The aims of the randomized controlled trial (RCT) proposed in this application are to: develop a computerized decision support system (CDSS) that will prompt primary care physicians to implement treatment recommendations based upon AHCPR practice guidelines for depression; compare the clinical outcomes and costs of care for 240 depressed patients; and examine the impacts of disseminating computerized guidelines on physician practices, attitudes, and knowledge. Based on the literature, the application posits that a successful computerized decision support system has the following characteristics: contains evidence-based, expert-panel-derived practice guidelines tailored for local use; uses patient-specific information to customize feedback; presents choices that allow for use of clinical judgment; provides online instruction with the option of accessing more detailed information; enables action at the time of the encounter; and provides brief specific treatment reminders both to patients and physicians. The development of the computerized decision support system for depression is described in this application with examples of screens presented. This system will be placed in the existing electronic medical record, Logician, which is integrated with the Medical ARchival System (MARS). An example of an encounter form from Logician is appended to the application. The primary care setting was chosen for the project since that is where most depression cases are treated. The primary hypothesis is that a higher proportion of depressed patients cared for by physicians using the computerized decision support system will recover within six months of diagnosis compared with patients cared for by physicians not exposed to the complete system. Other hypotheses are that physicians using the system will initiate depression treatment sooner, be more likely to provide treatment consistent with the AHCPR guidelines, have a greater increase in knowledge about the practice guidelines, and achieve similar clinical outcomes at a lower average cost. The variables and covariates of chief interest in the study relate to patient recovery from depression and amounts, types, and costs of treatment. Recovery will be determined by Hamilton Scale ratings (a) equal to or below 7, and (b) that decrease significantly after three and six months of treatment. The costs of treatment (Utilization Questionnaire and MARS) will be based on the Physician Marketplace Statistics and average wholesale prices for drugs. All costs for the six-month period will be summed to ascertain the total cost of care for each patient. These and related data will be analyzed with univariate and multivariate techniques after testing underlying assumptions about their distributions and relationships.
