Clinical and nutritional predictors of hospital readmission within 30 days

Absence of association between low calf circumference, adjusted or not for adiposity, and ICU mortality in critically ill adults: A secondary analysis of a cohort study

BACKGROUND

Despite its correlation with skeletal muscle mass and its predictive value for adverse outcomes in clinical settings, calf circumference is a metric underexplored in intensive care. We aimed to determine whether adjusting low calf circumference for adiposity provides prognostic value superior to its unadjusted measurement for intensive care unit (ICU) mortality and other clinical outcomes in critically ill patients.

METHODS

In a secondary analysis of a cohort study across five ICUs, we assessed critically ill patients within 24 h of ICU admission. We adjusted calf circumference for body mass index (BMI) (25-29.9, 30-39.9, and ≥40) by subtracting 3, 7, or 12 cm from it, respectively. Values ≤34 cm for men and ≤33 cm for women identified low calf circumference.

RESULTS

We analyzed 325 patients. In the primary risk-adjusted analysis, the ICU death risk was similar between the low and preserved calf circumference (BMI-adjusted) groups (hazard ratio, 0.90; 95% CI, 0.47-1.73). Low calf circumference (unadjusted) increased the odds of ICU readmission 2.91 times (95% CI, 1.40-6.05). Every 1-cm increase in calf circumference as a continuous variable reduced ICU readmission odds by 12%. Calf circumference showed no significant association with other clinical outcomes.

CONCLUSION

BMI-adjusted calf circumference did not exhibit independent associations with ICU and in-hospital death, nor with ICU and in-hospital length of stay, compared with its unadjusted measurement. However, low calf circumference (unadjusted and BMI-adjusted) was independently associated with ICU readmission, mainly when analyzed as a continuous variable.

Prevalence of malnutrition based on global leadership initiative in malnutrition criteria for completeness of diagnosis and future risk of malnutrition based on current malnutrition diagnosis: systematic review and meta-analysis

Background

The proposal of the global leadership initiative in malnutrition (GLIM) criteria has received great attention from clinicians. The criteria are mainly used in the research environment and have the potential to be widely used in the clinic in the future. However, the prevalence of malnutrition and risk of future malnutrition based on a current diagnosis of malnutrition are worth exploring.

Methods

A systematic search of PubMed, Embase, and the Cochrane Library was performed from the earliest available date to 1 February 2023. According to the diagnostic criteria of the GLIM, we analysed the prevalence of malnutrition by directly adopting the GLIM criteria for diagnosis without a previous nutritional risk screening (one-step approach) and by adopting the GLIM criteria for diagnosis after a nutritional risk screening (two-step approach). The main outcome was the prevalence of malnutrition based on the one-and two-step approaches. Secondary outcomes were the future risk of malnutrition based on the GLIM diagnosis, including mortality within and beyond 1 year. primary outcomes were pooled using random-effects models, and secondary outcomes are presented as hazard ratios (HRs) and 95% confidence intervals (CIs).

Results

A total of 64 articles were included in the study, including a total of 47,654 adult hospitalized patients and 15,089 malnourished patients based on the GLIM criteria. Malnutrition was diagnosed by the one-step approach in 18 studies and by the two-step approach in 46 studies. The prevalence of malnutrition diagnosed by the one-and two-step approaches was 53% (95% CI, 42%-64%) and 39% (95% CI, 0.35%-0.43%), respectively. The prevalence of malnutrition diagnosed by the GLIM criteria after a nutritional risk screening was quite different; the prevalence of malnutrition diagnosed by the Nutritional Risk Screening 2002 (NRS2002) GLIM tool was 35% (95% CI, 29%-40%); however, the prevalence of malnutrition diagnosed by the Mini Nutrition Assessment (MNA) GLIM tool was 48% (95% CI, 35%-62%). Among the disease types, the prevalence of malnutrition in cancer patients was 44% (95% CI, 36%-52%), while that in acute and critically ill patients was 44% (95% CI, 33%-56%). The prevalence in patients in internal medicine wards was 40% (95% CI, 34%-45%), while that in patients in surgical wards was 47% (95% CI, 30%-64%). In addition, the mortality risk within 1 year (HR, 2.62; 95% CI, 1.95-3.52; I² = 77.1%) and beyond 1 year (HR, 2.04; 95% CI, 1.70-2.45; I² = 59.9%) of patients diagnosed with malnutrition by the GLIM criteria was double that of patients with normal nutrition.

Conclusion

The prevalence of malnutrition diagnosed by the GLIM criteria after a nutritional risk screening was significantly lower than the prevalence of malnutrition diagnosed directly by the GLIM criteria. In addition, the mortality risk was significantly greater among malnourished patients assessed by the GLIM criteria.Systematic review registration: identifier CRD42023398454.

