Live From AMIA: Translating Diabetes Decision Support Tools Into the Safety Net | David Raths | Healthcare Blogs Skip to content Skip to navigation

Live From AMIA: Translating Diabetes Decision Support Tools Into the Safety Net

November 16, 2014
| Reprints
Kaiser’s diabetes quality improvement intervention applied to 11 safety net clinics

Can clinical decision support (CDS) tools and approaches developed over time in a sophisticated integrated health system be adopted for use in community health centers with few IT resources? Researchers in Portland, Ore., say the answer is yes.

I was late arriving for the Sunday sessions at the AMIA Symposium in Washington, D.C., (thanks, Amtrak!), yet the first day was not a disappointment. I saw a great presentation about efforts in Oregon to translate EHR-based CDS tools developed at Kaiser Permanente to 11 Portland-area community clinics in the OCHIN practice-based research network.

Calling it the first study of its kind, Rachel Gold, Ph.D., M.P.H., an investigator in Kaiser’s Center for Health Research, described an implementation in community clinics of an EHR-based CDS strategy developed in Kaiser’s integrated care setting.

Kaiser’s diabetes quality improvement intervention – the “ALL Initiative” uses alerts, panel management data rosters, and order sets. Gold called the initiative one of Kaiser Permanente’s “jewels.” It was highly effective at Kaiser, reducing heart attacks as much as 60 percent in a targeted population over a very few years. The researchers adapted this intervention for implementation in 11 community clinics that share a common Epic EHR database through OCHIN and studied its impact on rates of guideline-based medication prescribing among diabetic patients.

Gold said the study was not as simple as taking Kaiser’s tools and applying them directly to the clinics.  “We asked for them,” she said, “but they are proprietary, so we translated the concept and replicated the strategy, which is to identify patients and prescribe the medications.”

She explained that the translation faced several significant barriers, including resistance from providers. Some said they had learned to ignore automated alerts altogether. The alerts also had to be melded with existing workflow. Additionally, the translated tool addressed only one aspect of diabetes care rather than all aspects of care. Nevertheless, Gold said, the use of CDS tools was associated with significant improvements in rates of guideline-based prescribing. There were two groups of clinics in the study, ones that adopted early and another that adopted later. The percent of “early” clinic patients appropriately prescribed a statin increased from 59 percent to 68 percent in 12 months (a 15 percent relative increase); the percent appropriately prescribed ACE-inhibitors, from 69 percent to 76 percent (a 10 percent relative increase). The “late” clinics showed no concurrent change until their implementation began one year later, but one year post-implementation, the late clinics’ guideline-concordant prescribing rates had improved from 57 percent to 68 percent (statins), and 64 percent to 74 percent (ACE-inhibitors).

I’ve written previous articles about efforts to share clinical decision support solutions between health systems and different EHRs. Here is an exciting example of translating rules and approaches in one EHR and health system into the safety net clinic setting. Perhaps it will lead to other such efforts. 

More reports from AMIA tomorrow!


The Health IT Summits gather 250+ healthcare leaders in cities across the U.S. to present important new insights, collaborate on ideas, and to have a little fun - Find a Summit Near You!


See more on