Clinical Decision Making Fellowship

Fellows support a clinical decision making consultation service which examines complex management dilemmas in individual patients and provides advice in framing complex or unusual problems, particular ones that strongly depend on patients' preferences or where tradeoffs (e.g., length of life vs quality of life). Recommendations are based on literature review and formal decision analysis. Typical consults involve patients with a clinical indication for anticoagulation but with risk factors for bleeding complication or multiple co-morbidities facing possible surgery.

Fellows develop skills in computerized decision analysis in medical decision making, including decision tree construction, Markov model development, Monte Carlo simulation, Bayesian interpretation of diagnostic tests, the measurement of patient preferences, cost-effectiveness analysis, literature review, evidence-based medicine and meta-analysis.

These skills enable fellows to pursue clinical research questions examining the thorough and efficient evaluation of diagnostic possibilities, the value or information content of the medical history, physical examination and diagnostic tests, determination of the optimal diagnostic test or test sequence, selection of the optimal therapy, and evaluation of new medical technologies (tests, devices or drugs).

On a policy level, fellows also become involved with health policy and cost-effectiveness analysis. Previous such projects have ranged from coronary artery revascularization, anticoagulation for atrial fibrillation or prosthetic heart valves, postmenopausal hormone replacement therapy, screening for cancer, treatment for hepatitis, and prenatal genetic testing.

Our Division has been involved with numerous consensus conferences and guideline development processes for the National Institutes of Health (NIH), American College of Chest Physicians (ACCP), American College of Physicians (ACP), and Agency for Health Care Policy and Research (AHCPR) sponsored Patient Outcome Research Teams (PORTs) on Chronic Ischemic Heart Disease (IHD) and Diabetes Mellitus. Most recently, we are a part of the New England Medical Center AHCPR Evidence-based Practice Center awarded to Dr Joseph Lau.

A typical decision analysis involves seven steps:

1. Defining the clinical question
2. Specifying the alternative strategies
3. Defining the outcomes that might result from each alternative strategy
4. Estimating the likelihood of each outcome
5. Determining the value of each outcome (e.g., life expectancy or quality-adjusted life expectancy or costs)
6. Folding back and averaging out to estimate the average expected value of each strategy
7. Performing sensitivity analysis by varying each parameter used in the analysis over a broad range to determine the robustness of the base case decision results.

Each component of the analysis is based on the synthesis of the best available data in the literature coupled with an understanding of the underlying pathophysiology and current clinical practice patterns.

Each component (probability or mortality rate) in the decision analysis requires a thorough examination of the medical literature, examining study populations and outcomes and pooling data when appropriate. These literature reviews form the basis for creating decision trees (simple trees, Markov models or Monte Carlo Simulations), estimating probabilities, estimating life expectancy, performing utility and functional status assessments, defining test characteristics, performing meta-analysis or ROC analysis, and determining the optimal operating point on the ROC curve.

Beyond research, the Division is passionate about teaching and training the next generation of physicians and researchers in medical informatics, evidence-based medicine, and decision analysis. Funded by a National Library of Medicine supported medical informatics training center from 1980 to 2012, the Division of Clinical Decision Making trained over 70 fellows in decision sciences and health outcomes research. Trainees have also received PhD, MSc, MPH, and SM degrees, as well as grant awards. In addition, another 50 postgraduate physicians including a Fulbright Scholar have visited the Division to receive training. Since 2012, the Division has continued to accept physicians and researchers as visiting scholars and fellows. Many of these fellows and visiting physicians have moved on to leadership positions in informatics, policy making or education.

Frequent collaborators include Harry Selker, MD, Deat of the Tufts Clinical Translational Science Institute; Andrew Levey, MD and Klemens Meyer, MD, Division of Nephrology; David Snydman, MD, Division of Infectious Disease and Geographic Medicine; Peter Neumann, ScD and Joshua Cohen, PhD, Center for the Evaluation of Value and Risk in Health, Laurel Leslie, MD, MPH, Director of Community Engagement at Tufts Clinical and Translational Science Institute, and Deeb Salem, MD, Chair, Department of Medicine.  In addition, our Division collaborates with the Institute for Clinical Research and Health Policy Studies and in particular its Center for Clinical Evidence Synthesis, Center for the Evaluation of Value and Risk in Health (CEVR) and Research Design Center/Biostatistics Research Center.

The Division's expertise includes the use of clinical decision analysis, cost-effectiveness analysis, patient preference or utility assessment for quality of life, literature synthesis and meta-analysis, medical informatics. At the population level, it has been involved with technology assessment, guideline development, health outcome analysis, consensus conferences, expert panels, clinical informatics, clinical decision support, quality of care assessment, performance measures, decisision aids, shared decision making, theory of constraints, and telemedicine. Members of the division have participated in translating evidence into quality improvement and quality performance measures for consideration and adoption by the National Quality Forum.

The Division uses the following techniques for computerized decision analyses and simulation in medical decision making: decision tree construction, Markov model development, Monte Carlo simulation, Bayesian interpretation of diagnostic tests, the measurement of patient preferences, cost-effectiveness analysis, literature review, meta-analysis and discrete event simulation.

Many of our former fellows have moved on to leadership positions in informatics, policy-making, teaching or their clinical subspecialty. Most former trainees are general internists, but the fellowship complements training in other clinical specialties. Thus, former trainees include physicians in cardiology, nephrology, infectious diseases, hepatology, gastroenterology, hematology, oncology, pediatrics, cardiothoracic surgery, pathology, radiology, psychiatry and neurology. View a list of our alumni and where they have gone in their careers.