The Impact of Weight Loss Prior to Hospital Readmission

Hospital readmissions place a burden on hospitals. Reducing the readmission number and duration will help reduce the burden. Weight loss might affect readmission risk, especially the risk of an early (<30 days) readmission. This study sought to identify the predictors and the impact of weight loss prior to a delayed readmission (>30 days). Body mass index (BMI) was measured during the index admission and first readmission. Patients, after their readmission, were assessed retrospectively to identify the characteristics of those who had lost >5% weight prior to that readmission. Length of stay (LOS), time spent in the intensive care unit (ICU) and the one-year mortality of those patients who lost weight were compared to the outcomes of those who remained weight-stable using multilevel mixed-effects regression adjusting for BMI, Charlson comorbidity index (CCI), ICU hours and relative stay index (RSI). Those who were at risk of weight loss prior to readmission were identifiable based upon their age, BMI, CCI and LOS. Of 1297 patients, 671 (51.7%) remained weight-stable and 386 (29.7%) lost weight between admissions. During their readmission, those who had lost weight had a significantly higher LOS (IRR 1.17; 95% CI 1.12, 1.22: p < 0.001), RSI (IRR 2.37; 95% CI 2.27, 2.47: p < 0.001) and an increased ICU LOS (IRR 2.80; 95% CI 2.65, 2.96: p < 0.001). This study indicates that weight loss prior to a delayed readmission is predictable and leads to worse outcomes during that readmission.

Development and Internal Validation of an Interpretable Machine Learning Model to Predict Readmissions in a United States Healthcare System

(1) One in four hospital readmissions is potentially preventable. Machine learning (ML) models have been developed to predict hospital readmissions and risk-stratify patients, but thus far they have been limited in clinical applicability, timeliness, and generalizability. (2) Methods: Using deidentified clinical data from the University of California, San Francisco (UCSF) between January 2016 and November 2021, we developed and compared four supervised ML models (logistic regression, random forest, gradient boosting, and XGBoost) to predict 30-day readmissions for adults admitted to a UCSF hospital. (3) Results: Of 147,358 inpatient encounters, 20,747 (13.9%) patients were readmitted within 30 days of discharge. The final model selected was XGBoost, which had an area under the receiver operating characteristic curve of 0.783 and an area under the precision-recall curve of 0.434. The most important features by Shapley Additive Explanations were days since last admission, discharge department, and inpatient length of stay. (4) Conclusions: We developed and internally validated a supervised ML model to predict 30-day readmissions in a US-based healthcare system. This model has several advantages including state-of-the-art performance metrics, the use of clinical data, the use of features available within 24 h of discharge, and generalizability to multiple disease states.

A scoping review on the GLIM criteria for malnutrition diagnosis: Understanding how and for which purpose it has been applied in studies on hospital settings

AIMS

This scoping review aimed to identify and map the literature on malnutrition diagnosis made using the GLIM criteria in hospitalized patients.

METHODS

The scoping review was conducted using the Joanna Briggs Institute's methodology. We searched PubMed, Embase, Scopus, and Web of Science (until 16 April 2022) to identify studies based on the 'population' (adults or elderly patients), 'concept' (malnutrition diagnosis by the GLIM criteria), and 'context' (hospital settings) framework. Titles/abstracts were screened, and two independent reviewers extracted data from eligible studies.

RESULTS

Ninety-six studies were eligible (35.4% from China, 30.2% involving oncological patients, and 30.5% with prospective data collection), 32 followed the two-step GLIM approach, and 50 applied all the criteria. All the studies evaluated body mass index (BMI), while 92.7% evaluated weight loss; 77.1%, muscle mass; 93.8%, inflammation; and 70.8%, energy intake. A lack of details on the methods adopted for criterion evaluation was observed in five (muscle mass evaluation) to 40 studies (energy intake evaluation). The frequency of the use of the GLIM criteria ranged from 22.2% (frequency of low BMI) to 84.7% (frequency of inflammation), and the malnutrition prevalence ranged from 0.96% to 87.9%. Less than 30% of studies aimed to assess the GLIM criterion validity, eight studies cited the guidance on validation of the GLIM criteria, and a minority implemented it.

CONCLUSIONS

This map of studies on the GLIM criteria in hospital settings demonstrated that they are applied in a heterogeneous manner, with a wide range of malnutrition prevalence. Almost 50% of the studies applied all the criteria, while one-third followed the straightforward two-step approach. The recommendations of the guidance on validation of the criteria were scarcely adhered to. The gaps that need to be explored in future studies have been highlighted.

Time to readmission and associated factors after post treatment discharge of severe acute malnourished under-five children in Pawe General Hospital

Background

Relapse or repeated episodes is the admission of a child with the diagnosis of severe acute malnutrition (SAM) after being discharged to a status of treated and cured from a stabilizing center. A child may experience more than one episode of SAM depending on the improvement of the underlying comorbidity. Thus, this study aimed to estimate the time to readmission of SAM and associated factors for under-five children in North West Ethiopia.

