitalizations. Increasing age, higher CCI scores, increasing LOS, biliary stenting, and post-ERCP hemorrhage were independent predictors of inpatient mortality for 30-day readmissions.

CONCLUSION

After index ERCP, the 30-day remission rate was 10.15%. Compared to index hospitalizations, 30-day readmissions had higher odds of inpatient mortality.

Association of vena cava filters and catheter-directed thrombolysis for deep vein thrombosis with hospital readmissions

BACKGROUND

Acute deep vein thrombosis (DVT) affects >350,000 patients each year in the United States. Contemporary rehospitalization rates and predictors of acute DVT have not been well-characterized. We aimed to evaluate the all-cause 30-day readmission rate and its association with catheter-directed thrombolysis and vena cava filters in patients with proximal and caval DVT.

METHODS

Patients with an index hospitalization for acute proximal lower extremity DVT were evaluated for unplanned readmission rates at 30 days using the Nationwide Readmission Database from 2016 to 2017. We used Cox proportional hazard model to determine the predictors of 30-day readmissions and their association with inferior vena cava (IVC) filter and CDT use.

RESULTS

We identified 58,306 adult patients with an index hospitalization for acute proximal DVT. The unplanned 30-day rehospitalization rate was 14.7% (95% confidence interval [CI], 14.5-15.0%). There were 4995 patients (10.0%) who underwent CDT and 6085 (12.2%) who underwent IVC filter placement. In multivariable analysis, only CDT was associated with a lower hazard for rehospitalization (hazard ratio [HR], 0.77; 95% CI, 0.71-0.84; P < .001), whereas IVC filter placement (HR, 1.26; 95% CI, 1.19-1.34; P < .001), Charlson Comorbidity Index of >3 (HR, 1.47; 95% CI, 1.38-1.56; P < .001), malignancy (HR, 1.45; 95% CI, 1.34-1.57; P < .001), and length of stay >5 days (HR, 1.39; 95% CI, 1.33-1.46; P < .001), and acute kidney injury (HR, 1.18; 95% CI, 1.11-1.25; P < .001) were associated with higher readmission rates.

CONCLUSIONS

The 30-day unplanned rehospitalization rate continues to be high in patients with acute proximal DVT. CDT was associated with lower rehospitalization rates, whereas IVC filter placement was associated with increased rehospitalization rates.

Identifying the Prevalence and Causes of 30-Day Hospital Readmission in Children: A Case Study from a Tertiary Pediatric Hospital

Introduction

The objectives of this study were to determine the prevalence of unplanned readmissions in the pediatric population within 30 days of discharge, identify the possible reasons behind them, and develop a predictive model for unplanned admissions.

Methods

A retrospective chart review study of 25,211 patients was conducted to identify the prevalence of readmissions occurring within 30 days of discharge from the King Abdullah Specialized Children's Hospital (KASCH) in Riyadh, Saudi Arabia, between Jan 1, 2019, and Dec 31, 2021. The data were collected using the BestCare electronic health records system and analyzed using Jamovi statistical software version 1.6.

Results

Among the 25,211 patients admitted to the hospital during the study period, the prevalence of unplanned readmission within 30 days was 1291 (5.12%). Of the 1291 patients, 1.91% had subsequent unplanned readmissions. In 57.8% of the cases, the cause of the first unplanned readmission was related to the cause of the first admission, and in 90.64% of the cases, the cause of the subsequent unplanned readmission was related to the cause of the first unplanned readmission. The most common reason for the first unplanned readmission was postoperative complications (18.75%), whereas pneumonia (10.81%) was the most common reason for subsequent unplanned readmissions. Most patients with subsequent unplanned readmissions were also found to have either isolated central nervous system pathology or chronic complex medical conditions.

Conclusion

Internationally, the rate of unplanned readmissions in pediatric patients has been estimated to be 6.5% within 30 days, which is comparable to the results of our study (5.12%). Most of the causes of first and subsequent unplanned readmission were found to be related to primary admission. The diagnosis/causes of readmission vary depending on the patient's age. A predictive model for pediatric readmission should be established so that preventive measures can be implemented.

Does a Community Care programme reach a high health need population and high users of acute care hospital services in Melbourne, Australia? An observational cohort study

OBJECTIVE

The Community Care programme is an initiative aimed at reducing hospitalisations and emergency department (ED) presentations among patients with complex needs. We aimed to describe the characteristics of the programme participants and identify factors associated with enrolment into the programme.

DESIGN

This observational cohort study was conducted using routinely collected data from the National Centre for Healthy Ageing data platform.

SETTING

The study was carried out at Peninsula Health, a health service provider serving a population in Melbourne, Victoria, Australia.

PARTICIPANTS

We included all adults with unplanned ED presentation or hospital admission to Peninsula Health between 1 November 2016 and 31 October 2017, the programme's first operational year.

OUTCOME MEASURES

Community Care programme enrolment was the primary outcome. Participants' demographics, health factors and enrolment influences were analysed using a staged multivariable logistic regression.

RESULTS

We included 47 148 adults, of these, 914 were enrolled in the Community Care programme. Participants were older (median 66 vs 51 years), less likely to have a partner (34% vs 57%) and had more frequent hospitalisations and ED visits. In the multivariable analysis, factors most strongly associated with enrolment included not having a partner (adjusted OR (aOR) 1.83, 95% CI 1.57 to 2.12), increasing age (aOR 1.01, 95% CI 1.01 to 1.02), frequent hospitalisations (aOR 7.32, 95% CI 5.78 to 9.24), frequent ED visits (aOR 2.0, 95% CI 1.37 to 2.85) and having chronic diseases, such as chronic pulmonary disease (aOR 2.48, 95% CI 2.06 to 2.98), obesity (aOR 2.06, 95% CI 1.39 to 2.99) and diabetes mellitus (complicated) (aOR 1.75, 95% CI 1.44 to 2.13). Residing in aged care home and having high socioeconomic status) independently associated with reduced odds of enrolment.

CONCLUSIONS

The Community Care programme targets patients with high-readmission risks under-representation of individuals residing in residential aged care homes warrants further investigation. This study aids service planning and offers valuable feedback to clinicians about programme beneficiaries.

Readmission Rates After Acute Respiratory Distress Syndrome in Children

Importance

An increasing number of children survive after acute respiratory distress syndrome (ARDS). The long-term morbidity affecting these survivors, including the burden of hospital readmission and key factors associated with readmission, is unknown.

Objective

To determine 1-year readmission rates among survivors of pediatric ARDS and to investigate the associations of 3 key index hospitalization factors (presence or development of a complex chronic condition, receipt of a tracheostomy, and hospital length of stay [LOS]) with readmission.

Design, Setting, and Participants

This retrospective cohort study used data from the commercial or Medicaid IBM MarketScan databases between 2013 and 2017, with follow-up data through 2018. Participants included hospitalized children (aged ≥28 days to <18 years) who received mechanical ventilation and had algorithm-identified ARDS. Data analysis was completed from March 2022 to March 2023.

Exposures

Complex chronic conditions (none, nonrespiratory, and respiratory), receipt of tracheostomy, and index hospital LOS.

Main Outcomes and Measures

The primary outcome was 1-year, all-cause hospital readmission. Univariable and multivariable Cox proportional hazard models were created to test the association of key hospitalization factors with readmission.

Results

One-year readmission occurred in 3748 of 13 505 children (median [IQR] age, 4 [0-14] years; 7869 boys [58.3%]) with mechanically ventilated ARDS who survived to hospital discharge. In survival analysis, the probability of 1-year readmission was 30.0% (95% CI, 29.0%-30.8%). One-half of readmissions occurred within 61 days of discharge (95% CI, 56-67 days). Both respiratory (adjusted hazard ratio [aHR], 2.69; 95% CI, 2.42-2.98) and nonrespiratory (aHR, 1.86; 95% CI, 1.71-2.03) complex chronic conditions were associated with 1-year readmission. Placement of a new tracheostomy (aHR, 1.98; 95% CI, 1.69-2.33) and LOS 14 days or longer (aHR, 1.87; 95% CI, 1.62-2.16) were associated with readmission. After exclusion of children with chronic conditions, LOS 14 days or longer continued to be associated with readmission (aHR, 1.92; 95% CI, 1.49-2.47).

