Solving the COVID19 Decision Crisis: Moving from Data-Driven Decisions to Decision-Driven Data
Right now, getting vaccines “into arms” is the world’s greatest pandemic focus. But everything could change. Says Andy Slavitt,
“The headlines we see today…[describe a] massive over-demand and under-supply [of the COVID19 vaccine]…it could easily flip…we could have a massive oversupply, and we have more vaccines than people could possibly need, and now we’re trying to figure out how to get people to take them…each problem is equally important.”Andy Slavitt, temporary senior advisor to the Biden administration to support the Covid-19 response, speaking on In the Bubble
How can we stay agile enough to adapt to this kind of change, which is the rule—not the exception—within the pandemic? Slavitt adds,
“We’ve been running this whole Covid response on whack-a-mole…from the beginning we never saw around corners…[and didn’t] anticipate very well.”Ibid
This “whack-a-mole” problem means that we fix one problem and it creates another somewhere else. We’re missing the big picture.
This is a symptom of a number of problems we face as humanity. These include Covid-19, climate change, and global inequality. They:
- Are invisible
- Involve exponential behavior dynamics
- Involve actions that play out over long chains of cause-and-effect in time and space
- Involve complex systems interactions including feedback and winner-take-all mechanisms
- Can only be solved with great science and data.
People have a hard time understanding situations like these, and making good decisions within them.
Imagine a community health organization, looking to spend limited dollars to improve Covid-19 outcomes in an underserved community. Or a hospice, which is making difficult decisions regarding whether to allow family members to visit their loved ones. Or a university risk officer, who’s been given a budget to spend on safety so that students can come back to school.
We’re asking decision makers to become health experts, placing them in the impossible position of making life-and-death decisions with limited information and expertise. And sending them a great data set is like delivering chocolate, not chocolate cake. Government guidelines, data, and industry standards are limited in important ways, including the fact that they are one-size-fits-all approaches that may not be a good fit for an individual situation, and that can asymmetrically negatively impact disadvantaged populations.
We can do much better. Just as computers connect us through social media to friends and family it can also bridge from “under the hood” technology like artificial intelligence and simulation to decision makers, helping them substantially with this otherwise extremely challenging—if not impossible—task.
Good decision making is humanity’s most underutilized sustainable resource; it’s time to tap into it to solve Covid-19, and more.
A recipe for tackling wicked problems
We’ve done this before, in marketing and, increasingly, in sports, financial services, and medicine. Slavitt’s stepping away for 130 days from his In the Bubble podcast, and guest host is Dr. Robert Wachter, Professor and Chair, Department of Medicine, University of San Francisco, and author of The Digital Doctor. Wachter gets it: “[My] older son Doug does baseball analytics…through him I’ve learned what an industry looks like if they take advantage of all the information they have in their computers and use it to make better decisions, which we don’t do well enough in health care.” [emphasis mine]
The good news is that a solution to the Covid-19 crisis will be reusable to other situations that share these characteristics, including the climate crisis, poverty, and global inequality. To solve problems like these requires that we shift from optimizing silos of practice to integrating them, and we must use data in new ways, along with computers to process that data and to connect us together for shared understanding and coordinated action. This is especially needed because complexity creates an environment that’s hospitable to grifters, who can work “in the dark” to incent actions that achieve hidden agendas.
Covid-19 requires that we solve all links in a chain
Here’s an emerging picture of how this fits together. A worldwide ecosystem of experts is beginning to supply pieces of this puzzle, which can be reused and integrated without the need to reinvent the wheel.
Keeping it simple for policy and decision makers
Although this picture can look complex, the combination of the elements in circles make up an ‘engine’, which can be operated from a ‘driver’s seat’ that doesn’t require expertise in the underlying technologies. One way to do this is to take a page from the book of airline pilots: to use an immersive simulation engine that shows the impact of decisions, as shown below (here’s a climate example, too). This is not unlike the sophisticated phone or computer you’re using to read these words: there’s a lot going on “under the hood” that you don’t need to understand in detail, but are nonetheless necessary. What’s new is that creating custom simulations like this, for your particular situation, is hugely less expensive than it has been in the past.
As someone who’s delivered applied machine learning systems for decades, I’m painfully aware of their limitations, in particular. We need to integrate and orchestrate multiple machine learning models, and to integrate them with other disciplines and with human decision makers as well.
To date, the lack of separation between what’s required to use the specialties in the circle above and what’s needed to be build them has created a big barrier to integration. Car drivers can no longer be expected to double as automotive engineers, and this situation is similar. And nobody can possibly be an expert in AI, simulation, economics, and Computational Fluid Dynamics. For this reason, a key part of getting them to work together is to separate the “how” from the “what” for each one. This “system of systems” approach is core to the engineering of all complex objects, and good decisions about Covid-19 are no different.
Another problem to date is that we’ve been paying disproportionate attention to certain links in this chain (like vaccines, medical studies, and data) and systematically undervaluing others (like behavior change and collaborative linked-up decision making.) There are a lot of reasons for this focusing within specialties and silos, and ignoring their interconnections. A “neo generalist” mindset is an important shift to move to the connected-up thinking that’s required to solve Covid-19, and more.
The vaccine alone won’t save us, nor will data or medicine. It’s time to take a systematic approach to multi-level, evidence-based decisions seriously, along with connected-up support for them, so we can look effectively look around corners towards the end of the pandemic.
Although decision intelligence and orchestration are both young fields (just over ten years old), they are the right fit for the Covid-19 challenge. What better time to address them to move to the next level of maturity in how we make evidence-based decisions, from data-driven decisions to decision-driven data?
Please subscribe for upcoming posts about the emerging decision intelligence and orchestration ecosystem for Covid-19.