Abstract
BackgroundReadmission after discharge from a hospital is disruptive and costly, regardless of the reason. However, it can be particularly problematic for psychiatric patients, so predicting which patients may be readmitted is critically important but also very difficult. Clinical narratives in psychiatric electronic health records (EHRs) span a wide range of topics and vocabulary; therefore, a psychiatric readmission prediction model must begin with a robust and interpretable topic extraction component.ResultsWe designed and evaluated multiple multilayer perceptron and radial basis function neural networks to predict the sentences in a patient’s EHR that are associated with one or more of seven readmission risk factor domains that we identified. In contrast to our baseline cosine similarity model that is based on the methodologies of prior works, our deep learning approaches achieved considerably better F1 scores (0.83 vs 0.66) while also being more scalable and computationally efficient with large volumes of data. Additionally, we found that integrating clinically relevant multiword expressions during preprocessing improves the accuracy of our models and allows for identifying a wider scope of training data in a semi-supervised setting.ConclusionWe created a data pipeline for using document vector similarity metrics to perform topic extraction on psychiatric EHR data in service of our long-term goal of creating a readmission risk classifier. We show results for our topic extraction model and identify additional features we will be incorporating in the future.
Highlights
Readmission after discharge from a hospital is disruptive and costly, regardless of the reason
Starting with the approach taken by McCoy et al [16], who used aggregate cosine similarity scores to compute domain similarity directly from a Term Frequency – Inverse Document Frequency scores (TF-IDF) vector space model, we extend this method by training a suite of three-layer multilayer perceptron (MLP) and radial basis function (RBF) neural networks using a variety of parameters to compare performance
Despite prior research indicating that similar classification tasks to ours are more effectively performed by RBF networks [28,29,30], we find that our MLP model performs marginally better with significantly less computational complexity (i.e. k-means and width calculations)
Summary
Readmission after discharge from a hospital is disruptive and costly, regardless of the reason. Traditional strategies to study readmission risk factors rely on clinical observation and manual retrospective chart review [9, 10]. This approach, benefitting from clinical expertise, does not scale well for large data sets, is effort-intensive, and lacks automation. More robust, and cheaper alternative approach based on Natural Language Processing (NLP) has been developed and met with some success in other medical fields [11] This approach has seldom been applied in psychiatry because of the unique characteristics of psychiatric medical record content
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
