Deep Learning Analysis of Polish Electronic Health Records for Diagnosis Prediction in Patients with Cardiovascular Diseases

Abstract

Electronic health records naturally contain most of the medical information in the form of doctor’s notes as unstructured or semi-structured texts. Current deep learning text analysis approaches allow researchers to reveal the inner semantics of text information and even identify hidden consequences that can offer extra decision support to doctors. In the presented article, we offer a new automated analysis of Polish summary texts of patient hospitalizations. The presented models were found to be able to predict the final diagnosis with almost 70% accuracy based just on the patient’s medical history (only 132 words on average), with possible accuracy increases when adding further sentences from hospitalization results; even one sentence was found to improve the results by 4%, and the best accuracy of 78% was achieved with five extra sentences. In addition to detailed descriptions of the data and methodology, we present an evaluation of the analysis using more than 50,000 Polish cardiology patient texts and dive into a detailed error analysis of the approach. The results indicate that the deep analysis of just the medical history summary can suggest the direction of diagnosis with a high probability that can be further increased just by supplementing the records with further examination results.

Artificial Intelligence–Enabled Analysis of Statin-Related Topics and Sentiments on Social Media

“…To assess the sentiments for each post, we used a pretrained BERT model, called RoBERTa, that was trained on social media posts . RoBERTa offers multiclass labels (ie, positive, neutral, or negative classification of text) and has been used in recent studies investigating health care problems using data from social media . To quantify how sentiments varied across topics and groups, we transformed sentiment labels to scores: from negative to −1, neutral to 0, and positive to 1 .…”

“… 21 RoBERTa offers multiclass labels (ie, positive, neutral, or negative classification of text) and has been used in recent studies investigating health care problems using data from social media. 28 , 29 , 30 , 31 To quantify how sentiments varied across topics and groups, we transformed sentiment labels to scores: from negative to −1, neutral to 0, and positive to 1. 32 Full details on algorithm and model choice, input data handling, and output transformation are provided in the eMethods in Supplement 1 .…”

Year 2022 in Medical Natural Language Processing: Availability of Language Models as a Step in the Democratization of NLP in the Biomedical Area

“…• Malayalam: generation of synoptic clinical reports [39]; • Polish: prediction of cardiovascular diseases in electronic health records [40]; • (Brazilian) Portuguese: description of an annotated clinical corpus [41], ICD-10 coding [42]; • Serbian: sentiment analysis in COVID-19 tweets [43]; • Spanish: ICD-coding [10, 44], negation and uncertainty detection in clinical narratives [45], training and evaluation of word embeddings for the clinical domain [46]; • Swedish: ICD-10 coding [44];…”

“…and institutions (like MIMIC-III), as well as data from social media, hospitals, bibliographical datasets, clinical trials, etc. The research in other languages is possible mainly thanks to the availability of data from social media [7,9,19,20,22,38,43,47] and documents from local hospitals [10,13,14,17,18,23,25,27,36,37,40,42]. Besides, this set of works in languages other than English relies on the dedicated language models, which cover a great variety of languages by now.…”

“…Information 40,87,88,89], mortality [90], risks [91], and patient outcomes [89,92,93]. IE and prediction tasks will certainly continue to perform on clinical and health data in the next years.…”