Leadership
e-Newsletter
Intelligence Unit Special Reports Special Events Subscribe Sponsored Departments Follow Us

Twitter Facebook LinkedIn RSS

Forget Predictive Modeling, Try Leadership Forecasting

Philip Betbeze, for HealthLeaders Media, January 3, 2014

"Predicting is trend analysis," he says. "You have databases, you look at past data, find a mathematical equation that fits it, and use that to extrapolate into the future. You look at how the relationships have worked in the past and you come up with your prediction."

But data from the past quickly becomes obsolete in the rapidly changing healthcare business environment. From changes in reimbursement protocols to the advent of the Obamacare exchanges to a multitude of other drastic changes in how healthcare services are paid for and evaluated, predicting has lost its efficacy, Bauer argues.

"If the underlying circumstances that created that data have changed, then extrapolating from that data gives you stupid answers," he says.

Instead of attempting to predict outcomes based on old data, healthcare leaders should rely more heavily on forecasting, which is less precise but a more accurate gauge of how a particular innovation in care protocols or basic business strategy is likely to pay off.

Using forecasting is more helpful in both healthcare transformation as well as with the weather because there are too many variables involved to accurately predict an outcome.

"Weather is literally unpredictable, so you need a different way to look at it, which is by estimating probabilities and possibilities and weighing them," he says.

A Realm of Possibilities
The forecaster approaches the future as a realm of possibilities that can go any number of different ways. That results in a less specific glance at a possible future, but a much more accurate one, that will prevent unwise decision-making.

1 | 2 | 3 | 4 | 5

Comments are moderated. Please be patient.

4 comments on "Forget Predictive Modeling, Try Leadership Forecasting"


Jeffrey C. Bauer, Ph.D. (1/14/2014 at 4:53 PM)
While I appreciate Dr. Fischer's interest in my book, I respectfully disagree with his observations. He could not have made several of his comments if he had read the book first. For example, Upgrading Leadership's Crystal Ball clearly differentiates forecasting from scenario planning. Forecasting is a specific tool for estimating the realm of future possibilities; scenario planning focuses effort on preparing for a limited number of possible outcomes generally based on predictions, not forecasts. The last chapter provides considerable information about the strategic implications of this difference. Contrary to Dr. Fischer's assumption, my book also addresses the different circumstances under which predicting and forecasting are relevant in health care today. Mr. Betbeze's interview accurately captures my view that health care leaders are more likely to steer their organizations in better directions when they use possibility-driven forecasts [INVALID] not data-driven predictions [INVALID] to envision opportunities like vertically integrated, multi-stakeholder partnerships rather than predictable, horizontal networks of physicians and hospitals that have generally failed to solve the problems of cost and quality. Upon reading the book, Dr. Fischer will find extensive discussions of the modeling methodologies he suggests I should "learn about." I've been teaching and applying them for 40+ years and come to the conclusion that forecasting best fits the new and unprecedented realities of the 21st century.

Wayne G. Fischer, PhD (1/6/2014 at 12:36 PM)
What the quoted "expert" is describing is called scenario planning - it is a well-known and well-documented methodology. A classic case was Royal Dutch Shell's use of it to capitalize on the major disruption of crude oil prices back in the 1970s by OPEC. Either the author or the quoted expert misuse / confuse precision and accuracy, causing this reader to wonder if they really have any meaningful grasp of the subject matter. And the fact that if a system fundamentally changes, a model based on data before that change is no longer valid is a first principal of which anyone who has done any modeling is well aware. "Prediction," contrary to the quoted expert's implication, has a very important role to play in healthcare - operationally, tactically and strategically...because there are *many* types of modeling and prediction methodologies. I suggest the author and quoted expert learn about a few of them.

Frank Poggio (1/6/2014 at 11:11 AM)
Great piece, excellent points. Particularly liked his comment; "If the underlying circumstances that created that data have changed, then extrapolating from that data gives you stupid answers". In medicine the baseline and historical data are changing all the time. Medical protocols can get 'outdated' in a matter of months. Many of Mr. Bauer's points predict, as I do, that Big Data and Analytics will be the next HIT sink hole. Big Data and Analytics – The Next HIT Boondoggle! See: http://www.kelzongroup.com/Big_Data.html Frank Poggio The Kelzon Group