Data Integration for Future Medicine (DIFUTURE)

Big Data in Medicine, the Present and Hopefully the Future

The emergence of data coming from different venues, as several “omic” approaches, is providing already compelling evidence that the smart use of this information could provide invaluable information to prevent, diagnose and treat human diseases. However, the most daunting challenges remain ahead, as the explosive accumulation of data from additional perspectives, including social graphs, biosensors, and imaging, promise to deliver crucial information that could be exploited for the improvement of the entire human race, both in developed, and developing countries, optimizing health expenses and reaching also the less fortunate sections of the societies. And yet, formidable challenges remain, that pertain for the most part to the collection of the data, their organization, and most relevantly their integration. Here we provide few, pointed examples to the present relevance of these big data approaches in human health as well potential road maps toward the implementation of broader data collections and analyses.

KETOS: Clinical decision support and machine learning as a service - A training and deployment platform based on Docker, OMOP-CDM, and FHIR Web Services

Background and objective

To take full advantage of decision support, machine learning, and patient-level prediction models, it is important that models are not only created, but also deployed in a clinical setting. The KETOS platform demonstrated in this work implements a tool for researchers allowing them to perform statistical analyses and deploy resulting models in a secure environment.

Methods

The proposed system uses Docker virtualization to provide researchers with reproducible data analysis and development environments, accessible via Jupyter Notebook, to perform statistical analysis and develop, train and deploy models based on standardized input data. The platform is built in a modular fashion and interfaces with web services using the Health Level 7 (HL7) Fast Healthcare Interoperability Resources (FHIR) standard to access patient data. In our prototypical implementation we use an OMOP common data model (OMOP-CDM) database. The architecture supports the entire research lifecycle from creating a data analysis environment, retrieving data, and training to final deployment in a hospital setting.

Results

We evaluated the platform by establishing and deploying an analysis and end user application for hemoglobin reference intervals within the University Hospital Erlangen. To demonstrate the potential of the system to deploy arbitrary models, we loaded a colorectal cancer dataset into an OMOP database and built machine learning models to predict patient outcomes and made them available via a web service. We demonstrated both the integration with FHIR as well as an example end user application. Finally, we integrated the platform with the open source DataSHIELD architecture to allow for distributed privacy preserving data analysis and training across networks of hospitals.

Conclusion

The KETOS platform takes a novel approach to data analysis, training and deploying decision support models in a hospital or healthcare setting. It does so in a secure and privacy-preserving manner, combining the flexibility of Docker virtualization with the advantages of standardized vocabularies, a widely applied database schema (OMOP-CDM), and a standardized way to exchange medical data (FHIR).

The French Health Data Hub and the German Medical Informatics Initiatives: Two National Projects to Promote Data Sharing in Healthcare

OBJECTIVE

The diversity and volume of health data have been rapidly increasing in recent years. While such big data hold significant promise for accelerating discovery, data use entails many challenges including the need for adequate computational infrastructure and secure processes for data sharing and access. In Europe, two nationwide projects have been launched recently to support these objectives. This paper compares the French Health Data Hub initiative (HDH) to the German Medical Informatics Initiatives (MII).

METHOD

We analysed the projects according to the following criteria: (i) Global approach and ambitions, (ii) Use cases, (iii) Governance and organization, (iv) Technical aspects and interoperability, and (v) Data privacy access/data governance.

RESULTS

The French and German projects share the same objectives but are different in terms of methodologies. The HDH project is based on a top-down approach and focuses on a shared computational infrastructure, providing tools and services to speed projects between data producers and data users. The MII project is based on a bottom-up approach and relies on four consortia including academic hospitals, universities, and private partners.

CONCLUSION

Both projects could benefit from each other. A Franco-German cooperation, extended to other countries of the European Union with similar initiatives, should allow sharing and strengthening efforts in a strategic area where competition from other countries has increased.

