Search

We set out to build a powerful configuration layer for semi-technical analysts and end users on top of relational database technology. This configuration layer empowers data analysts to define data feeds and transformations, data mart schemas, as well as query builder display properties. A data exploration tool that can readily be configured by both analysts and end-users is more likely to meet users’ needs at a lower cost than fully custom software while providing greater data quality and auditability than spreadsheets.

Behavioral and biomedical researchers value data integrity, data exploration, and data sharing, but traditional research databases have unacceptable trade-offs among these important data management functions. This case study examines the design, development, and evaluation of RexMart, an open-source application designed to optimize data integrity, exploration, and sharing with easily accessible querying tools including the powerful query language HTSQL.

Disease- and procedure-specific registries are powerful tools for the improvement of long-term patient care. Unfortunately, registries are typically too expensive for smaller practices and organizations to develop and maintain. This case study examines an affordable and scalable alternative, RexDB®, an open-source solution that can integrate institutional (EHR) and primary research data.

Understanding the scope of research activities in a scientific community requires leveraging data from multiple public and private sources. However, data from different sources, such as PubMed and NIH RePORTER, may be difficult to link because data are organized differently, are of inconsistent or poor quality or lack appropriate identifiers. The standard solution to these challenges is to use an Extract, Transform and Load process (ETL), which pulls the data out of the original data source, changes its structure and content, then stores it in the target database. However, ETL processes tend to be brittle, expensive, and difficult to run. We needed a data-integration solution that would remain flexible and inexpensive, and would support an expanding list of ad hoc data sources to support decisions about scientific funding.

Making decisions about scientific direction, such as funding of research programs, requires a variety of information sources. Science officers require knowledge of multiple internal sources about grants, as well as multiple external sources about grants and publications. Access to a wide range of systems can increase costs and reduce the speed of decisions. Different kinds of tasks (use cases) often result in very different software solutions, such as customized “dashboards”. These attempts usually fail due to long delivery times, measured in months or years. By the time the dashboard is delivered, decisions and data have changed. Our approach is to produce a platform for rapidly configuring decision support dashboards that can be delivered in a matter of days or hours.

The Reusable Online Assessment Data Service (ROADS) is a lightweight and secure tool for publishing complex clinical assessments on the web and for collecting data from several different sources, including self-administered questionnaires from study participants and paper form data collected with dual data entry verification. ROADS makes it simple for researchers to design, maintain, and administer assessments with complex structures and that use a wide variety of question types.

Despite tremendous progress in web-development technologies, building custom data-driven web applications remains out of reach for most research programs. Application development costs too much, takes too long, and requires too many technical skills. A technology platform that significantly speeds and simplifies development would bridge the gap between limited budgets and capabilities and researchers’ need for powerful web tools for interacting with data.

The collaborative, multidisciplinary, and often longitudinal nature of autism research requires a data acquisition platform that is secure, scalable, extensible, and reusable across multiple studies and domains. We evaluated existing systems and found that none fit the needs of complex interdisciplinary or translational projects. No existing platform offered the combination of features that included interactive, user-driven configuration; open-source licensing; and thoughtful support for integrated data management across studies, data types, and sites. To address this deficit we packaged and open-sourced the Research Exchange and Acquisition Platform (RexAcquire), a secure Electronic Data Capture (EDC) system with integrated data management features.

We developed and used two open-source tools, the HyperText Rapid Application Framework (HTRAF) and HyperText Structured Query Language (HTSQL), to rapidly create four custom dashboards for an observational study. We gathered initial requirements, created mock-ups and solicited two rounds of feedback from four users: Clinician, Research Assistant, Clinic Coordinator, and Intake Coordinator. Prototypes were developed in one day, fully functional dashboards in five days. Developers were able to modify dashboards within hours. Qualitative feedback from users revealed that they liked the dashboards and found them easy to use.

We produced a platform for rapidly configuring decision support dashboards that can be delivered in days or hours. The platform was created using an agile methodology over 10 weekly iterations. We developed a data warehouse for storing data about scientific research, added a web-native data-access layer (HTSQL) and a rapid web-application framework (HTRAF). The prototype was developed in two weeks and included a total of 21 different screens covering 5 use cases for usability reviews. The final tool contained 191,000 total records from multiple internal and external grants and publications sources. Specifically, the application included 136,000 NIH records and 39,800 PubMed Publications, 7,800 scientists across 1,300 institutions and 2,400 internal records.