Most approaches to machine learning from electronic health data can only predict a single endpoint. The ability to simultaneously simulate dozens of patient characteristics is a crucial step towards personalized medicine for Alzheimer’s Disease. Here, we use an unsupervised machine learning model called a Conditional Restricted Boltzmann Machine (CRBM) to simulate detailed patient trajectories. We use data comprising 18-month trajectories of 44 clinical variables from 1909 patients with Mild Cognitive Impairment or Alzheimer’s Disease to train a model for personalized forecasting of disease progression. We simulate synthetic patient data including the evolution of each sub-component of cognitive exams, laboratory tests, and their associations with baseline clinical characteristics. Synthetic patient data generated by the CRBM accurately reflect the means, standard deviations, and correlations of each variable over time to the extent that synthetic data cannot be distinguished from actual data by a logistic regression. Moreover, our unsupervised model predicts changes in total ADAS-Cog scores with the same accuracy as specifically trained supervised models, additionally capturing the correlation structure in the components of ADAS-Cog, and identifies sub-components associated with word recall as predictive of progression.

Expanding the use of RWE in regulatory decision making and increasing uses of real-world data by sponsors will fill the gaps that are critically needed for drug development and safety.

The Office of National Coordinator for Health Information Technology final rule implementing the interoperability and information blocking provisions of the 21st Century Cures Act requires support for two SMART (Substitutable Medical Applications, Reusable Technologies) application programming interfaces (APIs) and instantiates Health Level Seven International (HL7) Fast Healthcare Interoperability Resources (FHIR) as a lingua franca for health data. We sought to assess the current state and near-term plans for the SMART/HL7 Bulk FHIR Access API implementation across organizations including electronic health record vendors, cloud vendors, public health contractors, research institutions, payors, FHIR tooling developers, and other purveyors of health information technology platforms. We learned that many organizations not required through regulation to use standardized bulk data are rapidly implementing the API for a wide array of use cases. This may portend an unprecedented level of standardized population-level health data exchange that will support an apps and analytics ecosystem. Feedback from early adopters on the API’s limitations and unsolved problems in the space of population health are highlighted.

This article proposes a comprehensive set of data standards to address the submission of clinical data based on the work of the Clinical Data Interchange Standards Consortium (CDlSC). The present components of the CDISC submission standards are described and related to Food and Drug Administration guidelines on submission as expressed in the 1999 guidance documents on regulatory submissions and the 2005 guidance on the electronic Common Technical Document (eCTD). The relative rdes of the CDlSC Study Data Tabulation Model (SDTM); the Case Report Tabulation Data Description Specification (define.xml), which is based on the CDlSC Operational Data Model (ODM); and the Analysis Dataset Models (ADaM) in meeting submission data requirements are next described. Finally, this article discusses implementation alternatives for sponsors, fiture directions, and the potential benefit of current projects to improve the submission process.

The results of the study indicate that increasing use of standards could translate into improvements in time, costs, and overall approval rates. The study also observed an uptake in the use of eClinical technologies that could potentially create efficiencies and streamline operational processes.