Developing Prediction Models for 30-Day Unplanned Readmission Among Children With Medical Complexity

Unplanned Readmissions in Children with Medical Complexity in Saudi Arabia: A Large Multicenter Study

Background

Children with medical complexity (CMC) account for a substantial proportion of healthcare spending, and one-third of their expenditures are due to readmissions. However, knowledge regarding the healthcare-resource utilization and characteristics of CMC in Saudi Arabia is limited.

Objectives

To describe hospitalization patterns and characteristics of Saudi CMC with an unplanned 30-day readmission.

Methodology

This retrospective study included Saudi CMC (aged 0-14 years) who had an unplanned 30-day readmission at six tertiary centers in Riyadh, Jeddah, Dammam, Alahsa, and Almadina between January 2016 and December 2020. Hospital-based inclusion criteria focused on CMC with multiple complex chronic conditions (CCCs) and technology assistance (TA) device use. CMC were compared across demographics, clinical characteristics, and hospital-resource utilization.

Results

A total of 9139 pediatric patients had unplanned 30-day readmission during the study period, of which 680 (7.4%) met the inclusion criteria. Genetic conditions were the most predominant primary pathology (66.3%), with one-third of cases (33.7%) involving the neuromuscular system. During the index admission, pneumonia was the most common diagnosis (33.1%). Approximately 35.1% of the readmissions were after 2 weeks. Pneumonia accounted for 32.5% of the readmissions. After readmission, 16.9% of patients were diagnosed with another CCC or received a new TA device, and the in-hospital mortality rate was 6.6%.

Conclusion

The rate of unplanned 30-day readmissions in children with medical complexity in Saudi Arabia is 7.4%, which is lower than those reported from developed countries. Saudi children with CCCs and TA devices were readmitted approximately within similar post-discharge time and showed distinct hospitalization patterns associated with specific diagnoses. To effectively reduce the risk of 30-day readmissions, targeted measures must be introduced both during the hospitalization period and after discharge.

How are children with medical complexity being identified in epidemiological studies? A systematic review

BACKGROUND

There are different definitions to identify/classify children with medical complexity (CMC). We aimed to investigate and describe the definitions used to classify CMC in epidemiological studies.

METHODS

PubMed, SciELO, LILACS, and EMBASE were searched from 2015 to 2020 (last updated September 15th, 2020) for original studies that presented the definition used to classify/identify CMC in the scientific research method. We applied the Preferred Reporting Items for Systematic Reviews and Meta-Analyses methodology. From the included studies, the following were identified: first author, year of publication, design, population, study period, the definition of CMC used, limitations, and strengths.

RESULTS

Nine hundred and sixty-seven records were identified in the searched databases, and 42 met the inclusion criteria. Of the 42 studies included, the four most frequent definitions used in the articles included in this review were classification of CMC into nine diagnostic categories based on the International Classification of Diseases, Ninth Revision (ICD-9) (35.7%, 15 articles); update of the previous classification for ICD-10 codes with the inclusion of other conditions in the definition (21.4%, nine articles); definition based on a medical complexity algorithm for classification (16.7%, seven articles); and a risk rating system (7.1%, three articles).

CONCLUSIONS

CMC definitions using diagnostic codes were more frequent. However, several limitations were found in its uses. Our research highlighted the need to improve health information systems to accurately characterize the CMC population and promote the provision of comprehensive care.

Predictive Modeling to Identify Children With Complex Health Needs At Risk for Hospitalization

BACKGROUND

Identifying children at high risk with complex health needs (CCHN) who have intersecting medical and social needs is challenging. This study's objectives were to (1) develop and evaluate an electronic health record (EHR)-based clinical predictive model ("model") for identifying high-risk CCHN and (2) compare the model's performance as a clinical decision support (CDS) to other CDS tools available for identifying high-risk CCHN.

