A Systematic Review of Highly Generalizable Risk Factors for Unplanned 30-Day All-Cause Hospital Readmissions

Abstract

Background: Hospital readmissions are common and expensive. Numerous global efforts have been devoted to predicting readmissions. However, for many reasons including the variations in the studied populations and inconsistent definitions of readmissions, the outcomes of some studies can hardly be generalized to other studies inside or outside the same country or region. Objective:The objective was to identify highly generalizable risk factors for unplanned 30-day all-cause hospital readmissions to guide the selection of baseline predictor variables in different readmission studies.Methods: In July 2017, PubMed was searched to identify articles pertaining to the risk factors for unplanned 30-day all-cause hospital readmissions. To identify potentially eligible risk factors, characteristics of the selected studies were manually extracted. The generalizability of the risk factors was assessed with predefined criteria.Results: 13 articles were eligible for the review. A total of 42 risk factors were identified and 34 of them were found to be highly generalizable. Conclusions:The 34 risk factors are not specific to any populations or places, and the corresponding predictor variables can serve as baseline variables in readmission prediction studies. No major difference has been observed between the risk factors identified inside and outside the United States except that US studies appeared to prefer composite comorbidity measures. All the reviewed studies have used traditional statistical regression-based methods to identify risk factors and more applications of data mining techniques are expected in this field.

A multi-modal machine learning approach towards predicting patient readmission

“…Discharge disposition was also a strong feature, including discharge to a short-term hospital and psychiatric hospital (compared to discharge to home or self care which was not a strong feature of readmission). Research also indicates that the type of index admission such as emergency department admissions (here referred to Admit Source-Self Referral) are strongly associated with readmission [3] An unexpected finding was that almost all patients who were discharged to a rehabilitation hospital were readmitted within 30 days (844 of 925 patients, 92%); this was among the strongest features associated with readmission. This clinical setting accommodates patient with greater medical complexity requiring professional nursing and rehabilitation services for problems such as hip fracture and stroke recovery.…”

“…Discharge disposition was also a strong feature, including discharge to a short-term hospital and psychiatric hospital (compared to discharge to home or self care which was not a strong feature of readmission). Research also indicates that the type of index admission such as emergency department admissions (here referred to Admit Source-Self Referral) are strongly associated with readmission [3]…”

“…Another potential factor that might lead to unplanned readmission is the presence of one or more co-morbidities or the use of high risk medications. Comorbidity, defined as the coexistence of two or more medical conditions in a patient, is commonly included in risk prediction models to account for patient heterogeneity and chronic disease burden that is associated with more complex care and poor clinical outcomes [3].…”

Health Disparities of Black Americans Hospitalized for Decompensated Liver Cirrhosis

“…Chronic comorbid conditions 30,31 and complications of liver cirrhosis 32 are both predictors of hospital readmission and poor healthcare outcomes. In patients with cirrhosis, hepatic encephalopathy (HE) and ascites are most strongly associated with readmission within 30 or 90 days.…”

Predicting 30-Day Hospital Readmission for Diabetes Patients using Multilayer Perceptron

“…This review, far from being exhaustive, succeeded in extracting a limited number of optimum performing models among the current body of literature. Despite providing additional knowledge contribution to previous systematic reviews, the listed models are still limited in term of performance [9], [17]. As such, this review reveals a clear gap in term of model's classification performance, model generalizability and data pre-processing.…”