question

I wish I had a few striking, earth shattering, and novel solutions to your dilemma of investigating alternate OR processes/schedules, etc.. Unfortunately, where the OR is concerned, I've faced the patient sequencing problem too many times in too many environments to believe that there's a single, much less an array, of solutions to the problem.

Here's my point...

We can't "optimize" an OR schedule. Most of the time, it's an NP Complete problem. There are too many scheduling alternatives, considerations, too many points of view, and too many options for the sequencing of patients. What we can do, however, if the modeling is done correctly and takes into account those elements I mentioned above, perform one very valuable function. We can make every interested party directly or indirectly connected with OR and patient management operations, AWARE of the breadth and impact of any and all management options and decisions brought to bear on OR functions. If you give us a patient stream replete with all of the characteristics normally associated with surgical care (i.e. estimates of case means and variances, estimates of PACU recovery times, estimates of surgical service personnel availability, estimates of OR clean up and set up time, etc.), we can return a clear and unequivocal picture of what any given surgical day might look like from a dozen perspectives that include, waiting times, unplanned delays, manpower and location utilization rates, cost, service rates, time-based census, etc.. In short, whether it's the scheduling clerk who'd like to know if there's any discretionary time left for an additional surgical case, to the comptroller who'd like to know how much unreimbursed time to expect from that days schedule, we can provide an "accurate" estimate of any representative variable they prefer. And, if those estimates don't seem favorable, we can alter the model, the patient stream, the schedule, and any other factor(s) contained in the model, to evaluate the strength and wisdom of any alternation to the process. That's the business we're in.

As a lead modeler in the world of healthcare simulation (my opinion, of course!), I used to think I was in the business of optimizing the way everything worked. In some cases, I might even get close but, in all honesty, that's not the case. Know why? One principle reason is because there's a major, contributing component of all healthcare systems, that's difficult, sometimes impossible in its entirety, to model. Know what that is? Human BEHAVIOR. It's the one ingredient in all human systems that causes those systems to be differentiated from their non-human cousins. In short, a good Healthcare Systems Engineer (HSE), using the latest LEAN and JIT techniques, will always focus on controlling patient care systems so that they, as closely as possible, look as much like either an assembly line or pipeline. They think they're accomplishing that chore by removing as much human discretion and involvement in the conduct of care as possible. Unfortunately, as hard as we work to eliminate the impact of everything from opinions, to human foibles, to founded and unfounded preferences, It's impossible to eliminate every behavior every day. That's why, unlike a "good, "HSE, a GREAT HSE doesn't attempt to eliminate or control behavior, rather they attempt to understand it and incorporate it in every model and action of consequence and purpose. And how do they do that? By staying AWARE.

So, we're not really in the business of optimizing or speeding up anything. Rather, we're in the business of making those who depend on us for realistic and competent recommendations, AWARE of what's going on, what the impact is of changing processes and policies, and what they can expect in the way of returns for the investment of time they devote to management and the goals they're attempting to achieve.

I'm not disappointed by that realization, not at all, even though it took two decades to float to the top of my pool of reason (smiling).

Now, all that remains is to determine how we can take advantage of that realization.

To help along those lines, I've attached a "How-to" paper I assembled two years ago that deals with modeling the OR. At the end of the paper, you'll find a bibliography I assembled that I think you'll find both interesting and useful. If you'd like to investigate the problem further, Let me know ([email protected]) and I'll send you a comprehensive OR demo model Cliff King and I created to demonstrate some ways to set up your model to investigate a few alternative scheduling techniques!