Methods

An institution-based retrospective cohort study was employed in 760 files of under-five children spanning from 2014/15 to 2019/20. The data extraction tool was developed from SAM treatment guidelines and medical history sheets. Epi Data version 3.2 and STATA version 14 were used for data entry and final analysis, respectively. After checking all assumptions, the multivariable Cox Proportional Hazard model was fitted to the isolated independent predictors for time to readmission. A categorical variable with p < 0.05 was considered a risk factor for the relapse of SAM.

Result

The mean (± SD) age of participant children was 27.8 (± 16.5) months with mean (± SD) time to relapse of SAM cases were 30.4(± 21.39) weeks posttreatment discharge. The overall incidence density rate of relapse was determined as 10.8% (95% CI 8.3; 12.6). The average time (± SD) for treatment recovery from the first admission of the SAM case was 28.8(± 18.7) days. Time of readmission was significantly associated with living in rural resident (AHR 5⋅3 = 95% CI, 2⋅95, 13⋅87, p = 0.021), having HIV infection (AHR6⋅8 = 95%CI; 4.1–11.9 p = 0.001), and first admission with edema (AHR = 3.5 = 95% CI; 1.92, 6.2, p = 0.018).

Conclusion

Nearly one in every ten severely acute malnourished under-five children relapsed within a mean time to relapse 30.4(± 21.39) weeks posttreatment discharge. Time to relapse was significantly associated with being a rural resident for children, having edema during the first admission, and being HIV-infected cases. A protocol ought to be drafted for extending Supplementary Nutrition in Acute Malnutrition management program following discharge is highly needed.

Supplementary Information

The online version contains supplementary material available at 10.1186/s41043-022-00308-8.

Nationwide hospital admission data statistics and disease-specific 30-day readmission prediction

Purpose

Hospital readmission prediction uses historical patient visit data to train machine learning models to predict risk of patients being readmitted after the discharge. Data used to train models, such as patient demographics, disease types, localized distributions etc., play significant roles in the model performance. To date, many methods exist for hospital readmission prediction, but answers to some important questions still remain open. For example, how will demographics, such as gender, age, geographic, impact on readmission prediction? Do patients suffering from different diseases vary significantly in their readmission rates? What are the nationwide hospital admission data characteristics? and how do hospital speciality, ownership, and locations impact on their readmission rates? In this study, we carry systematic investigations to answer the above questions, and propose a predictive modeling framework to predict disease-specific 30-day hospital readmission.

Methods

We first implement statistics analysis by using National Readmission Databases (NRD) with over 15 million hospital visits. After that, we create features and disease-specific readmission datasets. An ensemble learning framework is proposed to conduct hospital readmission prediction and Friedman test and Nemenyi post-hoc test is used to validate our proposed method.

Results

Using National Readmission Databases (NRD), with over 15 million hospital visits, as our testbed, we summarize nationwide patient admission data statistics, in related to demographic, disease types, and hospital factors. We use feature engineering to design 526 representative features to model each patient visit. Our studies found that readmission rates vary significantly from diseases to diseases. For six diseases studied in our research, their readmission rates vary from 1.832 (Pneumonia) to 8.761% (Diabetes). Using random sampling and voting approaches, our study shows that soft voting outperforms hard voting on majority results, especially for AUC and balanced accuracy which are the main measures for imbalanced data. Random under sampling using 1.1:1 for negative:positive ratio achieves the best performance for AUC, balanced accuracy, and F1-score.

Conclusion

This paper carries out systematic studies to understand US nationwide hospital readmission data statistics, and further designs a machine learning framework for disease-specific 30-day hospital readmission prediction. Our study shows that hospital readmission rates vary significantly with respect to different disease types, gender, age groups, any other factors. Gradient boosting achieves the best performance for disease specific hospital readmission prediction.

Malnutrition in Hospitalized Old Patients: Screening and Diagnosis, Clinical Outcomes, and Management

Malnutrition in hospitalized patients heavily affects several clinical outcomes. The prevalence of malnutrition increases with age, comorbidities, and intensity of care in up to 90% of old populations. However, malnutrition frequently remains underdiagnosed and undertreated in the hospital. Thus, an accurate screening to identify patients at risk of malnutrition or malnourishment is determinant to elaborate a personal nutritional intervention. Several definitions of malnutrition were proposed in the last years, affecting the real frequency of nutritional disorders and the timing of intervention. Diagnosis of malnutrition needs a complete nutritional assessment, which is often challenging to perform during a hospital stay. For this purpose, various screening tools were proposed, allowing patients to be stratified according to the risk of malnutrition. The present review aims to summarize the actual evidence in terms of diagnosis, association with clinical outcomes, and management of malnutrition in a hospital setting.