METHODS

This retrospective cohort study included children aged 0 to 20 years with established care within a single health system. The model development/validation cohort included 33 months (January 1, 2016-September 30, 2018) and the testing cohort included 18 months (October 1, 2018-March 31, 2020) of EHR data. Machine learning methods generated a model that predicted probability (0%-100%) for hospitalization within 6 months. Model performance measures included sensitivity, positive predictive value, area under receiver-operator curve, and area under precision-recall curve. Three CDS rules for identifying high-risk CCHN were compared: (1) hospitalization probability ≥10% (model-predicted); (2) complex chronic disease classification (using Pediatric Medical Complexity Algorithm [PMCA]); and (3) previous high hospital utilization.

RESULTS

Model development and testing cohorts included 116 799 and 27 087 patients, respectively. The model demonstrated area under receiver-operator curve = 0.79 and area under precision-recall curve = 0.13. PMCA had the highest sensitivity (52.4%) and classified the most children as high risk (17.3%). Positive predictive value of the model-based CDS rule (19%) was higher than CDS based on the PMCA (1.9%) and previous hospital utilization (15%).

CONCLUSIONS

A novel EHR-based predictive model was developed and validated as a population-level CDS tool for identifying CCHN at high risk for future hospitalization.

Trends and Burden of Firearm-Related Injuries Among Children and Adolescents: A National Perspective

INTRODUCTION

Firearm-related injuries in America have been under increasing scrutiny over the last several years. Few studies have examined the burden of these injuries in the pediatric population. The objective of this study was to describe the incidence of firearm-related injuries in hospitalized pediatric patients in the United States and identify the risk factors associated with readmission in this young population.

METHODS

The Nationwide Readmission Database was examined from 2010 to 2017. Pediatric patients (aged ≤18 y) who survived their index hospitalization for any firearm injury were analyzed to determine incidence rate, case fatality rate, risk factors for 30-d readmission, and financial health care burden.

RESULTS

There were 35,753 pediatric firearm injuries (86.8% male) with an overall incidence rate of 10.49 (95% confidence interval [CI]: 9.26-11.71) per 100,000 pediatric hospitalizations. Adolescents aged >12 y had the highest incidence rate (60.51, 95% CI: 55.19-65.84). In-hospital mortality occurred in 1948 cases (5.5%), with higher case fatality rates in males. There were 1616 (5.7%) unplanned 30-d readmissions. Multivariate analysis showed abdominal firearm injuries (hazard ratio: 1.13, 95% CI: 1.03-1.24; P = 0.006) and longer length of stay (hazard ratio: 1.27, 95% CI: 1.04-1.55; P = 0.016) were associated with a greater risk of 30-d readmission. The median health care cost for firearm-related injuries was $36,535 (interquartile range: $19,802-$66,443), 22% of which was due to readmissions. Cost associated with 30-d readmissions was $7978 (interquartile range: $4305-$15,202).

CONCLUSIONS

Firearm-related injury is a major contributor to pediatric morbidity, mortality, and health care costs. Males are disproportionately affected by firearm injury, but females are more likely to require unplanned 30-d readmissions. Interventions should target female sex, injuries of suicidal intent, psychiatric comorbidities, prolonged index hospitalization, and abdominal injuries.

Performance Characteristics of a Machine-Learning Tool to Predict 7-Day Hospital Readmissions

OBJECTIVES

To develop an institutional machine-learning (ML) tool that utilizes demographic, socioeconomic, and medical information to stratify risk for 7-day readmission after hospital discharge; assess the validity and reliability of the tool; and demonstrate its discriminatory capacity to predict readmissions.

PATIENTS AND METHODS

We performed a combined single-center, cross-sectional, and prospective study of pediatric hospitalists assessing the face and content validity of the developed readmission ML tool. The cross-sectional analyses used data from questionnaire Likert scale responses regarding face and content validity. Prospectively, we compared the discriminatory capacity of provider readmission risk versus the ML tool to predict 7-day readmissions assessed via area under the receiver operating characteristic curve analyses.