Conclusions and Relevance

In this retrospective cohort study of children with ARDS who survived to discharge, important factors associated with readmission included the presence or development of chronic medical conditions during the index admission, tracheostomy placement during index admission, and index hospitalization of 14 days or longer. Future studies should evaluate whether postdischarge interventions (eg, telephonic contact, follow-up clinics, and home health care) may help reduce the readmission burden.

Identification of Factors Associated With 30-day Readmissions After Posterior Lumbar Fusion Using Machine Learning and Traditional Models: A National Longitudinal Database Study

STUDY DESIGN

A retrospective cohort study.

OBJECTIVE

To identify the factors associated with readmissions after PLF using machine learning and logistic regression (LR) models.

SUMMARY OF BACKGROUND DATA

Readmissions after posterior lumbar fusion (PLF) place significant health and financial burden on the patient and overall health care system.

MATERIALS AND METHODS

The Optum Clinformatics Data Mart database was used to identify patients who underwent posterior lumbar laminectomy, fusion, and instrumentation between 2004 and 2017. Four machine learning models and a multivariable LR model were used to assess factors most closely associated with 30-day readmission. These models were also evaluated in terms of ability to predict unplanned 30-day readmissions. The top-performing model (Gradient Boosting Machine; GBM) was then compared with the validated LACE index in terms of potential cost savings associated with the implementation of the model.

RESULTS

A total of 18,981 patients were included, of which 3080 (16.2%) were readmitted within 30 days of initial admission. Discharge status, prior admission, and geographic division were most influential for the LR model, whereas discharge status, length of stay, and prior admissions had the greatest relevance for the GBM model. GBM outperformed LR in predicting unplanned 30-day readmission (mean area under the receiver operating characteristic curve 0.865 vs. 0.850, P <0.0001). The use of GBM also achieved a projected 80% decrease in readmission-associated costs relative to those achieved by the LACE index model.

CONCLUSIONS

The factors associated with readmission vary in terms of predictive influence based on standard LR and machine learning models used, highlighting the complementary roles these models have in identifying relevant factors for the prediction of 30-day readmissions. For PLF procedures, GBM yielded the greatest predictive ability and associated cost savings for readmission.

LEVEL OF EVIDENCE

3.

Effects of a Multimodal Transitional Care Intervention in Patients at High Risk of Readmission: The TARGET-READ Randomized Clinical Trial

Importance

Hospital readmissions are frequent, costly, and sometimes preventable. Although these issues have been well publicized and incentives to reduce them introduced, the best interventions for reducing readmissions remain unclear.

Objectives

To evaluate the effects of a multimodal transitional care intervention targeting patients at high risk of hospital readmission on the composite outcome of 30-day unplanned readmission or death.

Design, Setting, and Participants

A single-blinded, multicenter randomized clinical trial was conducted from April 2018 to January 2020, with a 30-day follow-up in 4 medium-to-large-sized teaching hospitals in Switzerland. Participants were consecutive patients discharged from general internal medicine wards and at higher risk of unplanned readmission based on their simplified HOSPITAL score (≥4 points). Data were analyzed between April and September 2022.

Interventions

The intervention group underwent systematic medication reconciliation, a 15-minute patient education session with teach-back, a planned first follow-up visit with their primary care physician, and postdischarge follow-up telephone calls from the study team at 3 and 14 days. The control group received usual care from their hospitalist, plus a 1-page standard study information sheet.

Main Outcomes and Measures

Thirty-day postdischarge unplanned readmission or death.

Results

A total of 1386 patients were included with a mean (SD) age of 72 (14) years; 712 (51%) were male. The composite outcome of 30-day unplanned readmission or death was 21% (95% CI, 18% to 24%) in the intervention group and 19% (95% CI, 17% to 22%) in the control group. The intention-to-treat analysis risk difference was 1.7% (95% CI, -2.5% to 5.9%; P = .44). There was no evidence of any intervention effects on time to unplanned readmission or death, postdischarge health care use, patient satisfaction with the quality of their care transition, or readmission costs.

Conclusions and Relevance

In this randomized clinical trial, use of a standardized multimodal care transition intervention targeting higher-risk patients did not significantly decrease the risks of 30-day postdischarge unplanned readmission or death; it demonstrated the difficulties in preventing hospital readmissions, even when multimodal interventions specifically target higher-risk patients.

Trial Registration

ClinicalTrials.gov Identifier: NCT03496896.

Effectiveness of Transition Care Intervention Targeted to High-Risk Patients to Reduce Readmissions: Study Protocol for the TARGET-READ Multicenter Randomized-Controlled Trial

Hospital readmissions within 30 days represent a burden for the patients and the entire health care system. Improving the care around hospital discharge period could decrease the risk of avoidable readmissions. We describe the methods of a trial that aims to evaluate the effect of a structured multimodal transitional care intervention targeted to higher-risk medical patients on 30-day unplanned readmissions and death. The TARGET-READ study is an investigator-initiated, pragmatic single-blinded randomized multicenter controlled trial with two parallel groups. We include all adult patients at risk of hospital readmission based on a simplified HOSPITAL score of ≥4 who are discharged home or nursing home after a hospital stay of one day or more in the department of medicine of the four participating hospitals. The patients randomized to the intervention group will receive a pre-discharge intervention by a study nurse with patient education, medication reconciliation, and follow-up appointment with their referring physician. They will receive short follow-up phone calls at 3 and 14 days after discharge to ensure medication adherence and follow-up by the ambulatory care physician. A blind study nurse will collect outcomes at 1 month by phone call interview. The control group will receive usual care. The TARGET-READ study aims to increase the knowledge about the efficacy of a bundled intervention aimed at reducing 30-day hospital readmission or death in higher-risk medical patients.

Comparison of 30-day planned and unplanned readmissions in a tertiary teaching hospital in China

PURPOSE

The purpose of this study was to analyze and compare the clinical characteristics of patients with 30-day planned and unplanned readmissions and to identify patients at high risk for unplanned readmissions. This will facilitate a better understanding of these readmissions and improve and optimize resource utilization for this patient population.

METHODS

A retrospective cohort descriptive study was conducted at the West China Hospital (WCH), Sichuan University from January 1, 2015, to December 31, 2020. Discharged patients (≥ 18 years old) were divided into unplanned readmission and planned readmission groups according to 30-day readmission status. Demographic and related information was collected for each patient. Logistic regression analysis was used to assess the association between unplanned patient characteristics and the risk of readmission.

RESULTS

We identified 1,118,437 patients from 1,242,496 discharged patients, including 74,494 (6.7%) 30-day planned readmissions and 9,895 (0.9%) unplanned readmissions. The most common diseases of planned readmissions were antineoplastic chemotherapy (62,756/177,749; 35.3%), radiotherapy sessions for malignancy (919/8,229; 11.2%), and systemic lupus erythematosus (607/4,620; 13.1%). The most common diseases of unplanned readmissions were antineoplastic chemotherapy (2038/177,747; 1.1%), age-related cataract (1061/21,255; 5.0%), and unspecified disorder of refraction (544/5,134; 10.6%). There were statistically significant differences between planned and unplanned readmissions in terms of patient sex, marital status, age, length of initial stay, the time between discharge, ICU stay, surgery, and health insurance.

CONCLUSION

Accurate information on 30-day planned and unplanned readmissions facilitates effective planning of healthcare resource allocation. Identifying risk factors for 30-day unplanned readmissions can help develop interventions to reduce readmission rates.

[Predictors for Early Unplanned Readmissions]

Predictors for Early Unplanned Readmissions Abstract. Unplanned rehospitalizations represent a major burden for patients, their relatives and the healthcare system. Since the introduction of the SwissDRG in 2012, financial incentives for hospitals have been promoted to forestall readmissions. Not every patient is at risk for rehospitalization. Affected patients can be identified by predictors from various areas in order to implement adequate interventions and avoid readmissions. Predictors can be directly related to patients as in the case of polypharmacy, multiple comorbidities or related to gender, but also provider-related and system-related. Early follow-up visits or a pre-discharge medication review are cited as effective interventions.

Evaluation of an Automated Text Message-Based Program to Reduce Use of Acute Health Care Resources After Hospital Discharge

IMPORTANCE

Posthospital contact with a primary care team is an established pillar of safe transitions. The prevailing model of telephone outreach is usually limited in scope and operationally burdensome.

OBJECTIVE

To determine whether a 30-day automated texting program to support primary care patients after hospital discharge is associated with reductions in the use of acute care resources.

