Moving Complex Systems Toward a Goal
I recently told several of my Mayo colleagues that I'm an outsider who has landed in a new environment where the degree of collaboration, dialog, and friendly debate is amazing. The following a kind of guest blog, started by Dr. Michael Joyner, who is also a writer
He wrote to a number of us
"Great paper written by chairman of bank of England in middle of financial crisis. Pertinent to anyone interested in nudging complex systems towards a goal."
Below are the responses:
Let’s move away from risk prediction a bit and apply these concepts to our world. The world of innovating and translating new solutions (hugely complex and risky) in the medical care industry (more complex and regulated than even the financial industry) to treat human disease (much of it psychological/societal/evolutionary) is off the charts with risk.
I love the statement of fighting complexity with simplicity – it intrinsically feels correct - so what should our heuristics be for advancing medical care?
Modern medical innovation is complex, perhaps too complex. Regulation of modern translation of medical innovation to products is complex, almost certainly too complex. That configuration spells trouble. As you do not fight fire with fire, you do not fight complexity
with complexity. Because complexity generates uncertainty, not risk, it requires a regulatory response plan to enable innovation and translation grounded in simplicity, not complexity.
So what is our plan and our heuristics? My suggestion to get us started in honor of Mr. Haldane:
1) You must move out of the harbor and get in the race in order to win (win being to develop new solutions in medical care)
a. The race has been underway for some time, so you are behind already, so get going and move with speed
b. A race is the best analogy – we should move quickly in just about every respect. “Fail quickly” is a trite but true statement
c. How many things are stuck in endless cycles of planning or less than serious (not capitalized or staffed anywhere near the need) attempts?
2) Almost all of your strategies, plans, predictions will turn out to be incorrect along the way, no matter how much planning you do – you will assuredly have many failures as you progress - how you respond and your ability recalibrate course and keep progressing to the goal is all that matters in terms of achieving success
a. The good news is at this point you are at least in the race
b. As with the best entrepreneurs, if you have not failed at least a few times you should not be trusted with anything – very easy to say and terribly hard to live it when you are deploying capital
c. The art in all this is determining if the failure is due to execution, if it is then you have the wrong people and should switch out the team – or vision/goal, if it is then time to go back to the harbor. The first determination (execution) is not too hard to make but hard to execute (needing to switch out the team).
d. Extremely hard to determine if a group or individual that has failed should be cast aside or given new resources to try again – all of history teaches us that the best folks who we celebrate as geniuses failed many times before succeeding. Of course, fools will fail as many times as you allow them.
e. No easy answers here other than did the group or individual at least execute well and move with efficiency and speed, that is necessary component but not sufficient
3) You should study the strategies and successes and failures of others, but because successes and failures are the result of an enormously complex environment with variables too numerous to count, we should not study them so much or lend them credence to an extent that we “over-fit” our analysis and plan.
a. Successes are tainted with survivorship bias – be careful when determining why and how they were successful
b. Failures are often due to a lack of ability to recalibrate from the inevitable setback or a poor team in charge of executing than a mistake in vision/goal
4) You will need capital sufficient to start and recalibrate at least twice
a. Undercapitalization is a guaranteed route to total failure
b. Perfect Oscar Wilde quote in the article. We often know the price of everything (salaries and the cost of a team) but the value of nothing (if not a great team staffed as required then we will surely fail, so what are we tabulating)?
5) The humans involved in the endeavor will override everything else – need to find and bet on the best team possible
a. All of the above will be evaluated, capitalized, planned, recalibrated and executed by a team of humans – they will determine success or failure
b. What makes life exciting and fun is that this means complexity and unpredictability – how do we find and support the right folks? How do we determine that the inevitable failure is the result of the vision/plan or of the humans charged with executing it? Can we hand additional resources to the team/individual who have failed twice already? Bill Gates, Steve Jobs and Jeff Bezos, etc. failed many times – what did others see in them to continue to provide resources?
c. Key question for us all is how to identify and recruit the right team – from the very top to the bottom – and hold all accountable for efficient and effective progress (not success)
The Manhattan Project is a great example of survivorship bias – we celebrate it (in a way) because it was successful. However, that does not mean that throwing $22B and a lot of really smart people at an objective means you will succeed. Indeed, that approach has failed more often. It is in the intangibles – I would argue that the sheer pressure that it must succeed (because of the fear of what if we lost that race) meant the accountability and demand for speed was very large – folks got out of the harbor and recalibrated at high speed (or so I imagine). Pretty hard to re-create that environment – hence the Haldane teaching that it is dangerous to extrapolate characteristics in a complex world
The keys to success when moving complex systems toward a goal are:
1) Engineering problem, not a theory problem.
2) Clear definition of success.
3) Response to a real existential threat.
4) One customer.
Much of what we propose as innovation does not clearly meet any of these criteria. Manhattan and Apollo did. The Transcontinental railroad is another example that meets several or all of these characteristics.
And that's what a group of wonderful colleagues debated on a Sunday morning.
By John Halamka and Paul Cerrato — Mayo Clinic Platform_Connect is transforming how patient data is used to generate innovative diagnostic tools and treatment options.
By John Halamka and Paul Cerrato — What happens when ChatGPT-4 and a human cardiologist are asked to diagnose the same patient? The results are quite revealing.
By John Halamka and Paul Cerrato — Several thought leaders and stakeholders have joined forces to create GoodDx.org, a searchable database that has the potential to reduce the human suffering affecting millions of Americans.