RESULTS

Overall, 80% (15 of 20) of hospitalists reported being somewhat to very confident with their ability to accurately predict readmission risk; 53% reported that an ML tool would influence clinical decision-making (face validity). The ML tool variable exhibiting the highest content validity was history of previous 7-day readmission. Prospective provider assessment of risk of 413 discharges showed minimal agreement with the ML tool (κ = 0.104 [95% confidence interval 0.028-0.179]). Both provider gestalt and ML calculations poorly predicted 7-day readmissions (area under the receiver operating characteristic curve: 0.67 vs 0.52; P = .11).

CONCLUSIONS

An ML tool for predicting 7-day hospital readmissions after discharge from the general pediatric ward had limited face and content validity among pediatric hospitalists. Both provider and ML-based determinations of readmission risk were of limited discriminatory value. Before incorporating similar tools into real-time discharge planning, model calibration efforts are needed.

Mortality During Readmission Among Children in United States Children's Hospitals

OBJECTIVE

To identify demographic, clinical, and hospital factors associated with mortality on readmission within 180 days following an inpatient hospitalization.

STUDY DESIGN

We conducted a retrospective cohort study including 33 US children's hospitals in the Pediatric Health Information System from January 2010 to June 2020. Our primary outcome was death during readmission within 180 days of an index hospitalization among children aged 0-18 years. Illness severity during the index hospitalization was defined according to the All Patient-Refined Diagnosis-Related Group-categorized illness severity (ie, minor, moderate, or major/extreme). We performed multivariable logistic regression analysis to identify factors during the index hospitalization associated with mortality during readmission.

RESULTS

Among 2 677 111 children discharged, 337 385 (12.6%) were readmitted within 180 days of the index hospitalization and 2913 (0.8%) died during readmission. More than one-quarter (26.2%) of deaths among children who were readmitted and died occurred within 10 days after discharge from the index hospitalization. Factors independently associated with mortality during readmission included multiple complex chronic conditions, index admissions lasting >7 days, moderate or severe/extreme illness during the index hospitalization, and public insurance. Children whose race was reported as Black had greater odds of mortality during readmission compared with children of other races.

CONCLUSIONS

Among hospitalized children, several demographic and clinical factors present during index hospitalizations were associated with mortality during readmission. Greater odds of mortality during readmission among children whose race was reported as Black likely reflects disparities in social determinants of health and clinical care. Interventions to reduce mortality during readmission may target high-risk populations in the period immediately following discharge.

Associations between comorbidity-related functional limitations and pneumonia outcomes

BACKGROUND

Underlying comorbidities are common in children with pneumonia.

OBJECTIVE

To determine associations between comorbidity-related functional limitations and risk for severe pneumonia outcomes.

DESIGN, SETTING, AND PARTICIPANTS

We prospectively enrolled children <18 years with and without comorbidities presenting to the emergency department with clinical and radiographic pneumonia at two institutions. Comorbidities included chronic conditions requiring daily medications, frequent healthcare visits, or which limited age-appropriate activities. Among children with comorbidities, functional limitations were defined as none or mild, moderate, and severe.

MAIN OUTCOMES AND MEASURES

Outcomes included an ordinal severity outcome, categorized as very severe (mechanical ventilation, shock, or death), severe (intensive care without very severe features), moderate (hospitalization without severe features), or mild (discharged home), and length of stay (LOS). Multivariable ordinal logistic regression was used to examine associations between comorbidity-related functional limitations and outcomes, while accounting for relevant covariates.

RESULTS

A cohort of 1116 children, including 452 (40.5%) with comorbidities; 200 (44.2%) had none or mild functional limitations, 93 (20.6%) moderate, and 159 (35.2%) had severe limitations. In multivariable analysis, comorbidity-related functional limitations were associated with the ordinal severity outcome and LOS (p < .001 for both). Children with severe functional limitations had tripling of the odds of a more severe ordinal (adjusted odds ratio [aOR]: 3.01, 95% confidence interval [2.05, 4.43]) and quadrupling of the odds for longer LOS (aOR: 4.72 [3.33, 6.70]) as compared to children without comorbidities.

CONCLUSION

Comorbidity-related functional limitations are important predictors of disease outcomes in children with pneumonia. Consideration of functional limitations, rather than the presence of comorbidity alone, is critical when assessing risk of severe outcomes.