DESIGN, SETTING, AND PARTICIPANTS

This cohort study used a difference-in-differences approach at 2 academic primary care practices in Philadelphia from January 27 through August 27, 2021. Established patients of the study practices who were 18 years or older, were discharged from an acute care hospitalization, and received the usual transitional care management telephone call were eligible for the study. At the intervention practice, 604 discharges were eligible and 430 (374 patients, of whom 46 had >1 discharge) were enrolled in the intervention. At the control practice, 953 patients met eligibility criteria. The study period, including before and after the intervention, ran from August 27, 2020, through August 27, 2021.

EXPOSURE

Patients received automated check-in text messages from their primary care practice on a tapering schedule during the 30 days after discharge. Any needs identified by the automated messaging platform were escalated to practice staff for follow-up via an electronic medical record inbox.

MAIN OUTCOMES AND MEASURES

The primary study outcome was any emergency department (ED) visit or readmission within 30 days of discharge. Secondary outcomes included any ED visit or any readmission within 30 days, analyzed separately, and 30- and 60-day mortality. Analyses were based on intention to treat.

RESULTS

A total of 1885 patients (mean [SD] age, 63.2 [17.3] years; 1101 women [58.4%]) representing 2617 discharges (447 before and 604 after the intervention at the intervention practice; 613 before and 953 after the intervention at the control practice) were included in the analysis. The adjusted odds ratio (aOR) for any use of acute care resources after implementation of the intervention was 0.59 (95% CI, 0.38-0.92). The aOR for an ED visit was 0.77 (95% CI, 0.45-1.30) and for a readmission was 0.45 (95% CI, 0.23-0.86). The aORs for death within 30 and 60 days of discharge at the intervention practice were 0.92 (95% CI, 0.23-3.61) and 0.63 (95% CI, 0.21-1.85), respectively.

CONCLUSIONS AND RELEVANCE

The findings of this cohort study suggest that an automated texting program to support primary care patients after hospital discharge was associated with significant reductions in use of acute care resources. This patient-centered approach may serve as a model for improving postdischarge care.

Re-Admission of COVID-19 Patients Hospitalized with Omicron Variant-A Retrospective Cohort Study

In accordance with previous publications, re-admission rates following hospitalization of patients with COVID-19 is 10%. The aim of the current study was to describe the rates and risk factors of hospital re-admissions two months following discharge from hospitalization during the fifth wave due to the dominant Omicron variant. A retrospective cohort study was performed in Rabin Medical Center, Israel, from November 2021 to February 2022. The primary outcome was re-admissions with any diagnosis; the secondary outcome was mortality within two months of discharge. Overall, 660 patients were hospitalized with a diagnosis of COVID-19. Of the 528 patients discharged from a primary hospitalization, 150 (28%) were re-admitted. A total of 164 patients (25%) died throughout the follow-up period. A multi-variable analysis determined that elevated creatinine was associated with a higher risk of re-admissions. Rates of re-admissions after discharge during the Omicron wave were considerably higher compared to previous waves. A discharge plan for surveillance and treatment following hospitalization is of great importance in the management of pandemics.

Predictive Modeling of Hospital Readmission: Challenges and Solutions

Hospital readmission prediction is a study to learn models from historical medical data to predict probability of a patient returning to hospital in a certain period, e.g. 30 or 90 days, after the discharge. The motivation is to help health providers deliver better treatment and post-discharge strategies, lower the hospital readmission rate, and eventually reduce the medical costs. Due to inherent complexity of diseases and healthcare ecosystems, modeling hospital readmission is facing many challenges. By now, a variety of methods have been developed, but existing literature fails to deliver a complete picture to answer some fundamental questions, such as what are the main challenges and solutions in modeling hospital readmission; what are typical features/models used for readmission prediction; how to achieve meaningful and transparent predictions for decision making; and what are possible conflicts when deploying predictive approaches for real-world usages. In this paper, we systematically review computational models for hospital readmission prediction, and propose a taxonomy of challenges featuring four main categories: (1) data variety and complexity; (2) data imbalance, locality and privacy; (3) model interpretability; and (4) model implementation. The review summarizes methods in each category, and highlights technical solutions proposed to address the challenges. In addition, a review of datasets and resources available for hospital readmission modeling also provides firsthand materials to support researchers and practitioners to design new approaches for effective and efficient hospital readmission prediction.

The contributions of survey-based versus administrative measures of socioeconomic status in predicting type of post-acute care for hospitalized Medicare beneficiaries

OBJECTIVES

To assess and compare the associations between socioeconomic status (SES) measures from two sources (claims vs. survey data) and the type of post-acute care (PAC) locations following hospital discharge.

METHODS

This observational study included Medicare Fee-for-Service (FFS) beneficiaries age 65.5 years or older who participated in the Medicare Current Beneficiary Survey (MCBS) and were hospitalized in 2006-2011. Multiple data sets were used including: Area Deprivation Index; Medicare Cost Reports, Provider of Services files, and Area Health Resource File. Multinomial regression models estimated associations between beneficiary's SES and PAC type. SES measures came from surveys (income and education) and administrative records (dual enrollment and area deprivation). PAC types included home with self-care, home health agency, skilled nursing facility (SNF), or inpatient rehabilitation facility.

RESULTS

Low income and dual enrollment were associated with higher SNF use while living in a deprived area was associated with lower SNF use and higher use of home with self-care. Dual enrollment and area deprivation were associated with the largest differences.

CONCLUSIONS

If policies to modify payment based on SES are considered, administrative measures (dual enrollment and area deprivation) rather than survey measures (education and income) may be sufficient.

A Machine Learning Approach for Predicting Readmission Charges Billed by Hospitals

Background

The Centers for Medicare and Medicaid Services projects that health care costs will continue to grow over the next few years. Rising readmission costs contribute significantly to increasing health care costs. Multiple areas of health care, including readmissions, have benefited from the application of various machine learning algorithms in several ways.

Objective

We aimed to identify suitable models for predicting readmission charges billed by hospitals. Our literature review revealed that this application of machine learning is underexplored. We used various predictive methods, ranging from glass-box models (such as regularization techniques) to black-box models (such as deep learning–based models).

Methods

We defined readmissions as readmission with the same major diagnostic category (RSDC) and all-cause readmission category (RADC). For these readmission categories, 576,701 and 1,091,580 individuals, respectively, were identified from the Nationwide Readmission Database of the Healthcare Cost and Utilization Project by the Agency for Healthcare Research and Quality for 2013. Linear regression, lasso regression, elastic net, ridge regression, eXtreme gradient boosting (XGBoost), and a deep learning model based on multilayer perceptron (MLP) were the 6 machine learning algorithms we tested for RSDC and RADC through 10-fold cross-validation.

Results

Our preliminary analysis using a data-driven approach revealed that within RADC, the subsequent readmission charge billed per patient was higher than the previous charge for 541,090 individuals, and this number was 319,233 for RSDC. The top 3 major diagnostic categories (MDCs) for such instances were the same for RADC and RSDC. The average readmission charge billed was higher than the previous charge for 21 of the MDCs in the case of RSDC, whereas it was only for 13 of the MDCs in RADC. We recommend XGBoost and the deep learning model based on MLP for predicting readmission charges. The following performance metrics were obtained for XGBoost: (1) RADC (mean absolute percentage error [MAPE]=3.121%; root mean squared error [RMSE]=0.414; mean absolute error [MAE]=0.317; root relative squared error [RRSE]=0.410; relative absolute error [RAE]=0.399; normalized RMSE [NRMSE]=0.040; mean absolute deviation [MAD]=0.031) and (2) RSDC (MAPE=3.171%; RMSE=0.421; MAE=0.321; RRSE=0.407; RAE=0.393; NRMSE=0.041; MAD=0.031). The performance obtained for MLP-based deep neural networks are as follows: (1) RADC (MAPE=3.103%; RMSE=0.413; MAE=0.316; RRSE=0.410; RAE=0.397; NRMSE=0.040; MAD=0.031) and (2) RSDC (MAPE=3.202%; RMSE=0.427; MAE=0.326; RRSE=0.413; RAE=0.399; NRMSE=0.041; MAD=0.032). Repeated measures ANOVA revealed that the mean RMSE differed significantly across models with P<.001. Post hoc tests using the Bonferroni correction method indicated that the mean RMSE of the deep learning/XGBoost models was statistically significantly (P<.001) lower than that of all other models, namely linear regression/elastic net/lasso/ridge regression.

Conclusions

Models built using XGBoost and MLP are suitable for predicting readmission charges billed by hospitals. The MDCs allow models to accurately predict hospital readmission charges.

