Publications
Length of Stay Prediction With Standardized Hospital Data From Acute and Emergency Care Using a Deep Neural Network
Abstract
Objective: Length of stay (LOS) is an important metric for the organization and scheduling of care activities. This study sought to propose a LOS prediction method based on deep learning using widely available administrative data from acute and emergency care and compare it with other methods. Patients and Methods: All admissions between January 1, 2011 and December 31, 2019, at 6 university hospitals of the Hospices Civils de Lyon metropolis were included, leading to a cohort of 1,140,100 stays of 515,199 patients. Data included demographics, primary and associated diagnoses, medical procedures, the medical unit, the admission type, socio-economic factors, and temporal information. A model based on embeddings and a Feed-Forward Neural Network (FFNN) was developed to provide fine-grained LOS predictions per hospitalization step. Performances were compared with random forest and logistic regression, with the accuracy, Cohen kappa, and a Bland-Altman plot, through a 5-fold cross-validation. Results: The FFNN achieved an accuracy of 0.944 (CI: 0.937, 0.950) and a kappa of 0.943 (CI: 0.935, 0.950). For the same metrics, random forest yielded 0.574 (CI: 0.573, 0.575) and 0.602 (CI: 0.601, 0.603), respectively, and 0.352 (CI: 0.346, 0.358) and 0.414 (CI: 0.408, 0.422) for the logistic regression. The FFNN had a limit of agreement ranging from −2.73 to 2.67, which was better than random forest (−6.72 to 6.83) or logistic regression (−7.60 to 9.20). Conclusion: The FFNN was better at predicting LOS than random forest or logistic regression. Implementing the FFNN model for routine acute care could be useful for improving the quality of patients’ care.
Hospital length of stay prediction tools for all hospital admissions and general medicine populations: systematic review and meta-analysis
Abstract
Background: Unwarranted extended length of stay (LOS) increases the risk of hospital-acquired complications, morbidity, and all-cause mortality and needs to be recognized and addressed proactively. Objective: This systematic review aimed to identify validated prediction variables and methods used in tools that predict the risk of prolonged LOS in all hospital admissions and specifically General Medicine (GenMed) admissions. Method: LOS prediction tools published since 2010 were identified in five major research databases. The main outcomes were model performance metrics, prediction variables, and level of validation. Meta-analysis was completed for validated models. The risk of bias was assessed using the PROBAST checklist. Results: Overall, 25 all admission studies and 14 GenMed studies were identified. Statistical and machine learning methods were used almost equally in both groups. Calibration metrics were reported infrequently, with only 2 of 39 studies performing external validation. Meta-analysis of all admissions validation studies revealed a 95% prediction interval for theta of 0.596 to 0.798 for the area under the curve. Important predictor categories were co-morbidity diagnoses and illness severity risk scores, demographics, and admission characteristics. Overall study quality was deemed low due to poor data processing and analysis reporting. Conclusion: To the best of our knowledge, this is the first systematic review assessing the quality of risk prediction models for hospital LOS in GenMed and all admissions groups. Notably, both machine learning and statistical modeling demonstrated good predictive performance, but models were infrequently externally validated and had poor overall study quality. Moving forward, a focus on quality methods by the adoption of existing guidelines and external validation is needed before clinical application.
Hospital length of stay prediction for general surgery and total knee arthroplasty admissions: Systematic review and meta-analysis of published prediction models
Abstract
Objective Systematic review of length of stay (LOS) prediction models to assess the study methods (including prediction variables), study quality, and performance of predictive models (using area under receiver operating curve (AUROC)) for general surgery populations and total knee arthroplasty (TKA). Method LOS prediction models published since 2010 were identified in five major research databases. The main outcomes were model performance metrics including AUROC, prediction variables, and level of validation. Risk of bias was assessed using the PROBAST checklist. Results Five general surgery studies (15 models) and 10 TKA studies (24 models) were identified. All general surgery and 20 TKA models used statistical approaches; 4 TKA models used machine learning approaches. Risk scores, diagnosis, and procedure types were predominant predictors used. Risk of bias was ranked as moderate in 3/15 and high in 12/15 studies. Discrimination measures were reported in 14/15 and calibration measures in 3/15 studies, with only 4/39 externally validated models (3 general surgery and 1 TKA). Meta-analysis of externally validated models (3 general surgery) suggested the AUROC 95% prediction interval is excellent and ranges between 0.803 and 0.970. Conclusion This is the first systematic review assessing quality of risk prediction models for prolonged LOS in general surgery and TKA groups. We showed that these risk prediction models were infrequently externally validated with poor study quality, typically related to poor reporting. Both machine learning and statistical modelling methods, plus the meta-analysis, showed acceptable to good predictive performance, which are encouraging. Moving forward, a focus on quality methods and external validation is needed before clinical application.
