*271. Hypertension Therapy Decision Aids: Which Organizing Framework is Considered Most Useful?
KN Simpson, VERDICT Charleston
Objectives: Hypertension is a condition which often co-occur with other risk factors or chronic diseases, such as diabetes. A review of the evidence from controlled trials of reduction in total mortality, cardiovascular disease (CVD) mortality, and CVD events (Mulrow and Pignone, forthcoming) revealed that contingency table summaries by age groups for interventions for patients with hypertension and diabetes, hyperlipidemia, obesity, sedentary lifestyle, and smoking could include as many as 79,200 individual cells. To develop an organizing framework for constructing models to present data on risk, risk reduction, and potential benefits and harms from evidence tables so that these data can be used as a practical decision aid at the time when physicians discuss treatment choices with patients.
Methods: We used a qualitative approach to identify two classes of rules that could be used to systematically reduce the original set of evidence table.
Results: The rules for the model construction are listed here. Our iterative application of case studies indicate that the following is the order of decreasing importance for initial consideration: A1) Presence of especially abnormal levels of a risk factor: yes/no; A2) History of CVD: clinical evidence, positive tests, no CVD history; A3) Largest risk factor standardized for time horizon; B1) Patient's aversion to potential side effects of therapy; B2) Patient judgement of their likelihood of successful adherence to treatment; B3) cost of therapy. However, case studies are simplified versions of real patient encounters, and we do not know if the order preference will be the same one for clinicians using the beta-models for real encounters. This testing is underway.
Conclusions: Evidence-based medical decision making holds a promise for improving outcomes for many patients with complex chronic diseases. However, in areas where the evidence base is dense, it becomes difficult to present this evidence useful and parsimonious manner. Methods for translating evidence-data into practice are in their infancy, and much research need to be done in this area. The rules for organizing data must be examined both under case study, and under real practice conditions before they are imbedded in new decision aids. This poses a problem for the efficient construction of models.
Impact: This has potentially far reaching impacts on the care of patients with CVD risk factors and on the development process used for translating evidence into practice. Our results may be useful for informing work on decision aids in adjacent clinical areas, such as COPD, heath failure and other complex chronic health conditions.