Follow-up Psychiatric Care and Risk of Readmission in Patients with Serious Mental Illness in State Funded or Operated Facilities

Receipt of outpatient treatment within 30 days of discharge from psychiatric hospitalization is an established quality indicator; however, there is scant, mixed evidence as to whether it reduces the risk of readmission. We evaluated this question in patients hospitalized for schizophrenic, bipolar or depressive disorders using the Mental Health Treatment Episode Data Set (MH-TEDS), comprising patients in state-funded or -operated facilities and programs. We performed a 6-month, retrospective longitudinal cohort study including 44,761 patients with schizophrenic disorders, 45,413 patients with bipolar disorders, and 74,995 patients with depressive disorders with an index hospitalization between 2014 and 2018, stratified by whether they had at least one outpatient treatment admission in the first 30 days post-discharge. We used multivariable logistic regression to assess risk of readmission during days 31-180. We found that less than 10 percent of patients in the three cohorts received the recommended follow-up outpatient care. Furthermore, we found that schizophrenic and bipolar patients who did receive such care were no less likely to be readmitted than those not receiving such care (AOR = 0.96, 95% CI 0.87-1.06; AOR 1.06, 955 CI 0.98-1.14), and patients with depressive disorders receiving such care were more likely to be readmitted (AOR = 1.14, 95% CI 1.07-1.22). Thus, few patients received follow-up outpatient care within 30 days of discharge. When it occurred, such outpatient care was either not linked to reduced readmissions or was associated with increased readmissions. These findings suggest the need for more effective care processes in state-funded or -operated facilities.

Safety of Elderly Fallers: Identifying Associated Risk Factors for 30-Day Unplanned Readmissions Using a Clinical Data Warehouse

BACKGROUND

Hospital readmissions are a major problem in the older people as they are frequent, costly, and life-threatening. Falls among older adults are the leading cause of injury, deaths, and emergency department visits for trauma.

OBJECTIVE

The main objective was to determine risk factors associated with a 30-day readmission after index hospital admission for fall-related injuries.

METHODS

A retrospective nested case-control study was conducted. Data from elderly patients initially hospitalized for fall-related injuries in 2019, in 11 of the Greater Paris University Hospitals and discharged home, were retrieved from the clinical data warehouse. Cases were admission of elderly patients who subsequently experienced a readmission within 30 days after discharge from the index admission. Controls were admission of elderly patients who were not readmitted to hospital.

RESULTS

Among 670 eligible index admissions, 127 (18.9%) were followed by readmission within 30 days after discharge. After multivariate analysis, men sex (odds ratio [OR] = 2.29, 95% confidence interval [CI] = 1.45-3.61), abnormal concentration of C-reactive protein, and anemia (OR = 2.22, 95% CI = 1.28-3.85; OR = 1.85, 95% CI = 1.11-3.11, respectively) were associated with a higher risk of readmission. Oppositely, having a traumatic injury at index admission decreased this risk (OR = 0.47, 95% CI = 0.28-0.81).

CONCLUSIONS

Reducing early unplanned readmission is crucial, especially in elderly patients susceptible to falls. Our results indicate that the probability of unplanned readmission is higher for patients with specific characteristics that should be taken into consideration in interventions designed to reduce this burden.

A Score to Predict the Risk of Major Adverse Drug Reactions Among Multi-Drug Resistant Tuberculosis Patients in Southern Ethiopia, 2014–2019

Background

Adverse events (AE) contribute to poor drug adherence and withdrawal, which contribute to a low treatment success rate. AE are commonly reported among multi-drug resistance tuberculosis (MDR-TB) patients in Ethiopia. However, predictors of AE among MDR-TB patients were limited in Ethiopia. Thus, the current study aimed to develop and validate a score to predict the risks of major AE among MDR-TB patients in Southern Ethiopia.

Methods

A retrospective follow-up study design was employed among MDR-TB patients from 2014–2019 in southern Ethiopia at selected hospitals. A least absolute shrinkage and selection operator algorithm was used to select the most potent predictors of the outcome. The adverse event risk score was built based on the multivariable logistic regression analysis. Discriminatory power and calibration were checked to evaluate the performance of the model. Bootstrapping method with 100 repetitions was used for internal model validation.

Results

History of baseline khat use, long-term drug regimen use, and having coexisting disorders (co-morbidity) were predictors of AEs. The score has a satisfactory discriminatory power (AUC = 0.77, 95% CI: 0.68, 0.82) and a modest calibration (Prob > chi² = 0.2043). It was found to have the same c-statistics after validation by bootstrapping method of 100 repetitions with replacement.

Conclusion

A history of baseline khat use, co-morbidity, and long-term drug regimen use are helpful to predict individual risk of major adverse events in MDR-TB patients with a satisfactory degree of accuracy (AUC = 0.77).

Factors Related to Pediatric Readmissions of Four Major Diagnostic Categories in Hawai`i

Readmissions are a key quality measure for health care decision making and understanding variables associated with readmissions has become a crucial research area. This study identified patient-level factors that might be associated with pediatric readmissions using a database that included inpatient data from 2008 to 2017 from Hawai`i. Four major diagnostic categories with the most pediatric readmissions in the state were identified: respiratory, digestive, mental, and nervous system diseases and disorders. The associations between readmission and patient-level variables, such as age, sex, race/ethnicity, insurance status, and Charlson Comorbidity Index (CCI), were determined for each diagnosis and for overall readmissions. CCI and insurance were the strongest predictors when all diagnoses were combined. However, for some diagnoses, there was weak or no association between CCI, insurance, and readmission. This suggests that diagnosis-specific analysis of predictors of readmission may be more useful than looking at predictors of readmission for all diagnoses combined. While this study focused on patient variables, future studies should also incorporate how hospital variables may also be related to diagnosis.

Epidemiology of haemodialysis outcomes

Haemodialysis (HD) is the commonest form of kidney replacement therapy in the world, accounting for approximately 69% of all kidney replacement therapy and 89% of all dialysis. Over the last six decades since the inception of HD, dialysis technology and patient access to the therapy have advanced considerably, particularly in high-income countries. However, HD availability, accessibility, cost and outcomes vary widely across the world and, overall, the rates of impaired quality of life, morbidity and mortality are high. Cardiovascular disease affects more than two-thirds of people receiving HD, is the major cause of morbidity and accounts for almost 50% of mortality. In addition, patients on HD have high symptom loads and are often under considerable financial strain. Despite the many advances in HD technology and delivery systems that have been achieved since the treatment was first developed, poor outcomes among patients receiving HD remain a major public health concern. Understanding the epidemiology of HD outcomes, why they might vary across different populations and how they might be improved is therefore crucial, although this goal is hampered by the considerable heterogeneity in the monitoring and reporting of these outcomes across settings.

This Review examines the epidemiology of haemodialysis outcomes — clinical, patient-reported and surrogate outcomes — across world regions and populations, including vulnerable individuals. The authors also discuss the current status of monitoring and reporting of haemodialysis outcomes and potential strategies for improvement.

Nearly 4 million people in the world are living on kidney replacement therapy (KRT), and haemodialysis (HD) remains the commonest form of KRT, accounting for approximately 69% of all KRT and 89% of all dialysis.

Dialysis technology and patient access to KRT have advanced substantially since the 1960s, particularly in high-income countries. However, HD availability, accessibility, cost and outcomes continue to vary widely across countries, particularly among disadvantaged populations (including Indigenous peoples, women and people at the extremes of age).

Cardiovascular disease affects over two-thirds of people receiving HD, is the major cause of morbidity and accounts for almost 50% of mortality; mortality among patients on HD is significantly higher than that of their counterparts in the general population, and treated kidney failure has a higher mortality than many types of cancer. Patients on HD also experience high burdens of symptoms, poor quality of life and financial difficulties.

Careful monitoring of the outcomes of patients on HD is essential to develop effective strategies for risk reduction. Outcome measures are highly variable across regions, countries, centres and segments of the population. Establishing kidney registries that collect a variety of clinical and patient-reported outcomes using harmonized definitions is therefore crucial.

Evaluation of HD outcomes should include the impact on family and friends, and personal finances, and should examine inequities in disadvantaged populations, who comprise a large proportion of the HD population.

