uppose you were offered a medical treatment that would add one full year to your life. You would probably take it, right? But what if an alternative treatment with substantial side effects might add two years of life instead of just one? What if accepting the treatment went against your cultural or religious values, or would end up costing so much that it would prevent you from pursuing other preventive health measures? Or even if you'd heard from Aunt Susie that it hadn't helped Cousin Ethel one bit? Your once-straightforward decision just got a little more complicated, didn't it? But it would be hard to argue that you didn't become more informed as well.
Day after day, people make choices that profoundly affect their health or that of others, all too often without the benefit of information that might aid them in their decision making. Over the last three decades, however, a new breed of scientist has come forward to shed some light on the dimly lit process by which we make clinical choices, from within the individual doctor-patient setting on out to the broad-based policy arena. The health decision scientist's goal is to devise a more reasoned approach for making health care decisions--one that pulls in and quantifies all the available data on multiple factors like risks and benefits, costs and effects. "The philosophy behind decision science is that decisions are being made every day," says Milt Weinstein, the Henry J. Kaiser Professor of Health Policy and Management and Biostatistics at the Harvard School of Public Health. "Doctors make decisions about what treatments to use, drug companies make decisions about what drugs to develop, government agencies make decisions about what to cover for reimbursement. Decision science is focused on those decisions, the consequences of those decisions, the probabilities of different outcomes, and the value citizens attach to those outcomes."
The overall concept of basing decisions on as much information as possible is deceptively simple and not altogether new; decision analysis began appearing on the business school scene as early as the 1940s in the context of management and operations research. But its application to clinical decisions is a more recent development--one that owes much to the pioneering work of Weinstein. A policy analyst trained in decision science at the Kennedy School of Government, Weinstein was recruited to the School by Dean Howard Hiatt in the early 1970s to create the Center for the Evaluation of Clinical Procedures (CECLIP). Working closely with some of Boston's most distinguished physicians and health professionals, Weinstein wrote one of the field's seminal papers on coronary bypass surgery (then only an emerging technology) and co-authored a book on treatment and screening for hypertension. He and Harvey Fineberg, former dean of the School and provost of Harvard, also jointly penned the text, Clinical Decision Analysis, which formally got the field up and running. Essentially quantitative, Weinstein's research set out to incorporate all the factors involved in a particular clinical decision into an all-encompassing mathematical model; this modeling methodology is still at the heart of clinical decision analysis today.
Propelled by Weinstein's work, CECLIP gradually morphed into Center for the Analysis of Health Practices (CAHP), which began to embrace more curriculum development and instruction. Weinstein, himself, had begun teaching a decision analysis course at the School, attracting handfuls of students like Jeffrey Koplan, now the director of the CDC, and Eli Capilouto, current dean of the University of Alabama, Birmingham School of Public Health. "I would venture to say that a good fraction of people who do work in this area have been trained by Milt," notes Peter Neumann, S.D.'93, assistant professor of policy and decision science at the School, "and that's a great tribute to him. Even those who weren't trained by Milt were probably trained by somebody who was trained by him. In many ways, Milt helped invent this field."
But while the fundamentals of decision analysis were becoming more embedded into the School's infrastructure, CAHP was not; the center gradually dissolved, leaving the decision scientists at the School to affiliate themselves with various departments across the institution and without a central hub from which to work. So the situation remained until 1990, when John Graham, professor of policy and decision sciences at the School, founded the Harvard Center for Risk Analysis. Although his own research centered primarily around environmental health and automobile safety issues, Graham tried to draw Weinstein and others working in clinical decision making at the School into his fold. "At first I resisted," recalls Weinstein. "But the attraction was that at the Center for Risk Analysis you had people who brought the same perspective to the world--the idea that decisions have to be made about important health interventions where you have tremendous uncertainty, where value judgements are involved, and there are often different risk reductions and tradeoffs between increasing survival and improving quality of life. The same kinds of issues apply in medical care, in public health, in auto safety, in environmental health."
Weinstein was eventually won over, and the Center's Program for the Economic Evaluation of Medical Technologies (PEEMT) was born--with Weinstein and Neumann at its helm as director and deputy director, respectively. Two other faculty members, Associate Professor Karen Kuntz, S.D.'93, and Assistant Professor Sue Goldie, M.P.H.'97, have since joined their ranks making them, as Weinstein quips, "the four horsemen of decision science."
