The objective of this project is to develop and implement sophisticated point-of-care EHR-based clinical decision support that (a) identifies and (b) prioritizes all available evidence-based treatment options to reduce a given patient's cardiovascular risk (CVR). After developing the EHR-based decision support intervention, we will test its impact on CVR, the components of CVR, in a group randomized trial that includes 18 primary care clinics, 60 primary care physicians, and 18,000 adults with moderate or high CVR. This approach, if successful, will (a) improve chronic disease outcomes and reduce CVR for about 35% of the U.S. adult population, (b) maximize the clinical return on the massive investments that are increasingly being made in sophisticated outpatient EHR systems, and (c) provide a model for how to use EHR technology support to deliver """"""""personalized medicine"""""""" in primary care settings.
The objective of this project is to develop and implement sophisticated point-of-care EHR-based clinical decision support that (a) identifies and (b) prioritizes all available evidence-based treatment options to reduce a given patient's cardiovascular risk (CVR). The prioritized list of treatment options is provided in different formats to both the primary care physician (PCP) and patient at the time of each office visit made by a patient with moderate to high CVR and sub-optimally controlled and potentially reversible CVR factors. Available evidence-based treatment options are prioritized based on the magnitude of potential CVR reduction of each treatment option. This intervention strategy, referred to as Prioritized Clinical Decision Support (PCS), is specifically designed for widespread use in primary care settings and has the potential to substantially augment current efforts to control CVR in the 35% of American adults with 10-year Framingham CVR of 10% or higher. To assess the ability of the PCS intervention to reduce CVR in adults, we will randomize 18 primary care clinics with 60 primary care physicians (PCPs) and approximately 18,000 eligible adults with baseline Framingham 10-year risk of a major CV event (either heart attack or stroke) of 10% or more into one of two experimental conditions: Group 1 includes 9 clinics (with 30 PCPs and 9,000 patients) that will receive prioritized clinical decision support (PCS) to reduce CVR at the time of each clinical encounter made by an eligible adult. Group 2 includes 9 clinics (with 30 PCPs and 9,000 patients) that receive no study intervention and constitute a usual care control group. The study will formally test the hypothesis that after control for baseline CVR, post- intervention 10-year Framingham CVR will be better in Group 1 than Group 2 at 12 and 24 months after start of the intervention. In addition, impact of the intervention on specific components of CVR (BP, lipids, glucose, aspirin use, and smoking) will be assessed, and the cost-effectiveness of the intervention will be quantified. This innovative project builds upon 10 years of prior work by our research team, and extends prior successful EHR clinical decision support interventions by introducing prioritization, by providing decision support to both patients and PCPs at the time of the office visit, and by extending the decision support across the broad and critically important clinical terrain of CVR reduction. The results of this project, whether positive or negative, will extend our understanding of how to maximize the clinical return on massive public and private sector investments now being made in sophisticated outpatient EHR systems. If successful, this decision support tool could be broadly used to both standardize and personalize care delivered by case managers, pharmacists, and other providers in a wide range of care delivery configurations.
|O'Connor, P J; Sperl-Hillen, J M; Fazio, C J et al. (2016) Outpatient diabetes clinical decision support: current status and future directions. Diabet Med 33:734-41|
|Vock, David M; Wolfson, Julian; Bandyopadhyay, Sunayan et al. (2016) Adapting machine learning techniques to censored time-to-event health record data: A general-purpose approach using inverse probability of censoring weighting. J Biomed Inform 61:119-31|
|Wolfson, Julian; Bandyopadhyay, Sunayan; Elidrisi, Mohamed et al. (2015) A Naive Bayes machine learning approach to risk prediction using censored, time-to-event data. Stat Med 34:2941-57|
|O'Connor, Patrick J; Desai, Jay R; Butler, John C et al. (2013) Current status and future prospects for electronic point-of-care clinical decision support in diabetes care. Curr Diab Rep 13:172-6|