Coronavirus Disease 2019 and Hospital Readmissions: Patient Characteristics and Socioeconomic Factors Associated With Readmissions in an Urban Safety-Net Hospital System

BACKGROUND

It is not yet known whether socioeconomic factors (ie, social determinants of health) are associated with readmission following hospitalization for coronavirus disease 2019 (COVID-19).

METHODS

We conducted a retrospective cohort study of 6191 adult patients hospitalized with COVID-19 in a large New York City safety-net hospital system between March 1 and June 1, 2020. Associations between 30-day readmission and selected demographic characteristics, socioeconomic factors, prior health care utilization, and relevant features of the index hospitalization were analyzed using a multivariable generalized estimating equation model.

RESULTS

The readmission rate was 7.3%, with a median of 7 days between discharge and readmission. The following were risk factors for readmission: age 65 and older [adjusted odds ratio (aOR): 1.32; 95% confidence interval (CI): 1.13-1.55], history of homelessness, (aOR: 2.03 95% CI: 1.49-2.77), baseline coronary artery disease (aOR: 1.68; 95% CI: 1.34-2.10), congestive heart failure (aOR: 1.34; 95% CI: 1.20-1.49), cancer (aOR: 1.68; 95% CI: 1.26-2.24), chronic kidney disease (aOR: 1.74; 95% CI: 1.46-2.07). Patients' sex, race/ethnicity, insurance, and presence of obesity were not associated with increased odds of readmission. A longer length of stay (aOR: 0.98; 95% CI: 0.97-1.00) and use of noninvasive supplemental oxygen (aOR: 0.68; 95% CI: 0.56-0.83) was associated with lower odds of readmission. Upon readmission, 18.4% of patients required intensive care, and 13.7% expired.

CONCLUSION

We have found some factors associated with increased odds of readmission among patients hospitalized with COVID-19. Awareness of these risk factors, including patients' social determinants of health, may ultimately help to reduce readmission rates.

Machine learning for predicting readmission risk among the frail: Explainable AI for healthcare

Healthcare costs due to unplanned readmissions are high and negatively affect health and wellness of patients. Hospital readmission is an undesirable outcome for elderly patients. Here, we present readmission risk prediction using five machine learning approaches for predicting 30-day unplanned readmission for elderly patients (age ≥ 50 years). We use a comprehensive and curated set of variables that include frailty, comorbidities, high-risk medications, demographics, hospital, and insurance utilization to build these models. We conduct a large-scale study with electronic health record (her) data with over 145,000 observations from 76,000 patients. Findings indicate that the category boost (CatBoost) model outperforms other models with a mean area under the curve (AUC) of 0.79. We find that prior readmissions, discharge to a rehabilitation facility, length of stay, comorbidities, and frailty indicators were all strong predictors of 30-day readmission. We present in-depth insights using Shapley additive explanations (SHAP), the state of the art in machine learning explainability.

Association of the VACS Index With Hospitalization Among People With HIV in the NA-ACCORD

BACKGROUND

People with HIV (PWH) have a higher hospitalization rate than the general population. The Veterans Aging Cohort Study (VACS) Index at study entry well predicts hospitalization in PWH, but it is unknown if the time-updated parameter improves hospitalization prediction. We assessed the association of parameterizations of the VACS Index 2.0 with the 5-year risk of hospitalization.

SETTING

PWH ≥30 years old with at least 12 months of antiretroviral therapy (ART) use and contributing hospitalization data from 2000 to 2016 in North American AIDS Cohort Collaboration on Research and Design (NA-ACCORD) were included. Three parameterizations of the VACS Index 2.0 were assessed and categorized by quartile: (1) "baseline" measurement at study entry; (2) time-updated measurements; and (3) cumulative scores calculated using the trapezoidal rule.

METHODS

Discrete-time proportional hazard models estimated the crude and adjusted associations (and 95% confidence intervals [CIs]) of the VACS Index parameterizations and all-cause hospitalizations. The Akaike information criterion (AIC) assessed the model fit with each of the VACS Index parameters.

RESULTS

Among 7289 patients, 1537 were hospitalized. Time-updated VACS Index fitted hospitalization best with a more distinct dose-response relationship [score <43: reference; score 43-55: aHR = 1.93 (95% CI: 1.66 to 2.23); score 55-68: aHR = 3.63 (95% CI: 3.12 to 4.23); score ≥68: aHR = 9.98 (95% CI: 8.52 to 11.69)] than study entry and cumulative VACS Index after adjusting for known risk factors.

CONCLUSIONS

Time-updated VACS Index 2.0 had the strongest association with hospitalization and best fit to the data. Health care providers should consider using it when assessing hospitalization risk among PWH.

Miller N, L. Pecina J, D. Decker M, I. Wi C, Garrison GM. Association of Early and Late Hospital Readmissions with a Novel Housing-Based Socioeconomic Measure

Background

While socioeconomic status has been linked to hospital readmissions for several conditions, reliable measures of individual socioeconomic status are often not available. HOUSES, a new measure of individual socioeconomic status based upon objective public data about one's housing unit, is inversely associated with overall hospitalization rate but it has not been studied with respect to readmissions.

Purpose

To determine if patients in the lowest HOUSES quartile are more likely to be readmitted within 30 days (short-term) and 180 days (long-term).

Methods

A retrospective cohort study of 11 993 patients having 21 633 admissions was conducted using generalized linear mixed-effects models.

Results

HOUSES quartile did not show any significant association with early readmission. However, when compared to the lowest HOUSES quartile, the second quartile (OR = 0.90, 95%CI 0.83-0.98) and the third quartile (OR = 0.91, 95%CI 0.83-0.99) were associated with lower odds of late readmission while the highest quartile (OR = 0.91, 95%CI 0.82-1.01) was not statistically different.

Conclusion

HOUSES was associated with late readmission, but not early readmission. This may be because early readmissions are influenced by medical conditions and hospital care while late readmissions are influenced by ambulatory care and home-based factors. Since HOUSES relies on public county assessor data, it is generally available and may be used to focus interventions on those at highest risk for late readmission.

Hospital readmissions in Internal Medicine Specialty: Frequency, associated factors and outcomes

Objectives:

Hospital readmission has become a focus of national attention as a potential indicator of healthcare quality and has a significant financial impact on healthcare system. Limited data is available regarding readmissions to Internal Medicine specialty from our sub-continent. It is, therefore, essential to determine the frequency and factors leading to readmissions, in order to avoid preventable readmissions and improve quality of healthcare provision.

Methods:

This retrospective study reviewed adult discharges from Internal Medicine specialty between October 2018 and February 2019 at Aga Khan University Hospital. Out of 1,835 discharges, 491 were randomly selected after excluding expiries. The frequency, factors and outcomes of readmission were noted. The studied outcomes included length of stay and in-hospital mortality.

Results:

Out of 491 patients, 15.3% were readmitted within 30-days of their discharge. Most of the readmitted patients were females (56%) and elderly with a mean age 67.1±14.9 years. Majority of the patient who got readmitted had multi-morbidities (68%) and were of functional Class-II (39%).The mean length of stay for index and readmission was between 4-7days. Eighty-percent readmissions were discharged as planned, 13% on request and seven-percent left against medical advice in their index admission. The most common causes of readmission were persistence of symptoms (43%) and nosocomial infection (29%). Avoidable causes included hospital-associated pneumonia, urinary tract infections and septic shock. Mortality in readmitted patients was 12%.

Conclusions:

The causes of readmission is multi-factorial, including advanced age, multi-morbidities, persistence of symptoms and nosocomial infections. Early follow-ups should be advised to prevent avoidable readmissions.

A Comparison of Falls and Dizziness Handicap by Vestibular Diagnosis

PURPOSE

There is a paucity of data that directly compares the falls rate and dizziness handicap of different vestibular diagnoses. The purpose of this study is to compare the falls rate and dizziness handicap of common vestibular diagnoses encountered among a cohort of vestibular patients at a single institution.

METHOD

We conducted a retrospective cross-sectional study of patients evaluated for dizziness at a tertiary care center vestibular clinic between August 1, 2017, and March 19, 2019. Vestibular diagnosis, demographic variables, comorbidities, falls status, and Dizziness Handicap Inventory (DHI) were extracted from the medical record for analysis. Associations between vestibular diagnosis and falls history or DHI were evaluated using multivariate logistic and linear regression, respectively.

