A deep attention model to forecast the Length Of Stay and the in-hospital mortality right on admission from ICD codes and demographic data

Harerimana, Gaspard; Kim, Jong Wook; Jang, Beakcheol

doi:10.1016/j.jbi.2021.103778

Cited by 28 publications

(14 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For instance, Zebin & Chaussalet [12] used an autoencoder deep neural network to categorize short stays (0-7 days) and long stays (>7 days) using Medical Information Mart for Intensive Care III (MIMIC III) dataset [13] but the dataset lacked multi-modalities and methodology does not analyze cohort of risk factors responsible for extended LoS. Likewise, Harerimana et al [14] proposed an attention-based DL method to predict LoS and in-hospital mortality but this method also has limitations of using limited lab data and unavailability of radiological information. Rajkomar et al [15] proposed a three-tier approach by combining three DL models to predict hospital readmission and patients LoS, based on data belonging to patients with varying diseases, without identifying the cohort of risk factors affecting patient LoS.…”

Section: Introductionmentioning

confidence: 99%

Towards Predicting Length of Stay and Identification of Cohort Risk Factors Using Self-Attention Based Transformers and Association Mining: Covid-19 as Phenotype

Alam¹,

Ananbeh²,

Malik³

et al. 2023

Preprint

View full text Add to dashboard Cite

Predicting Length of Stay (LoS) and understanding its underlying factors is essential to minimize the risk of hospital-acquired conditions, improve financial, operational, and clinical outcomes, and to better manage future pandemics. The purpose of this study is to forecast patients’ LoS using a deep learning model and analyze cohorts of risk factors minimizing or maximizing LoS. We employed various pre-processing techniques, SMOTE-N to balance data, and Tab-Transformer model to forecast LoS. Finally, Apriori algorithm was applied to analyze cohorts of risk factors influencing LoS at hospital. The Tab-Transformer outperformed the base Machine Learning models with an F1-score (.92), precision (.83), recall (.93), and accuracy (.73) for discharge dataset, and F1-score (.84), precision (.75), recall (.98), and accuracy (.77) for deceased dataset. The association mining algorithm was able to identify significant risk factors/indicators belonging to lab, X-Ray, and clinical data such as elevated LDH, and D-Dimer, lymphocytes count, and comorbidities such as hypertension and diabetes responsible for extending patients LoS. It also reveals what treatments has reduced the symptoms of COVID-19 patients leading to reduction in LoS particularly when no vaccines or medication such as Paxlovid were available.

show abstract

Section: Introductionmentioning

confidence: 99%

Towards Predicting Length of Stay and Identification of Cohort Risk Factors Using Self-Attention Based Transformers and Association Mining: Covid-19 as Phenotype

Alam¹,

Ananbeh²,

Malik³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Zebin & Chaussalet ( Zebin & Chaussalet, 2019 ) proposed to use autoencoder dense neural networks to classify LOS into short (0–7 days) and long stays (>7 days) using the public MIMIC III dataset. Harerimana, Kim, and Jang (2021) proposed a hierarchical attention network to predict LOS and in-hospital mortality. The proposed model was able to leverage the patient anamnesis and free text diagnosis recorded on the first day for prediction purposes.…”

Section: Introductionmentioning

confidence: 99%

“…In other words, the aim is to group the LOS into multiple classes, e.g., short stay, medium stay and long stay, based on the number of days that the patient stays in the hospital. However, several studies have demonstrated that the LOS distributions are highly skewed to the right ( Harerimana et al, 2021 , Ma et al, 2020 ). This skewness indicates that the dataset becomes heavily imbalanced as only a few long LOS cases exist.…”

Section: Introductionmentioning

confidence: 99%

Time-to-event modeling for hospital length of stay prediction for COVID-19 patients

Wen

Rahman

Zhuang

et al. 2022

Machine Learning with Applications

View full text Add to dashboard Cite

Section: Introductionmentioning

confidence: 99%

“…Added to this is the technical challenge of understanding the concepts involved in the survival analysis and also of dealing with unbalanced data, as often occurs in factors such as hospital mortality and length of hospital stay. [11] This study aimed to estimate the expected survival time, and to evaluate the association and impact of different factors for the survival time of hospitalized patients, based on a cohort of patients hospitalized in CCUs.…”

Section: Introductionmentioning

confidence: 99%

Survival analysis of coronary care unit patients from MIMIC IV database

Lanzieri

Palmer

Gismondi

et al. 2022

Preprint

View full text Add to dashboard Cite

The profile of hospitalizations in coronary care units (CCU) includes patients with different age groups, multiple comorbidities and causes of hospitalization that may or may not be primarily cardiac. This study aimed to estimate survival time and evaluate the association and impact of different factors on this time in a cohort of patients admitted to CCU. A cohort of 7120 adult patients admitted to CCU was analyzed from a subset of data from the MIMIC-IV database (Medical Information Mart for Intensive Care, version 4). A descriptive analysis was performed using Kaplan-Meier survival analysis, with a Log-rank test to establish comparisons between groups. Survival regression was modeled using Cox’s proportional risk models for the multiple analysis. The p-value was defined as ¡ 0.05 as statistically significant. In patients who died during hospitalization, there was a higher average age, longer hospital stay, and a higher rate of heart and respiratory rate, all with p ¡ 0.001. Median overall survival was 28 days (95% CI 26-30 days). The survival probability curve presented a higher inclination in the first weeks, reaching a stable value close to 20% at 10 weeks after hospitalization. When Cox’s regression adjusted for age, gender and comorbidities was performed, hyperpotassemia was shown to be an independent risk factor for in-hospital mortality (RR = 1.22, 95% CI: 1.14-1.30) in this group of patients. These results reinforced that the electronic health record may contain, already in the first hours of hospitalization, relevant information to understand the progression of diseases and identify future directions for research. This study is expected to clarify important topics related to the MIMIC-IV database and enable further research using this patient database. Knowledge of the characteristics of the CCU population can allow better management of physical and human hospital resources.

show abstract

A deep attention model to forecast the Length Of Stay and the in-hospital mortality right on admission from ICD codes and demographic data

Cited by 28 publications

References 27 publications

Towards Predicting Length of Stay and Identification of Cohort Risk Factors Using Self-Attention Based Transformers and Association Mining: Covid-19 as Phenotype

Towards Predicting Length of Stay and Identification of Cohort Risk Factors Using Self-Attention Based Transformers and Association Mining: Covid-19 as Phenotype

Time-to-event modeling for hospital length of stay prediction for COVID-19 patients

Survival analysis of coronary care unit patients from MIMIC IV database

Contact Info

Product

Resources

About