BarrettNET-a prospective registry for risk estimation of patients with Barrett's esophagus to progress to adenocarcinoma

Risk stratification in patients with Barrett's esophagus (BE) to prevent the development of esophageal adenocarcinoma (EAC) is an unsolved task. The incidence of EAC and BE is increasing and patients are still at unknown risk. BarrettNET is an ongoing multicenter prospective cohort study initiated to identify and validate molecular and clinical biomarkers that allow a more personalized surveillance strategy for patients with BE. For BarrettNET participants are recruited in 20 study centers throughout Germany, to be followed for progression to dysplasia (low-grade dysplasia or high-grade dysplasia) or EAC for >10 years. The study instruments comprise self-administered epidemiological information (containing data on demographics, lifestyle factors, and health), as well as biological specimens, i.e., blood-based samples, esophageal tissue biopsies, and feces and saliva samples. In follow-up visits according to the individual surveillance plan of the participants, sample collection is repeated. The standardized collection and processing of the specimen guarantee the highest sample quality. Via a mobile accessible database, the documentation of inclusion, epidemiological data, and pathological disease status are recorded subsequently. Currently the BarrettNET registry includes 560 participants (23.1% women and 76.9% men, aged 22-92 years) with a median follow-up of 951 days. Both the design and the size of BarrettNET offer the advantage of answering research questions regarding potential causes of disease progression from BE to EAC. Here all the integrated methods and materials of BarrettNET are presented and reviewed to introduce this valuable German registry.

Data integration of electronic medical record under administrative decentralization of medical insurance and healthcare in China: a case study

In most regions of China, Electronic Medical Record (EMR) systems in hospitals are developed in an uncoordinated manner. Medical Insurance and Healthcare Administration are localised and organizations gather data from a functional management viewpoint without consideration of wider information sharing. Discontinuity of data resources is serious. Despite the government’s repeated emphasis on EMR data integration, little progress has been made, causing inconvenience to patients, but also significantly hindering data mining.

This exploratory investigation used a case study to identify bottlenecks of data integration and proposes countermeasures. Interviews were carried out with 27 practitioners from central and provincial governments, hospitals, and related enterprises in China. This research shows that EMR data collection without patients’ authorization poses a major hazard to data integration. In addition, non-uniform information standards and hospitals’ unwillingness to share data are also significant obstacles to integration. Moreover, friction caused by the administrative decentralization, as well as unsustainability of public finance investment, also hinders the integration of data resources.

To solve these problems, first, a protocol should be adopted for multi-stakeholder participation in data collection. Administrative authorities should then co-establish information standards and a data audit mechanism. Finally, measures are proposed for expanding data integration for multiplying effectiveness and adopting the Public-Private Partnerships model.

Digital networks for laboratory data: potentials, barriers and current initiatives

Medical care is increasingly delivered by multiple providers across healthcare sectors and specialties, leading to a fragmentation of the electronic patient record across organizations and vendor IT systems. The rapid uptake of wearables and connected diagnostic devices adds another source of densely collected data by the patients themselves. Integration of these data sources opens up several potentials: a longitudinal view of laboratory findings would close the gaps between individual provider visits and allow to more closely follow disease progression. Adding non-laboratory data (e.g. diagnoses, procedures) would add context and support clinical interpretation of findings. Case-based reasoning and disease-modelling approaches would allow to identify similar patient groups and classify endotypes. Realization of these potentials is, however, subject to several barriers, including legal and ethical prerequisites of data access, syntactic and semantic integration, comparability of items and user-centered presentation. The German Medical Informatics Initiative is presented as a current undertaking that strives to address these issues by establishing a national infrastructure for the secondary use of routine clinical data.

Health Information Systems - from Present to Future?

This article is part of the Focus Theme of Methods of Information in Medicine on the German Medical Informatics Initiative. Funded by the German Federal Ministry of Education and Research with about 150 million Euro in its currently starting development and networking phase this initiative has already a significant impact on the development of health information systems in Germany. In this Focus Theme two editorials introduce this initiative, one from the viewpoint of its funding institution and one from the initiative's accompanying institutions. Then the initiative's four consortia DIFUTURE (Data Integration for Future Medicine), HiGHmed (Heidelberg-Göttingen-Hannover Medical Informatics), MIRACUM (Medical Informatics in Research and Care in University Medicine), and SMITH (Smart Medical Information Technology for Healthcare) present their concepts and plans. For better readability their manuscripts all contain three major sections on governance and policies, on architectural framework and methodology, and on use cases. As the German Medical Informatics Initiative is a large national experiment, we are convinced that communicating on this initiative already at this early stage to an international audience is of importance.