Méthode globale de prédiction des durées de séjours hospitalières avec intégration des données incrémentales et évolutives
Abstract
Prédire la durée de séjour des patients est un enjeu important pour l'organisation des activités de soin dans les hôpitaux, notamment en termes de gestion des lits et de préparation de la sortie des patients. Faciliter l'organisation des activités de l'hôpital influence l'accès, la qualité et l'efficience des soins. Dans cette thèse, nous avons cherché à prédire la durée de séjour pour tous les patients de l'hôpital, à toutes les étapes qui composent leurs parcours de soins, à l'aide de données médico-administratives standardisées de Médecine, Chirurgie, Obstétrique qui sont collectées pour le remboursement des soins. Nous avons commencé par faire une revue systématique de la littérature sur les méthodes de prédiction des durées de séjours, afin de mieux comprendre la préparation des données, les différentes approches de prédiction et la façon de rapporter les résultats. Nous avons ensuite travaillé sur une méthode de prétraitement des données et déterminé si les embeddings peuvent représenter les concepts médicaux dans le cadre des prédictions de durées de séjours via un réseau de neurones. La capacité du réseau de neurones à correctement prédire la durée de séjour a été évaluée et comparée avec celle d'une forêt aléatoire et d'une régression logistique. Nos travaux montrent que la durée de séjour hospitalière peut être prédite au moyen d'un réseau de neurones avec des données médico-administratives standardisées disponibles pour tous les patients.
Predicting length of stay with administrative data from acute and emergency care: an embedding approach
Abstract
Hospital beds management is critical for the quality of patient care, while length of inpatient stay is often estimated empirically by physicians or chief nurses of medical wards. Providing an efficient method for forecasting the length of stay (LOS) is expected to improve resources and discharges planning. Predictions should be accurate and work for as many patients as possible, despite their heterogeneous profiles. In this work, a LOS prediction method based on deep learning and embeddings is developed by using generic hospital administrative data from a French national hospital discharge database, as well as emergency care. Data concerned 497 626 stays of 304 931 patients from 6 hospitals in Lyon, France, from 2011 to 2019. Results of a 5-fold cross-validation showed an accuracy of 0.73 and a kappa score of 0.67 for the embeddings method. This outperformed the baseline which used the raw input features directly.
Hospital Length of Stay Prediction Methods: A Systematic Review
Abstract
Objective: This systematic review sought to establish a picture of length of stay (LOS) prediction methods based on available hospital data and study protocols designed to measure their performance. Materials and Methods: An English literature search was done relative to hospital LOS prediction from 1972 to September 2019 according to the PRISMA guidelines. Articles were retrieved from PubMed, ScienceDirect, and arXiv databases. Information were extracted from the included papers according to a standardized assessment of population setting and study sample, data sources and input variables, LOS prediction methods, validation study design, and performance evaluation metrics. Results: Among 74 selected articles, 98.6% (73/74) used patients’ data to predict LOS; 27.0% (20/74) used temporal data; and 21.6% (16/74) used the data about hospitals. Overall, regressions were the most popular prediction methods (64.9%, 48/74), followed by machine learning (20.3%, 15/74) and deep learning (17.6%, 13/74). Regarding validation design, 35.1% (26/74) did not use a test set, whereas 47.3% (35/74) used a separate test set, and 17.6% (13/74) used cross-validation. The most used performance metrics were R2 (47.3%, 35/74), mean squared (or absolute) error (24.4%, 18/74), and the accuracy (14.9%, 11/74). Over the last decade, machine learning and deep learning methods became more popular (P = 0.016), and test sets and cross-validation got more and more used (P = 0.014). Conclusions: Methods to predict LOS are more and more elaborate and the assessment of their validity is increasingly rigorous. Re-
Doc2Vec on the PubMed corpus: study of a new approach to generate related articles
Abstract
PubMed is the biggest and most used bibliographic database worldwide, hosting more than 26M biomedical publications. One of its useful features is the \"similar articles\" section, allowing the end-user to find scientific articles linked to the consulted document in term of context. The aim of this study is to analyze whether it is possible to replace the statistic model PubMed Related Articles (pmra) with a document embedding method. Doc2Vec algorithm was used to train models allowing to vectorize documents. Six of its parameters were optimised by following a grid-search strategy to train more than 1,900 models. Parameters combination leading to the best accuracy was used to train models on abstracts from the PubMed database. Four evaluations tasks were defined to determine what does or does not influence the proximity between documents for both Doc2Vec and pmra. The two different Doc2Vec architectures have different abilities to link documents about a common context. The terminological indexing, words and stems contents of linked documents are highly similar between pmra and Doc2Vec PV-DBOW architecture. These algorithms are also more likely to bring closer documents having a similar size. In contrary, the manual evaluation shows much better results for the pmra algorithm. While the pmra algorithm links documents by explicitly using terminological indexing in its formula, Doc2Vec does not need a prior indexing. It can infer relations between documents sharing a similar indexing, without any knowledge about them, particularly regarding the PV-DBOW architecture. In contrary, the human evaluation, without any clear agreement between evaluators, implies future studies to better understand this difference between PV-DBOW and pmra algorithm.