constraint theory and pareto optimals linkedin…

Brian Gregory, MD, MBA

Constraint theory and Pareto optimals…

We’ve been using Constraint theory (TOC) as a way to derive Pareto optimal solutions in the OR for a couple of years now, and briefly mentioned the approach at a recent Academy of Business Research presentation we gave. Has anyone taken this approach elsewhere in hospitals? Examples, please.

11 days ago

Wayne Fischer • Sounds intriguing, Brian – but can you give us an explanation of what/how you do “Pareto optimal solutions using ToC?”

11 days ago• Like
Brian Gregory, MD, MBA

Brian Gregory, MD, MBA • Hi Wayne,
It’s a nice way to solve some problems. I was hoping to get more of a response to see if someone else uses a variant of the approach–and come up with some type of algorithm that could be applied anywhere.

If I’m the only one doing this, then I’ll just let the topic drop. I’m here for the feedback. 🙂

10 days ago

Wayne Fischer • Brian, you did it again: you did not answer my question / share your knowledge. Part of the “obligation” of each of us who hope to learn from others is sharing *our* knowledge and experiences…

10 days ago• Like

Robert Gordon • Brian, I have not heard of using Pareto optimality with TOC. But P.O. is a subtle concept or set of such, about marginal efficiency trade-offs. TOC has several sets of principles and tools. The only ways I can imagine them intersecting is in managing buffers or in some part of Vanishing Clouds. It might help people to respond if you gave a little more info. Or, maybe you mean “If you are doing it, you know what I mean.” That’s fair too. Regards,

10 days ago• Like
Brian Gregory, MD, MBA

Brian Gregory, MD, MBA • Ah….peer pressure.

ToC helps clarify the pay-off trade-offs when one agent’s ‘take’ is maximized to the exclusion of the others (subjugation, etc). Then change the chosen constraint (the agent), alter the activities (actions) that are needed to maximize throughput (defined by that agent to be the same as a prior agent>) to get your baselines.

You now have a list of baseline maximum for the throughput of each agent (surgeon, anesthesia, nurse, risk management) and a list of all activities and actions needed to achieve each. This gives you the information you need to do simulations that alter the actions that directly affect each maximum and lets you control (achieve Pareto optimals if you want) the throughput for each.

So…if you are doing it, you’ll understand what I just said.
I’d like to see another approach…if someone has one.

10 days ago

Wayne Fischer • Wait a mo…wait a mo’…

I thought ToC was all about maximizing *system* throughput, by identifying each constraint in turn and taking the appropriate “corrective” action for each – until the *system* is optimized.

What you describe sounds like sub-optimization of the system by optimizing each component (constraint) in turn – and comparing the results…???

9 days ago• Like

Thomas Jones • We’ve used ToC to identify throughput constraints in Surgery but I’m weak on the use of pareto optimals, is there a published version of your presentation somewhere?

8 days ago• Like
Brian Gregory, MD, MBA

Brian Gregory, MD, MBA • Sorry Thomas, no published version yet. It’s being revised, after seeing what people tended to focus on during the presentation (surprising to me was the interest in the ‘buy-in’ portion of the presentation).

I’m more into the nuts and bolts techniques of a methodical approach to figuring out the actions/activities that effect the constraints of one group versus others and the resulting tradeoffs when doing so. Swapping constraints (spontaneously and easily when needed) is also a good concept and ability to have when trying to maximize economic goals.

There are caveats to using ToC in surgery as I mentioned previously: you need to be aware that unless you’re a military hospital (or similar) with a captive work force of surgeons, anesthesiologists, and nurses…you are usually striking a balance between several constraints because of the different goals of those different groups.

You have to look for the pareto optimals (at the very least) to be sure that you aren’t overly subjugating several groups for the benefit of a single group’s constraint.

If the focus is on economic gain, then you also need to be aware that achieving a maximal total economic return may be through subjugating the direct gain of one or more groups…and that some form of indirect compensation (don’t know the legalities of this) would need to be given to achieve some pareto optimal solutions [guaranteed income for anesthesiologists is an example of this type of indirect compensation].

Lots of gaming can be done with this…as long as you have a methodical way of measuring the rewards and tradeoffs to each group–which is the focus of the paper (and the clarity to all parties which is why it helps buy-in).

Hospitals have been been trying to do this juggling for years, but haven’t had a particularly good way to maximize, quantify, and clarify these interactions–and hence bad ‘buy-in’. I suppose you could also think of this as a means to achieve transparency for all parties.

8 days ago

Robert Gordon • Brian, I know you are soliciting participation from people who might already be into this, so please excuse another intrusion or ignore it.