Factors associated with admission of children to an intensive care unit and readmission to hospital within 28 days of discharge: A population-based study

AIM

Hospital readmissions within 28 days are an important performance measurement of quality and safety of health care. The aims of this study were to examine the rates, trends and characteristics of paediatric intensive care unit admissions, and factors associated with readmissions to hospital within 28 days of discharge.

METHODS

This retrospective, population-based record linkage study included all children ≥28 days and <16 years old admitted to an intensive care unit (ICU) in a New South Wales (NSW) public hospital from 2004 to 2013. Data were sourced from the NSW Admitted Patients Data Collection and the NSW Registry of Births, Deaths and Marriages, Death Registration.

RESULTS

We identified 21 200 ICU admissions involving 17 130 children. Admissions increased by 24% over the study period with the greatest increase attributed to respiratory and musculoskeletal conditions. A higher proportion of children were <5 years, male, lived in major cities, were publicly insured and had chronic conditions. The median length of ICU stay was 42 h and overall hospital stay was 7 days. There were 905 deaths, two-thirds during the index admission with the leading causes being injuries, cancer and infections. Twenty-three per cent of ICU admissions were readmitted to hospital within 28 days of discharge. Associated independent factors were younger age, longer index hospital stay and emergency index admission. Children with chronic conditions of cancer and genitourinary disorders were more likely to be readmitted.

CONCLUSIONS

Identification of complex chronic conditions, consideration of long-term health planning and interventions intended to reduce readmission is warranted in order to reduce the burden to families and the health-care system.

Applicability of predictive models for 30-day unplanned hospital readmission risk in paediatrics: a systematic review

OBJECTIVES

To summarise multivariable predictive models for 30-day unplanned hospital readmissions (UHRs) in paediatrics, describe their performance and completeness in reporting, and determine their potential for application in practice.

DESIGN

Systematic review.

DATA SOURCE

CINAHL, Embase and PubMed up to 7 October 2021.

ELIGIBILITY CRITERIA

English or German language studies aiming to develop or validate a multivariable predictive model for 30-day paediatric UHRs related to all-cause, surgical conditions or general medical conditions were included.

DATA EXTRACTION AND SYNTHESIS

Study characteristics, risk factors significant for predicting readmissions and information about performance measures (eg, c-statistic) were extracted. Reporting quality was addressed by the 'Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis' (TRIPOD) adherence form. The study quality was assessed by applying six domains of potential biases. Due to expected heterogeneity among the studies, the data were qualitatively synthesised.

RESULTS

Based on 28 studies, 37 predictive models were identified, which could potentially be used for determining individual 30-day UHR risk in paediatrics. The number of study participants ranged from 190 children to 1.4 million encounters. The two most common significant risk factors were comorbidity and (postoperative) length of stay. 23 models showed a c-statistic above 0.7 and are primarily applicable at discharge. The median TRIPOD adherence of the models was 59% (P₂₅-P₇₅, 55%-69%), ranging from a minimum of 33% to a maximum of 81%. Overall, the quality of many studies was moderate to low in all six domains.

CONCLUSION

Predictive models may be useful in identifying paediatric patients at increased risk of readmission. To support the application of predictive models, more attention should be placed on completeness in reporting, particularly for those items that may be relevant for implementation in practice.

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.

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.

Predictors of pediatric readmissions among patients with neurological conditions

Background

Unplanned readmission is one of many measures of the quality of care of pediatric patients with neurological conditions. In this multicenter study, we searched for novel risk factors of readmission of patients with neurological conditions.

Methods

We retrieved hospitalization data of patients less than 18 years with one or more neurological conditions. This resulted in a total of 105,834 encounters from 18 hospitals. We included data on patient demographics, prior healthcare resource utilization, neurological conditions, number of other conditions/diagnoses, number of medications, and number of surgical procedures performed. We developed a random intercept logistic regression model using stepwise minimization of Akaike Information Criteria for variable selection.

