Informatics-based challenges of building collaborative healthcare research and analysis networks from rural community health centers

Abstract

We discuss informatics-based challenges of constructing large-scale collaborative networks for healthcare research and analysis from rural community health centers. These types of networks provide data access and analytic insights across multiple heterogeneous health centers for both healthcare professionals and biomedical researchers. Challenges fall into three general categories: data access, data integration, and technical infrastructure. Data access issues arise in balancing patient privacy, security, and utility; data integration issues persist from each site independently operating its desired electronic medical record; technical infrastructure challenges include creating an analysis and reporting hub capable of scaling across a large collaborative network. Other challenges, such as the difficulty of site recruitment, are important to discuss, but cannot be solved directly through informatics alone. We discuss these challenges and their potential solutions in the context of our implementation of the Kentucky Diabetes and Obesity Collaborative (KDOC). KDOC is a network of Federally-Qualified Community Health Centers (FQHCs) that established a collaborative infrastructure for research and analysis of obesity and diabetes in rural and under-served communities.

Using i2b2 to bootstrap rural health analytics and learning networks

“…Although this is an improvement from an organizational point of view, mapping a common standard remains challenging. In a previous study, we observed laboratory data coded according to internal standards [3] and successfully mapped internal labels to LOINC by using RELMA [15]. …”

“…A known challenge of collaborative networks in rural settings is the lack of information technology infrastructure at each individual site, which ultimately forces centralization at locations with adequate resources [3]. Data from the rural clinics were curated and hosted by its local regional academic research institution; as seen in Fig.…”

“…Interventions driven from electronic health record (EHR) data hold great promise in helping to achieve these goals, yet challenges remain surrounding EHR data standardization, sharing, and access. Data integration has been identified as a key informatics challenge in establishing collaborative healthcare data networks [3]. These informatics networks are constructed between healthcare centers to provide infrastructure for collaboration on many fronts including research, data sharing, analytical insights, and cross-site learning.…”

“…The challenge of connecting multiple independent healthcare sites stems from incompatibilities and inconsistencies in the underlying data model of EHRs. It is highly unlikely that each site in the network uses the same EHR unless a common EHR was developed in advance and in fact, even if the same EHR vendor is used, version incompatibilities exist [3]. Establishing a comprehensive common data model is a necessary first step to building a network that can support queries across its members; transforming data from the source into this common model becomes a secondary challenge that can be alleviated with heavy use of data standards.…”

Using closure tables to enable cross-querying of ontologies in database-driven applications

“…i2b2 is reported to be used by over half of all institutions receiving a Clinical and Translational Science Award (CTSA), over 60 academic medical centers, and over 10 international medical centers [5]. We have demonstrated that i2b2 can bootstrap rural health analytics and learning networks [6] and quickly give infrastructure to institutions that otherwise lack the resources to support clinical and health informatics research [7]. …”