RESULTS

A total of 283 patients met our inclusion criteria with the following diagnoses: benign paroxysmal positional vertigo (BPPV; n = 55), acoustic neuroma (n = 30), Ménière's disease (n = 28), multiple vestibular diagnoses (n = 15), vestibular migraine (n = 135), or vestibular neuritis (n = 20). After adjusting for age, sex, race, medications, and comorbidities, the odds of falling was 2.47 times greater (95% CI [1.08, 6.06], p = .039) and the DHI score was 11.66 points higher (95% CI [4.99, 18.33], p < .001) in those with vestibular migraine compared to those with BPPV. Other diagnoses were comparable to BPPV with respect to odds of falling and dizziness handicap.

CONCLUSIONS

Patients with vestibular migraine may suffer an increased risk of falls and dizziness handicap compared to patients with BPPV. Our findings highlight the need for timely evaluation and treatment of all patients with vestibular disease.

Extended Patient Alone Time in Emergency Department Leads to Increased Risk of 30-Day Hospitalization

OBJECTIVES

This study was conducted to describe patients at risk for prolonged time alone in the emergency department (ED) and to determine the relationship between clinical outcomes, specifically 30-day hospitalization, and patient alone time (PAT) in the ED.

METHODS

An observational cohort design was used to evaluate PAT and patient characteristics in the ED. The study was conducted in a tertiary academic ED that has both adult and pediatric ED facilities and of patients placed in an acute care room for treatment between May 1 and July 31, 2016, excluding behavioral health patients. Simple linear regression and t tests were used to evaluate the relationship between patient characteristics and PAT. Logistic regression was used to evaluate the relationship between 30-day hospitalization and PAT.

RESULTS

Pediatric patients had the shortest total PAT compared with all older age groups (86.4 minutes versus 131 minutes, P < 0.001). Relationships were seen between PAT and patient characteristics, including age, geographic region, and the severity and complexity of the health condition. Controlling for Charlson comorbidity index and other potentially confounding variables, a logistic regression model showed that patients are more likely to be hospitalized within 30 days after their ED visit, with an odds ratio (95% confidence interval) of 1.056 (1.017-1.097) for each additional hour of PAT.

CONCLUSIONS

Patient alone time is not equal among all patient groups. Study results indicate that PAT is significantly associated with 30-day hospitalization. This conclusion indicates that PAT may affect patient outcomes and warrants further investigation.

The critical role of care coordinators for persons with substance use disorder in rural settings: a case study

Many rural regions lack the basic fundamentals in healthcare for Opioid Use Disorder. We present a case of a dual-diagnosed, impoverished, adult female court-ordered to inpatient treatment in rural Kentucky. A care coordinator linked her to regional and community resources to address her health, environmental, and psychosocial needs, as well as provided needed transportation, coaching, and emotional support. As a result, she overcame the substantial barriers that each component of the care continuum presents in severely underserved areas. This case study highlights the critical role care coordination plays in reentry, its differences from urban areas, and its alignment with social work's core values.

Factors and experiences associated with unscheduled 30-day hospital readmission: A mixed method study

INTRODUCTION

Analysis of risk factors can pave the way for reducing unscheduled hospital readmissions and improve resource utilisation.

METHODS

This was a concurrent nested, mixed method study. Factors associated with patients readmitted within 30 days between 2011 and 2015 at the National University Hospital, Singapore (N=104,496) were examined. Fifty patients were sampled in 2016 to inform an embedded qualitative study. Narrative interviews explored the periods of readmissions and related experiences, contrasted against those of non-readmitted patients.

RESULTS

Neoplastic disease (odds ratio [OR] 1.91, 95% confidence interval [CI] 1.70-2.15), number of discharged medications (5 to 10 medications OR 1.21, 95% CI 1.14-1.29; ≥11 medications OR 1.80, 95% CI 1.66-1.95) and length of stay >7 days (OR 1.46, 95% CI 1.36-1.58) were most significantly associated with readmissions. Other factors including number of surgical operations, subvention class, number of emergency department visits in the previous year, hospital bill size, gender, age, Charlson comorbidity index and ethnicity were also independently associated with hospital readmissions. Although readmitted and non-readmitted patients shared some common experiences, they reported different psychological reactions to their illnesses and viewed hospital care differently. Negative emotions, feeling of being left out by the healthcare team and perception of ineffective or inappropriate treatment were expressed by readmitted patients.

CONCLUSION

Patient, hospital and system-related factors were associated with readmissions, which may allow early identification of at-risk patients. Qualitative analysis suggested several areas of improvement in care including greater empowerment and involvement of patients in care and decision making.

Development of Prediction Models for Unplanned Hospital Readmission within 30 Days Based on Common Data Model: A Feasibility Study

Background  Unplanned hospital readmission after discharge reflects low satisfaction and reliability in care and the possibility of potential medical accidents, and is thus indicative of the quality of patient care and the appropriateness of discharge plans.

Objectives  The purpose of this study was to develop and validate prediction models for all-cause unplanned hospital readmissions within 30 days of discharge, based on a common data model (CDM), which can be applied to multiple institutions for efficient readmission management.

Methods  Retrospective patient-level prediction models were developed based on clinical data of two tertiary general university hospitals converted into a CDM developed by Observational Medical Outcomes Partnership. Machine learning classification models based on the LASSO logistic regression model, decision tree, AdaBoost, random forest, and gradient boosting machine (GBM) were developed and tested by manipulating a set of CDM variables. An internal 10-fold cross-validation was performed on the target data of the model. To examine its transportability, the model was externally validated. Verification indicators helped evaluate the model performance based on the values of area under the curve (AUC).

Results  Based on the time interval for outcome prediction, it was confirmed that the prediction model targeting the variables obtained within 30 days of discharge was the most efficient (AUC of 82.75). The external validation showed that the model is transferable, with the combination of various clinical covariates. Above all, the prediction model based on the GBM showed the highest AUC performance of 84.14 ± 0.015 for the Seoul National University Hospital cohort, yielding in 78.33 in external validation.

Conclusions  This study showed that readmission prediction models developed using machine-learning techniques and CDM can be a useful tool to compare two hospitals in terms of patient-data features.

Targeted Clinical Interventions for Reducing Pediatric Readmissions

BACKGROUND

In this interventional study, we addressed the selection and application of clinical interventions on pediatric patients identified as at risk by a predictive model for readmissions.

METHODS

A predictive model for readmissions was implemented, and a team of providers expanded corresponding clinical interventions for at-risk patients at a freestanding children's hospital. Interventions encompassed social determinants of health, outpatient care, medication reconciliation, inpatient and discharge planning, and postdischarge calls and/or follow-up. Statistical process control charts were used to compare readmission rates for the 3-year period preceding adoption of the model and clinical interventions with those for the 2-year period after adoption of the model and clinical interventions. Potential financial savings were estimated by using national estimates of the cost of pediatric inpatient readmissions.

RESULTS

The 30-day all-cause readmission rates during the periods before and after predictive modeling (and corresponding 95% confidence intervals [CI]) were 12.5% (95% CI: 12.2%-12.8%) and 11.1% (95% CI: 10.8%-11.5%), respectively. More modest but similar improvements were observed for 7-day readmissions. Statistical process control charts indicated nonrandom reductions in readmissions after predictive model adoption. The national estimate of the cost of pediatric readmissions indicates an associated health care savings due to reduced 30-day readmission during the 2-year predictive modeling period at $2 673 264 (95% CI: $2 612 431-$2 735 364).

CONCLUSIONS

A combination of predictive modeling and targeted clinical interventions to improve the management of pediatric patients at high risk for readmission was successful in reducing the rate of readmission and reducing overall health care costs. The continued prioritization of patients with potentially modifiable outcomes is key to improving patient outcomes.

Predicting 30-Day Readmissions: Evidence From a Small Rural Psychiatric Hospital

To improve quality of care and patient outcomes, and to reduce costs, hospitals in the United States are trying to mitigate readmissions that are potentially avoidable. By identifying high-risk patients, hospitals may be able to proactively adapt treatment and discharge planning to reduce the likelihood of readmission. Our objective in this study was to derive and validate a predictive model of 30-day readmissions for a small rural psychiatric hospital in the northeast. However, this model can be adapted by other rural psychiatric hospitals-a context that has been understudied in the literature. Our sample consisted of 1912 adult inpatients (1281 in the derivation cohort and 631 in the validation cohort), who were admitted between August 1, 2014, and July 31, 2016. We used deidentified data from the hospital's electronic medical record, including physician orders and discharge summaries. These data were merged with community-level variables that reflected the availability of care in the patients' zip codes. We first considered the correlates of 30-day readmission in a regression framework. We found that the probability of readmission increased with the number of previous admissions (vs. no readmissions). Moreover, the probability of readmission was much higher for patients with a depressive disorder (vs. no depressive disorder), with another mood disorder (vs. no other mood disorder), and/or with a psychotic disorder (vs. no psychotic disorder). We used these associations to derive a predictive model, in which we used the regression coefficients to construct a score for each patient. We then estimated the predicted probability of 30-day readmission on the basis of that score. After validating the model, we discuss the implications for clinical practice and the limitations of our approach.