With an acronym that Weinstein acknowledges comes across better in print than out loud, PEEMT embodies much more than its technical moniker might suggest. For one, the term "medical technologies" is used in the broadest sense, to cover everything from clinical procedures, drugs, and devices to screening tests and health behavior changes, all with the goal of improving health in the population. And economic considerations, like cost-effectiveness, are only one piece--albeit a very important one--in the puzzle of clinical decision analysis that constitutes the program's work. "I think in general if we had unlimited resources and could spend money on any intervention that could provide any amount to benefit to anyone, then that would certainly be the way to go," says Kuntz. "But given that we have limited resources, the next question is how can we best spend our money to optimize the health of the population. That's what cost-effectiveness does. But again it's sort of the first step. On top of this cold, analytical approach, there are ethical issues, distributional issues, political issues that need to be addressed as well in order to bring to the table a fairly reasoned approach for comparing programs and interventions."
This multitude of layers to decision science is what makes PEEMT's research results sometimes difficult to accept. Decision scientists know that people are going to make health care decisions regardless of whether all the information about a public health intervention, clinical procedure, or drug is available to them. So decision analytic methods are not used to get the "right" answer but are instead the best answer possible given all the obtainable data at a specific point in time. Consequently, results are not set in stone nor are they always intuitive, which can make people wary of their validity; add in the four-letter word of "cost" and you sometimes become "like a skunk at a garden party," as Neumann puts it. "The entire discipline of risk and decision science focuses on consequences," notes Goldie. "Our research is rarely driven by a single research question in contrast to the more familiar clinical studies that employ hypothesis testing; rather most of us are involved in pursuing answers to complex policy questions that require simultaneous consideration of multiple factors. Because we're driven by outcomes in the real world, it's always somewhat muddy."
But this down and dirty approach to real-life problems can be part of the appeal of decision analysis too. PEEMT researchers are truly passionate about making a difference, be it communicating their results to the lay public, influencing policy, or enticing new health professionals into the area. The task is not without challenges--securing research funding, ensuring career development for junior faculty, attracting talented students, all in a field that is not historically in the mainstream of public health, is no mean feat. But these decision scientists are optimistic. This new field is quickly gaining structure and tradition, with universal standards to make the results of different studies easier to compare; Neumann has even created a new online database of cost-utility analyses available to both researchers and the public to aid in this task. Government organizations and insurance companies have already incorporated PEEMT results in policies and guidelines on AIDS treatments, cholesterol lowering, and cervical cancer screening. And decision science classes at the School are packed with students eager to understand a new branch of scientific thought--one that broadens the mind, not narrows it, because it involves so many variables, so many outcomes, and so many subjects.
Indeed, PEEMT faculty research covers everything from Alzheimer's disease (Neumann) to sexually transmitted diseases (Goldie) to cancer (Kuntz), just to name a few. And Weinstein? "Well I don't think there's a subject Milt hasn't covered," laughs Neumann. This diversity may explain the melange of individuals who pursue decision science as a career. "It's not a discipline where you're in third grade and say I want to grow up and be a decision scientist. People usually do a bunch of other things for a while, and all of a sudden this just speaks to them. They will often make big changes in their lives to pursue it," remarks Kuntz, who left behind six years of working in a big research lab in California based solely on brief blurb about the decision science track in the School's course catalogue. Neumann pursued a career in economic policy in Washington, D.C., until the decision analysis bug bit him and brought him to Harvard for his doctorate. Goldie, a physician formerly in clinical practice at Yale, tells hair-raising stories of commuting to the School from New Haven for three years to get her master's under Weinstein; she says simply, "I was struck."
So of all the potential clinical and public health problems out there, how do these decision scientists choose what to focus on in their research? For some, it's that issue that affected them way back on their first job; for others, it's the lure of the hot new technology hitting the clinical scene. But, ironically, for some of the PEEMT faculty the topics of study just fell into their laps--which may be for the best. Laughs Kuntz, "Never ask a decision scientist what they're going to have for dinner; all you'll get is, 'What are you going to have?' 'I don't know; what are you going to have?' We seem to be the worst decision makers." Probably because they're too busy helping us be better ones.
page is maintained by Development Communications in the Office of Resource
Updated January 2005
To contact us with suggestions, comments, and questions, please e-mail: email@example.com
Copyright, 2005, President and Fellows of Harvard College