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.

Currently, there is very little research aimed at developing medical knowledge extraction tools for major West Slavic languages (Czech, Polish, and Slovak). This project lays the groundwork for a general medical knowledge extraction pipeline, introducing the resource vocabularies available for the respective languages (UMLS resources, ICD-10 translations and national drug databases). It demonstrates the utility of this approach on a case study using a large proprietary corpus of Czech oncology records consisting of more than 40 million words written about more than 4,000 patients. After correlating MedDRA terms found in patients’ records with drugs prescribed to them, significant non-obvious associations were found between selected medical conditions being mentioned and the probability of certain drugs being prescribed over the course of the patient’s treatment, in some cases increasing the probability of prescriptions by over 250%. This direction of research, producing large amounts of annotated data, is a prerequisite for training deep learning models and predictive systems.