Results

The most important neurological conditions associated with unplanned pediatric readmissions include hydrocephalus, inflammatory diseases of the central nervous system, sleep disorders, disease of myoneural junction and muscle, other central nervous system disorder, other spinal cord conditions (such as vascular myelopathies, and cord compression), and nerve, nerve root and plexus disorders. Current and prior healthcare resource utilization variables, number of medications, other diagnoses, and certain inpatient surgical procedures were associated with changes in odds of readmission. The area under the receiver operator characteristic curve (AUROC) on the independent test set is 0.733 (0.722, 0.743).

Conclusions

Pediatric patients with certain neurological conditions are more likely to be readmitted than others. However, current and prior healthcare resource utilization remain some of the strongest indicators of readmission within this population as in the general pediatric population.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12883-020-02028-0.

The Role of Caregiver-Reported Risks in Predicting Adverse Pediatric Outcomes

OBJECTIVE

Certain social risk factors (e.g., housing instability, food insecurity) have been shown to directly and indirectly influence pediatric health outcomes; however, there is limited understanding of which social factors are most salient for children admitted to the hospital. This study examines how caregiver-reported social and medical characteristics of children experiencing an inpatient admission are associated with the presence of future health complications.

METHODS

Caregivers of children experiencing an inpatient admission (N = 249) completed a predischarge questionnaire designed to capture medical and social risk factors across systems (e.g., patient, caregiver, family, community, healthcare environment). Electronic health record (EHR) data were reviewed for child demographic data, chronic disease status, and subsequent emergency department visits or readmissions (i.e., acute events) 90 days postindex hospitalization. Associations between risk factors and event presence were estimated using odds ratios (ORs) and confidence intervals (CI), both unadjusted and adjusted OR (aOR) for chronic disease and age.

RESULTS

Thirty-three percent (N = 82) of children experienced at least one event. After accounting for child age and chronic disease status, caregiver perceptions of child's health being generally "poor" or "not good" prior to discharge (aOR = 4.7, 95% CI = 2.3, 9.7), having high care coordination needs (aOR = 3.2, 95% CI = 1.6, 6.1), and experiencing difficulty accessing care coordination (aOR = 2.5, 95% CI = 1.4, 4.7) were significantly associated with return events.

CONCLUSIONS

Caregiver report of risks may provide valuable information above and beyond EHR records to both determine risk of future health problems and inform intervention development.

Children dependent on respiratory support: A 10-year review from one pediatric postacute care hospital

OBJECTIVE

Pediatric postacute care hospitals (PACH) provide long-term care for children with medical complexity including children dependent on respiratory support. Descriptions of PACH respiratory care populations and outcomes, however, remain under-reported. Our aim was to describe demographics, respiratory outcome, and longitudinal trend of children with respiratory support admitted to a single PACH in the United States.

METHODS

Using electronic records from 2009 to 2018, data were examined for all children dependent on respiratory support. Children were identified for inclusion using respiratory level of care classifications (type of support) as outlined in hospital policy. Outcome was defined as change in level from first admission to final discharge. Number of admissions by level and year during the study timeframe were analyzed.

RESULTS

There were 1423 admissions for 767 children requiring respiratory support during the study timeframe. Children with higher respiratory classification level (eg, mechanical ventilation) at initial admission had more admissions to PACH (P < .001) and longer length of stays (P < .001). From first admission to final discharge, there was a significant change (reduction) in respiratory level (z = -4.588, P < .001). An increase in the overall number of admissions for children with respiratory support during the study timeframe was noted, with the largest increase for children requiring the highest level of support.

CONCLUSION

There has been a consistent increase in the number of children requiring respiratory support at admission to PACH. Reduction in respiratory support with postacute care occurs but children admitted with a higher level of support stay longer and experience multiple admissions.

Care Coordination Program for Children With Complex Chronic Conditions Discharged From a Rural Tertiary-Care Academic Medical Center

OBJECTIVES

Hospital discharge offers an opportunity to initiate coordination of follow-up care, preventing readmissions or emergency department (ED) recidivism. We evaluated how revisits and costs of care varied in a 12-month period between children in a care coordination program at our center (enrolled after hospital discharge with a tracheostomy or on a ventilator) and children with complex chronic condition discharges who were not enrolled.

