2011 National Meeting

1021 — Investigating Integration of Computerized Decision Support into Workflow at Three Benchmark Institutions

Doebbeling BN (R.L. Roudebush VA Medical Center) , Saleem JJ (R.L. Roudebush VA Medical Center), Haggstrom DA (R.L. Roudebush VA Medical Center), Militello LG (Applied Decision Science), Flanagan M (Indiana University), Kiess CL (Regenstrief Institute), Arbuckle N (University of Dayton Research Institute)

Integration of clinical decision support (CDS) into clinical workflow is a key factor influencing uptake and sustainability. Novel approaches to integrate CDS into workflow are needed if cognitive support is to reach its potential. To better understand the design of CDS and its integration into workflow, we studied the use of CDS for colorectal cancer (CRC) screening and follow-up at benchmark institutions.

We selected clinics affiliated with 4 medical centers (2 VAMCs) in three health systems, considered benchmark organizations in the use of HIT. We conducted rapid ethnographic observations, opportunistic interviews, key informant interviews, and focus groups to identify effective practices in integrating CDS into clinical workflow. All data were analyzed using a coding template based on the sociotechnical model. The coding template was modified as coding proceeded, and themes emerged from the data. Findings were integrated across sites into meaningful patterns, and the data were abstracted into emergent themes.

We found very different forms of EHRs and CDS across sites. Despite design differences, there were common generalizable barriers across sites. These included: 1) receiving and documenting “outside” exam results; 2) colorectal cancer screening CDS was at times not timely; 3) compliance issues (following through with testing); 4) poor EHR or CDS usability; 5) lack of coordination between primary care and the gastrointestinal specialists; and 6) the delivery of acute vs. preventive care. A comparison of these themes with the type of CDS typically used at the site supported the validity of the coding.

Our study demonstrates the importance of identifying effective strategies in the design, implementation and integration of CDS into workflow. Despite working in several different health systems, we found the barriers quite consistent. New CDS prototypes are needed which: 1) improve data organization and presentation, 2) integrate outside results, and 3) provide just-in-time education and cognitive support.

These findings demonstrate that effective design and integration of new technologies requires mindful iteration. Designing and testing prototypes based on these features may help inform development of the next generation of cognitive support for decision making.