If I understand what you are doing, it is already amazingly complex (even if phenomenologically simplified by relying on human judgement for parameter adjustments, as I guess would be unavoidable at this point in the development of your algorithms and software). It seems to be a kind of dynamic linear programming for allocation solutions subject to “constraints” (in this sense) that are in a constant state of change responsive both to quasi-stable participant agreements (with negotiated maxs and mins) and to feedback-related system structural change [forecasting through Bayesian chains?]. To a non-technical person like myself, this is mind-blowing.

I wonder whether I am even asking the right kind of question: Are you operating this in real time (hourly, daily, weekly) or is it an intellectual understanding of the principles that govern OR scheduling decisions in the longer term?

Regards

8 days ago• Like
Brian Gregory, MD, MBA

Brian Gregory, MD, MBA • Robert, my software lets me do it in real time (minute by minute if needed) for the OR. With the right graphics, it’s more like playing a game where you fit the pieces together. Of course, you can use if for trying out all types of scenarios for longer periods of time, too.

The more operations management concepts you understand, and the better you know the people (surgeons, nurses, anesthetists, floor) and other resources at your facility, the better you do with the game.

Think ‘World of ORcraft’, but much simpler to play. It’s more like ‘Tetrix’ with gantt charts and bar graphs. Tough to explain, but if you can play an online game, then this would be easy.

That’s how I implement constraints and pareto optimals in my situation for the OR. I’m still curious as to what others have done (if they’ve done anything).

8 days ago
Brian Gregory, MD, MBA

Brian Gregory, MD, MBA • Also, as with any good program…there are a lot of concepts embedded in the code. But one does not need software to begin implementing constraints and pareto optimals.

8 days ago

Thomas Jones • I can see why most folks focused on buy-in. The tradeoffs in “swapping
constraints” as you put it are hotly contested with lots of emotional and
cultural issues making change difficult

We are not a military hospital (or similar) with a captive work force of
surgeons, anesthesiologists, and nurses. We are usually striking a balance
between several constraints as you say because of the different goals of
different groups and overly subjugating several groups for the benefit of a
single group’s constraint.

That is the surgeon’s. Historically this facility has bent over backward to
facilitate the surgeon’s timetable and in the process making the hospital
staff waste time in numerous ways. We need to work toward a Pareto optimal
solution that allows the hospital to reduce expenses.

Sadly the economy and regulatory changes are driving the change, I would
have preferred a more altruistic driver, but whatever does the job. Right?

4 days ago• Like
Brian Gregory, MD, MBA

Brian Gregory, MD, MBA • Thomas, in the case of the OR (and probably many other scenarios) the goals, rewards, and trade-offs are often subtle (in the sense of not realized, but quite significant), but once transparent (necessitates seeing and understanding the significance) can be used for negotiating and buy-in by the parties.

In my experience in the OR, people often viciously fight for something that does not give them what they truly want. One of the skills of a good negotiator is to dissect and clarify the goals of competing parties so that the parties, themselves, help come up with solutions (buy-in).

You (as a transparent clarifier and negotiator) may need to discover what they ‘truly’ want, instead of what they say they want. Of course, some people want to be dictator of the world…and there is no real solution to that problem other than excluding them from your society (medical staff, hospital, government, etc).

Unfortunately, in the case of hospitals, administration often doesn’t realize how disruptive and detrimental particular individuals can be. It reminds me of a movie where the bad guy won’t let go of the gold…which causes him to fall off a cliff to his death. An administrator may not realize the true cost and rewards of individuals. Not sure how to deal with that…other than to get a more enlightened administrator or change facilities.

4 days ago
Brian Gregory, MD, MBA

Brian Gregory, MD, MBA • “Not sure how to deal with that…other than to get a more enlightened administrator or change facilities.”

In the case of OR scheduling, you can improve the outcome for surgeons, anesthesia, and nurses…hence a pareto optimal. But it takes a keen awareness of intra-case and inter-case structure (model) and software that makes it easy to schedule on a large scale.

The unusual benefit to this is that you can dramatically improve the goals for a surgeon through an active process, and (if needed for purposes of negotiating with a surgeon) take away that improvement by passive non-strategic scheduling (the current norm).

The more you understand the process, the more you can negotiate.

About Brian D Gregory MD, MBA

Board Certified Anesthesiologist for 30 years. TOC design and implement for 30 years. MBA from U of Georgia '90: Finance, Data Management, Risk Management. Practiced in multiple US states and Saudi Arabia at KFSH&RC and KFMC Taught residents in two locations. Worked with CRNAs for 20 years.
This entry was posted in Uncategorized and tagged , , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s