METHODS

Children ages 1 to 17 years were retrospectively included if they had a hospital discharge in 2017 with an International Classification of Diseases, 10th Revision code meeting complex chronic condition criteria or if they were in active follow-up with the care coordination program. Revisits and total costs of care were compared over 2018 for included patients.

RESULTS

Seventy patients in the program were compared with 56 patients in the control group. On bivariate analysis, the median combined number of hospitalizations and ED visits in 2018 was lower among program participants (0 vs 1; P = .033), and program participation was associated with lower median total costs of care in 2018 ($700 vs $3200; P = .024). On multivariable analysis, care coordination program participation was associated with 59% fewer hospitalizations in 2018 (incidence rate ratio: 0.41; 95% confidence interval: 0.23 to 0.75; P = .004) but was not significantly associated with reduced ED visits or costs.

CONCLUSIONS

The care coordination program is a robust service spanning the continuum of patient care. We found program participation to be associated with reduced rehospitalization, which is an important driver of costs for children with medical complexity.

Parent Perspectives During Hospital Readmissions for Children With Medical Complexity: A Qualitative Study

OBJECTIVES

Children with medical complexity (CMC) have high readmission rates, but relatively little is known from the parent perspective regarding care experiences surrounding and factors contributing to readmissions. We aimed to elicit parent perspectives on circumstances surrounding 30-day readmissions for CMC.

METHODS

We conducted 20 semistructured interviews with parents of CMC experiencing an unplanned 30-day readmission at 1 academic medical center between December 2016 and January 2018, asking about topics such as previous discharge experiences, medical services and resources, and home environment and social support. Interviews were recorded, professionally transcribed, and analyzed thematically by using a modified grounded theory approach.

RESULTS

Children ranged in age from 0 to 15 years, with neurologic complex chronic conditions being predominant (35%). Although the majority of parents did not identify any factors that they perceived to have contributed to readmission, themes emerged regarding challenges associated with chronicity of care and transitions of care that might influence readmissions, including frequency of hospital use, symptom confusion, lack of inpatient continuity, resources needed but not received, and difficulty filling prescriptions.

CONCLUSIONS

Parents identified multiple challenges associated with chronicity of medical management and transitions of care for CMC. Future interventions aiming to improve continuity and communication between admissions, ensure that home services are provided when applicable and prescriptions are filled, and provide comprehensive support for families in both the short- and long-term may help improve patient and family experiences while potentially decreasing readmissions.

Development and Validation of a Web-Based Pediatric Readmission Risk Assessment Tool

OBJECTIVES

Accurately predicting and reducing risk of unplanned readmissions (URs) in pediatric care remains difficult. We sought to develop a set of accurate algorithms to predict URs within 3, 7, and 30 days of discharge from inpatient admission that can be used before the patient is discharged from a current hospital stay.

METHODS

We used the Children's Hospital Association Pediatric Health Information System to identify a large retrospective cohort of 1 111 323 children with 1 321 376 admissions admitted to inpatient care at least once between January 1, 2016, and December 31, 2017. We used gradient boosting trees (XGBoost) to accommodate complex interactions between these predictors.

RESULTS

In the full cohort, 1.6% of patients had at least 1 UR in 3 days, 2.4% had at least 1 UR in 7 days, and 4.4% had at least 1 UR within 30 days. Prediction model discrimination was strongest for URs within 30 days (area under the curve [AUC] = 0.811; 95% confidence interval [CI]: 0.808-0.814) and was nearly identical for UR risk prediction within 3 days (AUC = 0.771; 95% CI: 0.765-0.777) and 7 days (AUC = 0.778; 95% CI: 0.773-0.782), respectively. Using these prediction models, we developed a publicly available pediatric readmission risk scores prediction tool that can be used before or during discharge planning.

CONCLUSIONS

Risk of pediatric UR can be predicted with information known before the patient's discharge and that is easily extracted in many electronic medical record systems. This information can be used to predict risk of readmission to support hospital-discharge-planning resources.