Predictive modelling of hospital readmission: Evaluation of different preprocessing techniques on machine learning classifiers

Hospital readmission is a major cost for healthcare systems worldwide. If patients with a higher potential of readmission could be identified at the start, existing resources could be used more efficiently, and appropriate plans could be implemented to reduce the risk of readmission. Therefore, it is important to predict the right target patients. Medical data is usually noisy, incomplete, and inconsistent. Hence, before developing a prediction model, it is crucial to efficiently set up the predictive model so that improved predictive performance is achieved. The current study aims to analyse the impact of different preprocessing methods on the performance of different machine learning classifiers. The preprocessing applied by previous hospital readmission studies were compared, and the most common approaches highlighted such as missing value imputation, feature selection, data balancing, and feature scaling. The hyperparameters were selected using Bayesian optimisation. The different preprocessing pipelines were assessed using various performance metrics and computational costs. The results indicated that the preprocessing approaches helped improve the model’s prediction of hospital readmission.

A prospective cohort study of two predictor models for 30-day emergency readmission in older patients

AIM

To undertake a prospective study of the accuracy of two models (LACE and BOOST) in predicting unplanned hospital readmission in older patients (>75 years).

METHODS

Data were collected from a single centre prospectively on 110 patients over 75 years old admitted to the acute medical unit. Follow-up was conducted at 30 days. The primary outcome was the c-statistic for both models.

RESULTS

The readmission rate was 32.7% and median age 82 years, and both BOOST and LACE scores were significantly higher in those readmitted compared with those who were not. C-statistics were calculated for both tools with BOOST score 0.667 (95% CI 0.559-0.775, P = .005) and LACE index 0.685 (95% CI 0.579-0.792, P = .002).

CONCLUSION

In this prospective study, both the BOOST and LACE scores were found to be significant yet poor, predictive models of hospital readmission. Recent hospitalisation (within the previous 6 months) was found to be the most significant contributing factor.

Published models that predict hospital readmission: a critical appraisal

INTRODUCTION

The number of readmission risk prediction models available has increased rapidly, and these models are used extensively for health decision-making. Unfortunately, readmission models can be subject to flaws in their development and validation, as well as limitations in their clinical usefulness.

OBJECTIVE

To critically appraise readmission models in the published literature using Delphi-based recommendations for their development and validation.

METHODS

We used the modified Delphi process to create Critical Appraisal of Models that Predict Readmission (CAMPR), which lists expert recommendations focused on development and validation of readmission models. Guided by CAMPR, two researchers independently appraised published readmission models in two recent systematic reviews and concurrently extracted data to generate reference lists of eligibility criteria and risk factors.

RESULTS

We found that published models (n=81) followed 6.8 recommendations (45%) on average. Many models had weaknesses in their development, including failure to internally validate (12%), failure to account for readmission at other institutions (93%), failure to account for missing data (68%), failure to discuss data preprocessing (67%) and failure to state the model's eligibility criteria (33%).

CONCLUSIONS

The high prevalence of weaknesses in model development identified in the published literature is concerning, as these weaknesses are known to compromise predictive validity. CAMPR may support researchers, clinicians and administrators to identify and prevent future weaknesses in model development.

Predicting all-cause unplanned readmission within 30 days of discharge using electronic medical record data: A multi-centre study

OBJECTIVE

To develop a predictive model for identifying patients at high risk of all-cause unplanned readmission within 30 days after discharge, using administrative data available before discharge.

MATERIALS AND METHODS

Hospital administrative data of all adult admissions in three tertiary metropolitan hospitals in Australia between July 01, 2015, and July 31, 2016, were extracted. Predictive performance of four mixed-effect multivariable logistic regression models was compared and validated using a split-sample design. Diagnostic details (Charlson Comorbidity Index CCI, components of CCI, and primary diagnosis categorised into International Classification of Diseases chapters) were added gradually in the clinically simplified model with socio-demographic, index admission, and prior hospital utilisation variables.

RESULTS

Of the total 99 470 patients admitted, 5796 (5.8%) were re-admitted through the emergency department of three hospitals within 30 days after discharge. The clinically simplified model was as discriminative (C-statistic 0.694, 95% CI [0.681-0.706]) as other models and showed excellent calibration. Models with diagnostic details did not exhibit any substantial improvement in predicting 30-days unplanned readmission.

CONCLUSION

We propose a 10-item predictive model to flag high-risk patients in a diverse population before discharge using readily available hospital administrative data which can easily be integrated into the hospital information system.

A cross-sectional study of factors predicting readmission in Thais with coronary artery disease

BACKGROUND

Coronary artery disease is a major cause of morbidity and mortality with high readmission rates. Hospital readmissions for coronary artery disease contribute to rising healthcare costs and are a marker of quality of care. Despite this, prior studies have found that readmission rates vary widely.

AIMS

This study aims to determine the impact of social support, depression, comorbidities, symptom severity, quality of life and readmission among coronary artery disease patients in Thailand.

METHODS

A total of 321 coronary artery disease patients from tertiary care hospitals across all regions of Thailand were recruited for this study. Data were analysed using multiple regression analysis.

RESULTS

The coefficient for social support (beta = -0.22) was found to be significant (p < 0.05), whereas comorbidity, symptom severity, depression and quality of life were not significant. Thus, social support was found to be the most significant predictive factor for readmission.

CONCLUSIONS

Accordingly, when designing effective nursing interventions, nurses should promote social support interventions for coronary artery disease patients to improve the quality of care, decrease readmission rates and improve patients' quality of life.

Effect of a Real-Time Risk Score on 30-day Readmission Reduction in Singapore

OBJECTIVE

To develop a risk score for the real-time prediction of readmissions for patients using patient specific information captured in electronic medical records (EMR) in Singapore to enable the prospective identification of high-risk patients for enrolment in timely interventions.

METHODS

Machine-learning models were built to estimate the probability of a patient being readmitted within 30 days of discharge. EMR of 25,472 patients discharged from the medicine department at Ng Teng Fong General Hospital between January 2016 and December 2016 were extracted retrospectively for training and internal validation of the models. We developed and implemented a real-time 30-day readmission risk score generation in the EMR system, which enabled the flagging of high-risk patients to care providers in the hospital. Based on the daily high-risk patient list, the various interfaces and flow sheets in the EMR were configured according to the information needs of the various stakeholders such as the inpatient medical, nursing, case management, emergency department, and postdischarge care teams.

RESULTS

Overall, the machine-learning models achieved good performance with area under the receiver operating characteristic ranging from 0.77 to 0.81. The models were used to proactively identify and attend to patients who are at risk of readmission before an actual readmission occurs. This approach successfully reduced the 30-day readmission rate for patients admitted to the medicine department from 11.7% in 2017 to 10.1% in 2019 (p < 0.01) after risk adjustment.

CONCLUSION

Machine-learning models can be deployed in the EMR system to provide real-time forecasts for a more comprehensive outlook in the aspects of decision-making and care provision.

Factors Associated With Discharge Destination in Community-Dwelling Adults Admitted to Acute General Medical Units

BACKGROUND AND PURPOSE

General medical patients often present to the hospital with medical, social, cognitive, and functional issues that may impact discharge destination. The aim of this study was to investigate the association between patient factors at hospital admission and discharge destination in general medical patients.

METHODS

This was a prospective, single-site observational study conducted on the general medical wards at a tertiary hospital. Inpatients admitted to the general medical unit and referred to physical therapy were included. Patients admitted from residential care were excluded.

MAIN OUTCOME MEASURES

Data were collected a median of 2 days (interquartile range: 1-3) from hospital admission and included demographics, comorbidities (Charlson Comorbidity Index), premorbid physical function (Blaylock Risk Assessment Screening Score, BRASS), current function (de Morton Mobility Index, DEMMI and Alpha Functional Independence Measure, AlphaFIM), and cognition (Rowland Universal Dementia Assessment Scale, RUDAS).

