Search

Data-driven medical care delivery must always respect patient privacy—a requirement that is not easily met. This issue has impeded improvements to healthcare software and has delayed the long-predicted prevalence of artificial intelligence in healthcare. Until now, it has been very difficult to share data between healthcare organizations, resulting in poor statistical models due to unrepresentative patient cohorts. Synthetic data, i.e., artificial but realistic electronic health records, could overcome the drought that is troubling the healthcare sector. Deep neural network architectures, in particular, have shown an incredible ability to learn from complex data sets and generate large amounts of unseen data points with the same statistical properties as the training data. Here, we present a generative neural network model that can create synthetic health records with realistic timelines. These clinical trajectories are generated on a per-patient basis and are represented as linear-sequence graphs of clinical events over time. We use a variational graph autoencoder (VGAE) to generate synthetic samples from real-world electronic health records. Our approach generates health records not seen in the training data. We show that these artificial patient trajectories are realistic and preserve patient privacy and can therefore support the safe sharing of data across organizations. [ABSTRACT FROM AUTHOR]

Parkinson's disease is a neurodegenerative disorder characterized by several motor symptoms that develop gradually: tremor, bradykinesia, limb rigidity, and gait and balance problems. While there is no cure, levodopa therapy has been shown to mitigate symptoms. A patient on levodopa experiences cycles in the severity of their symptoms, characterized by an ON state—when the drug is active—and an OFF state—when symptoms worsen as the drug wears off. The longitudinal progression of the disease is monitored using episodic assessments performed by trained physicians in the clinic, such as the Unified Parkinson's Disease Rating Scale (UPDRS). Lately, there has been an effort in the field to develop continuous, objective measures of motor symptoms based on wearable sensors and other remote monitoring devices. In this work, we present an effort towards such a solution that uses a single wearable inertial sensor to automatically assess the postural instability and gait disorder (PIGD) of a Parkinson's disease patient. Sensor data was collected from two independent studies of subjects performing the UPDRS test and then used to train and validate a convolutional neural network model. Given the typical limited size of such studies we also employed the use of generative adversarial networks to improve the performance of deep-learning models that usually require larger amounts of data for training. We show that for a 2-min walk test, our method's predicted PIGD scores can be used to identify a patient's ON/OFF states better than a physician evaluated on the same criteria. This result paves the way for more reliable, continuous tracking of Parkinson's disease symptoms. [ABSTRACT FROM AUTHOR]

Clinical risk prediction models powered by electronic health records (EHRs) are becoming increasingly widespread in clinical practice. With suicide-related mortality rates rising in recent years, it is becoming increasingly urgent to understand, predict, and prevent suicidal behavior. Here, we compare the predictive value of structured and unstructured EHR data for predicting suicide risk. We find that Naive Bayes Classifier (NBC) and Random Forest (RF) models trained on structured EHR data perform better than those based on unstructured EHR data. An NBC model trained on both structured and unstructured data yields similar performance (AUC = 0.743) to an NBC model trained on structured data alone (0.742, p = 0.668), while an RF model trained on both data types yields significantly better results (AUC = 0.903) than an RF model trained on structured data alone (0.887, p < 0.001), likely due to the RF model's ability to capture interactions between the two data types. To investigate these interactions, we propose and implement a general framework for identifying specific structured-unstructured feature pairs whose interactions differ between case and non-case cohorts, and thus have the potential to improve predictive performance and increase understanding of clinical risk. We find that such feature pairs tend to capture heterogeneous pairs of general concepts, rather than homogeneous pairs of specific concepts. These findings and this framework can be used to improve current and future EHR-based clinical modeling efforts. [ABSTRACT FROM AUTHOR]