Search

Background The medical subdomain of a clinical note, such as cardiology or neurology, is useful content-derived metadata for developing machine learning downstream applications. To classify the medical subdomain of a note accurately, we have constructed a machine learning-based natural language processing (NLP) pipeline and developed medical subdomain classifiers based on the content of the note. Methods We constructed the pipeline using the clinical NLP system, clinical Text Analysis and Knowledge Extraction System (cTAKES), the Unified Medical Language System (UMLS) Metathesaurus, Semantic Network, and learning algorithms to extract features from two datasets — clinical notes from Integrating Data for Analysis, Anonymization, and Sharing (iDASH) data repository (n = 431) and Massachusetts General Hospital (MGH) (n = 91,237), and built medical subdomain classifiers with different combinations of data representation methods and supervised learning algorithms. We evaluated the performance of classifiers and their portability across the two datasets. Results The convolutional recurrent neural network with neural word embeddings trained-medical subdomain classifier yielded the best performance measurement on iDASH and MGH datasets with area under receiver operating characteristic curve (AUC) of 0.975 and 0.991, and F1 scores of 0.845 and 0.870, respectively. Considering better clinical interpretability, linear support vector machine-trained medical subdomain classifier using hybrid bag-of-words and clinically relevant UMLS concepts as the feature representation, with term frequency-inverse document frequency (tf-idf)-weighting, outperformed other shallow learning classifiers on iDASH and MGH datasets with AUC of 0.957 and 0.964, and F1 scores of 0.932 and 0.934 respectively. We trained classifiers on one dataset, applied to the other dataset and yielded the threshold of F1 score of 0.7 in classifiers for half of the medical subdomains we studied. Conclusion Our study shows that a supervised learning-based NLP approach is useful to develop medical subdomain classifiers. The deep learning algorithm with distributed word representation yields better performance yet shallow learning algorithms with the word and concept representation achieves comparable performance with better clinical interpretability. Portable classifiers may also be used across datasets from different institutions. Keywords: Medical Decision Making; Computer-assisted; Natural Language Processing; Unified Medical Language System; Machine Learning; Deep Learning; Distributed Representation

The Shared Pathology Informatics Network (SPIN), a research initiative of the National Cancer Institute, will allow for the retrieval of more than 4 million pathology reports and specimens. In this paper, we describe the special query tool as developed for the Indianapolis/Regenstrief SPIN node, integrated into the ever-expanding Indiana Network for Patient care (INPC). This query tool allows for the retrieval of de-identified data sets using complex logic, auto-coded final diagnoses, and intrinsically supports multiple types of statistical analyses. The new SPIN/INPC database represents a new generation of the Regenstrief Medical Record system - a centralized, but federated system of repositories.

Disseminating information from large biomedical databases can be crucial for research. Often this data will be patient-specific, and therefore require that the privacy of the patient be protected. In response to this requirement, HIPAA released regulations for the dissemination of patient data. In many cases, the regulations are so restrictive as to render data useless for many purposes. We propose in this paper a model for obfuscation of data when served to a client application, that will make it extremely unlikely that an individual will be identified. At Partners Healthcare Inc, with over 1.4 million patients and 400 research clinician users, we implemented this model. Based on the results, we believe that a web-client could be made generally available using the proposed data obfuscation scheme that could allow general usage of large biomedical databases of patient information without risk to patient privacy.

With the advent of new web standards such as XML (the extensible markup language), there lies the promise of increased connectivity among informatics applications; such promise can only be realized however if the data models of the information sources are compatible. While work has been done in modeling the logical specification of automated guideline/reminder systems and the clinical data that form their input, little has been done on modeling the output of such systems. A data model for automated guideline output can help foster such compatibility by providing a common frame of reference. We have begun to utilize the “use case” approach by Jacobson et al to develop an object oriented information model of guideline output. We derived use cases from previously published implementations of guideline systems. We present our preliminary results.

This paper describes experience with the development of two systems for clinical knowledge access. Six design principles for implementing such systems in hypertext or related media are offered under the heading “Problem-based Knowledge Access”. These design principles concern: contextual indexing, categories of knowledge access, constraints on hypertext linking, graphical indexing, maintaining the illusion of “no navigation” and problem-based content hierarchies.