RESULTS

Between July 2016 and August 2017, 417 patients were recruited (53% female, median age: 81 years (interquartile range: 76-86). Of these, 245 patients were discharged directly home; 172 were not discharged home of whom 140 were discharged to a subacute temporary facility providing further opportunity for therapy and discharge planning. Patients discharged directly home had higher functional, mobility, and cognitive scores. Data were partitioned into training, validation, and test sets to provide unbiased estimates of sensitivity, specificity, receiver operating characteristic curve, and area under the curve. Models best associated with discharge were "DEMMI and toilet transfers" (sensitivity 82.1%, specificity 66.2%, area under the curve 83.8%, 95% confidence interval: 76.4-91.2) and "AlphaFIM and walking independence" (sensitivity: 66.7%, specificity: 83.1%, area under the curve: 81.5, 95% confidence interval: 73.2-89.7).

CONCLUSION

Two models were created that differentiated between discharge home and not home and had similar statistical measures of validity. Although the models require further validation, clinicians should consider whether identification of patients likely to be discharged home or not home is of greater relevance for their clinical setting.

Effects of an Intensive Discharge Intervention on Medication Adherence, Glycemic Control, and Readmission Rates in Patients With Type 2 Diabetes

OBJECTIVES

Patients with diabetes are at particularly high risk for adverse outcomes after hospitalization. The goals of this study were to design, implement, and evaluate a multipronged transitional care intervention among hospitalized patients with diabetes.

METHODS

We randomly assigned inpatients likely to be discharged home on insulin to an intensive transitional care intervention or usual care. The primary outcome was 90-day postdischarge insulin adherence, using prescription refill information to calculate a medication possession ratio. Unadjusted analyses were conducted using Wilcoxon rank sum; adjusted analyses used multivariable linear regression and weighted propensity scoring methods, with general estimating equations to account for clustering by admitting physician.

RESULTS

One hundred eighty patients participated. The mean (SD) medication possession ratio to all insulin types was 84.5% (22.6) among intervention and 76.4% (25.1) among usual care patients (difference = 8.1, 95% confidence interval = -1.0 to 17.2, P = 0.06), with a smaller difference for adherence to all medications (86.3% versus 82.0%). A1c levels decreased in both groups but was larger in the intervention arm (1.09 and 0.11, respectively) (difference = -0.98, 95% confidence interval = -2.03 to -0.07, P = 0.04). Differences between study arms were not significant for rates of hypoglycemic episodes, 30-day readmissions, or emergency department visits. In adjusted/clustered analyses, the difference in A1c reduction remained statistically significant, whereas differences in all other outcomes remained nonsignificant.

CONCLUSIONS

The intervention was associated with improvements in glycemic control, with nonsignificant trends toward greater medication adherence. Further research is needed to optimize and successfully implement interventions to improve patient safety and health outcomes during care transitions.

Factors influencing readmission among Thais with myocardial infarction

Background

Readmission among patients with myocardial infarction is costly, and it has become a marker of quality of care. Therefore, factors related to readmission warrant examination.

Objective

This study aimed at examining factors influencing readmission in Thai with myocardial infarction.

Methods

This was a cross-sectional study with 200 participants randomly selected from five regional hospitals in Thailand. All research tools used indicated acceptable validity and reliability. Linear Structural Relationship version 8.72 was used for the data analysis.

Results

The findings showed that the hypothesized model with social support, depression, symptom severity, comorbidity, and quality of life could explain 4% (R² = 0.04) of the variance in readmission (χ² = 1.39, df = 2, p < 0.50, χ²/df = 0.69, GIF = 1.00, RMSEA = 0.00, SRMR = 0.01, and AGFI = 0.98). Symptom severity was the most influential factor that had a positive and direct effect on the readmission rate (0.06, p < 0.05).

Conclusion

These findings serve as an input to decrease readmission in patients with myocardial infarction by reducing the symptom severity and comorbidity and promoting a better quality of life.

Hospital Admission Location Prediction via Deep Interpretable Networks for the Year-Round Improvement of Emergency Patient Care

OBJECTIVE

This paper presents a deep learning method of predicting where in a hospital emergency patients will be admitted after being triaged in the Emergency Department (ED). Such a prediction will allow for the preparation of bed space in the hospital for timely care and admission of the patient as well as allocation of resource to the relevant departments, including during periods of increased demand arising from seasonal peaks in infections.

METHODS

The problem is posed as a multi-class classification into seven separate ward types. A novel deep learning training strategy was created that combines learning via curriculum and a multi-armed bandit to exploit this curriculum post-initial training.

RESULTS

We successfully predict the initial hospital admission location with area-under-receiver-operating-curve (AUROC) ranging between 0.60 to 0.78 for the individual wards and an overall maximum accuracy of 52% where chance corresponds to 14% for this seven-class setting. Our proposed network was able to interpret which features drove the predictions using a 'network saliency' term added to the network loss function.

CONCLUSION

We have proven that prediction of location of admission in hospital for emergency patients is possible using information from triage in ED. We have also shown that there are certain tell-tale tests which indicate what space of the hospital a patient will use.

SIGNIFICANCE

It is hoped that this predictor will be of value to healthcare institutions by allowing for the planning of resource and bed space ahead of the need for it. This in turn should speed up the provision of care for the patient and allow flow of patients out of the ED thereby improving patient flow and the quality of care for the remaining patients within the ED.

The Effects of a Multifaceted Intervention to Improve Care Transitions Within an Accountable Care Organization: Results of a Stepped-Wedge Cluster-Randomized Trial

BACKGROUND

Transitions from hospital to the ambulatory setting are high risk for patients in terms of adverse events, poor clinical outcomes, and readmission.

OBJECTIVES

To develop, implement, and refine a multifaceted care transitions intervention and evaluate its effects on postdischarge adverse events.

DESIGN, SETTING, AND PARTICIPANTS

Two-arm, single-blind (blinded outcomes assessor), stepped-wedge, cluster-randomized clinical trial. Participants were 1,679 adult patients who belonged to one of 17 primary care practices and were admitted to a medical or surgical service at either of two participating hospitals within a pioneer accountable care organization (ACO).

INTERVENTIONS

Multicomponent intervention in the 30 days following hospitalization, including inpatient pharmacist-led medication reconciliation, coordination of care between an inpatient "discharge advocate" and a primary care "responsible outpatient clinician," postdischarge phone calls, and postdischarge primary care visit.

MAIN OUTCOMES AND MEASURES

The primary outcome was rate of postdischarge adverse events, as assessed by a 30-day postdischarge phone call and medical record review and adjudicated by two blinded physician reviewers. Secondary outcomes included preventable adverse events, new or worsening symptoms after discharge, and 30-day nonelective hospital readmission.

RESULTS

Among patients included in the study, 692 were assigned to usual care and 987 to the intervention. Patients in the intervention arm had a 45% relative reduction in postdischarge adverse events (18 vs 23 events per 100 patients; adjusted incidence rate ratio, 0.55; 95% CI, 0.35-0.84). Significant reductions were also seen in preventable adverse events and in new or worsening symptoms, but there was no difference in readmission rates.

CONCLUSION

A multifaceted intervention was associated with a significant reduction in postdischarge adverse events but no difference in 30-day readmission rates.

Machine learning models to predict length of stay and discharge destination in complex head and neck surgery

BACKGROUND

This study develops machine learning (ML) algorithms that use preoperative-only features to predict discharge-to-nonhome-facility (DNHF) and length-of-stay (LOS) following complex head and neck surgeries.

METHODS

Patients undergoing laryngectomy or composite tissue excision followed by free tissue transfer were extracted from the 2005 to 2017 NSQIP database.

RESULTS

Among the 2786 included patients, DNHF and mean LOS were 421 (15.1%) and 11.7 ± 8.8 days. Four classification models for predicting DNHF with high specificities (range, 0.80-0.84) were developed. The generalized linear and gradient boosting machine models performed best with receiver operating characteristic (ROC), accuracy, and negative predictive value (NPV) of 0.72-0.73, 0.75-0.76, and 0.88-0.89. Four regression models for predicting LOS in days were developed, where all performed similarly with mean absolute error and root mean-squared errors of 3.95-3.98 and 5.14-5.16. Both models were developed into an encrypted web-based interface: https://uci-ent.shinyapps.io/head-neck/.

CONCLUSION

Novel and proof-of-concept ML models to predict DNHF and LOS were developed and published as web-based interfaces.