Expanding on the concepts of a clinical decision support (CDS) system to reduce medication errors at BWH and a rudimentary imaging CDS pilot initiated for inpatients in 1991, a web‐enabled CDS program was designed and launched in 1992 (US patent 6,029,138; held by BWH and licensed to Medicalis Corporation in 2000). By embedding CDS in physician workflow, we aimed to reduce inappropriate use of imaging through real‐time evidence‐based decision support at the point of care. By creating CDS‐enabled consequences for ignoring evidence presented in CDS, we devised physician‐led CDS‐enabled targeted interventions to enable, measure and assess adherence to evidence‐based practice, improve quality, and reduce unnecessary imaging and waste.
We named the initiative radiology medical management program3 to emphasize the goal of effectively managing imaging needs of patient populations. Highlighting the impact of our interventions to reduce unnecessary imaging, we were able to negotiate with our local payers (Blue Cross/Blue Shield, Tuft’s, Harvard Pilgrim) to avoid onerous payer pre‐authorization programs for providers who used our radiology medical management program beginning 2005. The integration of an enterprise scheduling module into the workflow of the ordering provider’s practice ensures that the imaging capacity created by the elimination of inappropriate testing as the result of CDS improves access for appropriate imaging services. The value delivered to the patient and the ordering physician through this convenient, web‐enabled, integrated workflow also reduces leakage (i.e. reduces the number of imaging studies referred to radiology providers outside BWH, BWFH and DFCI). A series of peer reviewed publications highlight the impact of this initiative. For example, we have observed a 12% sustained reduction in ambulatory high cost imaging per 1000 member months for a commercial payer population (2005‐2009)3, a 21% reduction in use of CT per severity of disease adjusted inpatient admission (2009‐2012)4, a 33% reduction in use of CT per 1000 ED visits (2007‐2012)5 and a 5%6‐7.5%7 reduction is repeat CT scans. Results of our condition‐specific interventions include a 12.3% reduction in use of MRI for ambulatory patients with low back pain8, a 20% reduction in use of chest CT pulmonary angiography (CTPA) for suspected pulmonary embolism in the ED,9 and a 12% reduction in use of CTPA for inpatients10. We have shown that CDS can be used to unambiguously measure, monitor and enhance provider adherence to evidence11 and national quality measures12. We have described best practices for imaging CDS13 and demonstrated that properly deployed, providers rarely enter erroneous clinical data to avoid onerous CDS interactions (to ‘game the CDS system’)14. We have also shown that the educational effect of CDS, even if based on high quality evidence, is unlikely to optimize practice or to reduce unwarranted variation among providers12,15. CDS‐enabled consequences for ignoring high quality evidence embedded in CDS will likely be needed to enhance the educational effect of CDS8,13. In parallel, the leakage rate for high cost imaging at BWH has plummeted to <1%, compared to 13‐25% leakage at other academic institutions, based on analysis of CMS claims data from the recently completed Medicare Imaging Demonstration. The pioneering work on imaging CDS at CEBI and BWH has helped influence public policy.
For over a century, a leader in patient care, medical education and research, with expertise in virtually every specialty of medicine and surgery.