Predicting Hospital Readmission among Diabetics using Deep Learning

Hammoudeh, Ahmad; Al‐Naymat, Ghazi; Ghannam, Ibrahim; Obied, Nadim

doi:10.1016/j.procs.2018.10.138

Cited by 46 publications

(28 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The proposed model consistently outperforms the other models on both accuracy (1 point of increase) and recall metrics (9 points of increase). The AUC equals the one obtained by [16] with CNN and outperforms the other models. Despite, achieving similar AUC metrics (95%), the CNN model, however, falls short in term of accuracy (92%) while precision and recall metrics weren't available for comparison.…”

Section: E Model Benchmarkingmentioning

confidence: 64%

“…Convolutional Neural Network presents deep learning as an efficient method for predicting hospital readmission of diabetic patients [16]. This model indeed achieves state of the art cstatistic performance of 95% and performs better than other machine learning models.…”

Section: Related Workmentioning

confidence: 91%

“…While under-sampling would be a preferable method, the dataset is not large enough to allow further reduction that could significantly reduce the model performance. Hence, SMOTE can be viewed as an efficient way to overcome imbalanced data as it also increases the sample size [16].…”

Section: Class Imbalancementioning

confidence: 99%

“…Such categories aim at assessing whether a change of medication occurred during the patient's admission. Several studies highlighted medication change as an influential factors towards readmission [14], [16], [24]. Hence, a new feature label "medication_change" will be engineered by counting all changes in medication for all records.…”

Section: ) Feature Engineeringmentioning

confidence: 99%

“…The proposed model will now be benchmarked with recent best performing models from the body of literature. The Convolutional Neural Network (CNN) proposed by [16], Random Forest Gini Algorithm from [14] and RNN developed by [15] will be used for this comparison, see Table 5.…”

Section: E Model Benchmarkingmentioning

confidence: 99%

See 4 more Smart Citations

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

Goudjerkan¹,

Jayabalan²

2019

ijacsa

View full text Add to dashboard Cite

Hospital readmission is considered a key metric in order to assess health center performances. Indeed, readmissions involve different consequences such as the patient's health condition, hospital operational efficiency but also cost burden from a wider perspective. Prediction of 30-day readmission for diabetes patients is therefore of prime importance. The existing models are characterized by their limited prediction power, generalizability and pre-processing. For instance, the benchmarked LACE (Length of stay, Acuity of admission, Charlson comorbidity index and Emergency visits) index traded prediction performance against ease of use for the end user. As such, this study propose a comprehensive pre-processing framework in order to improve the model's performance while exploring and selecting a prominent feature for 30-day unplanned readmission among diabetes patients. In order to deal with readmission prediction, this study will also propose a Multilayer Perceptron (MLP) model on data collected from 130 US hospitals. More specifically, the pre-processing technique includes comprehensive data cleaning, data reduction, and transformation. Random Forest algorithm for feature selection and SMOTE algorithm for data balancing are some example of methods used in the proposed pre-processing framework. The proposed combination of data engineering and MLP abilities was found to outperform existing research when implemented and tested on health center data. The performance of the designed model was found, in this regard, particularly balanced across different metrics of interest with accuracy and Area under the Curve (AUC) of 95% and close to the optimal recall of 99%.

show abstract

Section: E Model Benchmarkingmentioning

confidence: 64%

Section: Related Workmentioning

confidence: 91%

Section: Class Imbalancementioning

confidence: 99%

Section: ) Feature Engineeringmentioning

confidence: 99%

Section: E Model Benchmarkingmentioning

confidence: 99%

See 3 more Smart Citations