Phenotyping Hypotensive Patients in Critical Care Using Hospital Discharge Summaries

Efficient goal attainment and engagement in a care manager system using unstructured notes

OBJECTIVE

To improve efficient goal attainment of patients by analyzing the unstructured text in care manager (CM) notes (CMNs). Our task is to determine whether the goal assigned by the CM can be achieved in a timely manner.

MATERIALS AND METHODS

Our data consists of CM structured and unstructured records from a private firm in Orlando, FL. The CM data is based on phone interactions between the CM and the patient. A portion of the data has been manually annotated to indicate engagement. We present 2 machine learning classifiers: an engagement model and a goal attainment model.

RESULTS

We can successfully distinguish automatically between engagement and lack of engagement. Subsequently, incorporating engagement and features from textual information from the unstructured notes significantly improves goal attainment classification.

DISCUSSION

Two key challenges in this task were the time-consuming annotation effort for engagement classification and the limited amount of data for the more difficult goal attainment class (specifically, for people who take a long time to achieve their goals). We successfully explore domain adaptation and transfer learning techniques to improve performance on the under-represented classes. We also explore the value of using features from unstructured notes to improve the model and interpretability.

CONCLUSIONS

Unstructured CMNs can be used to improve accuracy of our classification models for predicting patient self-management goal attainment. This work can be used to help identify patients who may require special attention from CMs to improve engagement in self-management.

Assessing the Readability of Freely Available ICU Notes

Unstructured data stored in an electronic health record (EHR) system can be very informative but require techniques such as natural language processing to extract the information. Developing such techniques requires shared data, but clinical data are often not easy to access. A freely available intensive care unit database, MIMIC-III, was released in 2016 to address this issue and benefit the informatics research community. While the database has been utilized by a few studies, the text characteristics of the notes have not been summarized. In this study, we present the summary of the basic text characteristics and the readability of the MIMIC-III ICU notes. We further compare the results with our previous study where proprietary EHR notes were used. The results show that the text characteristics of MIMIC-III notes were comparable with proprietary EHR notes, although the note readability index was slightly lower. The clinical notes in MIMIC-III can be a viable option for researchers who are interested in clinicians' language use but have no access to proprietary EHR systems.

Estimating Patient's Health State Using Latent Structure Inferred from Clinical Time Series and Text

Modern intensive care units (ICUs) collect large volumes of data in monitoring critically ill patients. Clinicians in the ICUs face the challenge of interpreting large volumes of high-dimensional data to diagnose and treat patients. In this work, we explore the use of Hierarchical Dirichlet Processes (HDP) as a Bayesian nonparametric framework to infer patients' states of health by combining multiple sources of data. In particular, we employ HDP to combine clinical time series and text from the nursing progress notes in a probabilistic topic modeling framework for patient risk stratification. Given a patient cohort, we use HDP to infer latent "topics" shared across multimodal patient data from the entire cohort. Each topic is modeled as a multinomial distribution over a vocabulary of codewords, defined over heterogeneous data sources. We evaluate the clinical utility of the learned topic structure using the first 24-hour ICU data from over 17,000 adult patients in the MIMIC-II database to estimate patients' risks of in-hospital mortality. Our results demonstrate that our approach provides a viable framework for combining different data modalities to model patient's states of health, and can potentially be used to generate alerts to identify patients at high risk of hospital mortality.