1154 — Integration of the Electronic Health Record and a Clinical Decisions Support System Addressing High-Risk Medications in Older Adults

Fried TR, VA Connecticut; Neihoff K, VA Connecticut; Rajeevan N, VA Connecticut; Charpentier P, Yale School of Medicine; Miller P, VA Connecticut; Goldstein M, Palo Alto VA;

Objectives:
Widespread support exists for electronic exchange of health information between electronic health records (EHRs) and computerized clinical decision support systems (CDSS) to improve patient care. We set out to develop an integrated system to identify older adults at high risk of receiving potentially inappropriate medications (PIMs) and to provide patient-specific recommendations in order to improve medication management.

Methods:
An interdisciplinary team was assembled to design one system to extract data from CPRS in order to identify patients and to collect selected data elements for use in a second system, a web-based CDSS. Our team included expertise in bioinformatics, web and systems design, geriatric medicine, and geriatric pharmacy. To inform development of the CDSS, we conducted systematic reviews of the effects of polypharmacy and of multimorbidity on patient outcomes. We also conducted a Delphi Panel to inform approaches to identifying and correcting problems with medication regimens.

Results:
We developed a novel program to extract CPRS data to identify Veterans with upcoming PCP appointments who have risk factors for receiving PIMs, including: age ? 65 years, multiple medications, and multiple comorbidities. The CDSS integrates two sets of inputs: one from the EHR and one from patient assessment obtained through a patient interview. The CDSS generates a PDF clinician report that provides patient-specific recommendations from evidence and Delphi-based algorithms and is provided to the clinician. The report addresses: medication reconciliation discrepancies; problems with adherence, social support, and cognition; PIMs as identified using Beers and STOPP criteria; potential overtreatment of diabetes and hypertension; inappropriate dosing of renally excreted medications; and patient-reported side effects.

Implications:
We have demonstrated the feasibility of identifying patients at high risk for PIMs coming into primary care and automating the merge of CPRS data with patient assessment data to provide patient-specific medication recommendations to the PCP.

Impacts:
The CDSS was designed to improve the accuracy of medication list, decrease the use of inappropriate medications, and encourage clinician-patient communication around medication concerns. Implementation of the CDSS is in the pilot phase to assess the effect of the tool on medication prescribing and shared decision making around medications.