What PPI co-founder Ryan Hunter learned from his time in San Francisco's COVID Command Center.
In February 2020, I was leading a workshop where an activity required everyone to put their hands in a circle. When we did this, someone sneezed -- and my co-leader screamed. She'd been hearing something about a new virus going around.
A few weeks later, we would be thrust into the thick of San Francisco's COVID response, supporting the City's COVID testing infrastructure. A lab testing system built for low-volume, long-term surveillance needed to be transformed overnight into a well oiled machine working at 100x its prior volume.
Over the next months, the Lean methods we'd practiced and taught for the last 5 years would be battle tested. There were more than a few bumps along the way, but with our work and that of dozens of other public servants, San Francisco led the nation in testing capacity in those crucial early pandemic months.
Here's a few key moments and lessons from that story.
Lesson 1: Look at the whole value stream
I (Ryan) was initially deployed to assist the Public Health Lab (PHL). Every county in the country has one, and this is where the original CDC COVID assay was first sent.
For several days I followed PHL microbiologists, mapped the process of specimen analysis, and added words like "thermocycler" and "aliquot" to my vocabulary, all in an effort to make PHL more efficient. Until I realized something crucial.
PHL wasn't the problem.
Truth is, the lab was just one stop in the long chain of steps needed to get a test result. That chain is called a value stream in lean parlance -- it starts with a symptomatic patient who has to find a test site and get swabbed. That swab is transported to the lab, analyzed, and resulted (to whom?), and then the patient needs to take action on the result. And the problems were upstream of the lab.
We took a step back. We spent a couple days mapping out the complicated network of specimen collection sites, couriers, labs, and results communication. With this system map in hand, we could understand the flow of tests through the whole system -- and start to pinpoint problems that would lead, ultimately, to faster time-to-result.
Lesson 2: Identify bottlenecks
All problems are not created equal. In every process, there are places where resource constraints slow everything down, like an accident that blocks lanes of the freeway and makes a traffic jam. Making the process faster without addressing the bottleneck is like pouring more cars onto the freeway -- it just builds up more traffic before the accident.
In COVID testing, there were three main resource constraints:
Our supply of test kits
Sites for specimen collection
Laboratory analysis capacity
To scale up the system, we had to scale all of these together. Our staff in the labs helped organize schedules of staff and equipment to maximize analysis throughput. Other testing staff worked to set up large, publicly available test sites for specimen collection. But managing test kits was the thorn in our side for months.
Lesson 3: Manage scarce resources through kanban
Early in the COVID response, we were flying blind with test kits: When we stood up a new test site, we would deliver a trunkload of test kits and hope for the best. The Testing Branch's time was sucked up by urgent phone calls from test sites asking for more kits, leading to testing staff driving around the city delivering supplies.
When a resource is scarce, people hoard it. So sites that could get their hands on kits often took more than they needed, making scarce supplies even scarcer.
We needed a way to allocate our scarce kits fairly and thoughtfully, in a way that maximized access to testing while recouping our time. We took a step back.
In Lean, inventory is bad -- you don't want to keep extra supplies around. But you also don't want stuff to run out. The solution is a reorder trigger called kanban, that lets you resupply just in time.
Here's how kanban works:
Estimate need for supplies: From our system map, we knew all our test sites. We contacted them all to learn their daily testing need.
Set inventory levels based on need. For example, all sites should be supplied with enough kits to test for 10 days. At a large site, that might mean 2,500 kits, while 10 days supply for the Medical Examiner might mean just 50.
Set a reorder trigger. This trigger is the kanban: In this case, if inventory levels dropped below three days' worth, we would resupply.
Continuously improve. Our initial estimates would prove wrong. That's fine. When a site used supplies faster than expected, we would increase their daily testing need estimate. A site with highly variable volume might have a higher trigger level.
The result was a simple spreadsheet like this:
Site | Kits needed daily | Resupply to this level... | When supply drops below... | Current level |
Test site 1 | 250 | 2500 | 750 | 1200 |
Test site 2 | 30 | 300 | 90 | 75 |
Test site 3 | 75 | 750 | 225 | 750 |
In this example, Test Site 3 has just been resupplied. Test Site 2 is in need of a resupply, which needs to happen within two days for them not to run out. Test Site 1 will be flagged for resupply the day after tomorrow.
It was a simple system -- pulled together in under 48 hours -- but it ended hoarding, let us stretch our limited test kit capacity as far as it would go, and gave us data that we needed to make decisions about prioritization.
COVID testing is so different today: testing access is ubiquitous, and when people test at all, they're doing it at home with a rapid antigen test. At the time, though, these simple improvements got people more quickly into treatment, kicked off effective contact tracing, and ultimately saved lives.
Lesson 4: Mindset Matters
When I served as Chief of Testing, it meant long hours, frayed nerves, flaring tempers, thorny problems, and high stakes. Working in that environment took its toll on everyone.
And yet: In some ways, I wish we would stay in that same mindset. Because I saw the power of smart, passionate people who were driven by a common goal, tackling seemingly insurmountable problems, because lives were on the line. Those public servants somehow found a way to slash through years of bureaucracy and red tape, partly through the leeway of emergency declarations, but also through an imperative to do what was needed even if it was inconvenient or just wasn't "the way it's always been done."
But here's the thing: Our governments help people find jobs, keep them safe on the road, provide social safety nets, and care for the homeless every day. Lives are on the line every day.
And so we have an imperative: To figure out how to be responsive, compassionate, and creative in providing the best public service we can. To find creative solutions to tackle the new problems facing us today. To prioritize people over red tape, and systems thinking over firefighting.
When you show up with that mindset, you can do more than you ever imagined. You can build a kanban inventory system in two days.
Comments