Data management in testing times at UC Berkeley’s Innovative Genomics Institute

On 13 March 2020, with the impact of the COVID-19 pandemic starting to become horribly clear, an extraordinary meeting was called at the University of California, Berkeley (UCB) by biochemist Professor Jennifer Doudna.

At that meeting, Professor Doudna asked her fellow scientists to set aside their usual roles and research agendas for six months and instead volunteer their time, expertise and ideas to serve their local community during the crisis.

Several hundred scientists turned up in person, or called into the meeting from their own homes. That’s no surprise, given Doudna’s profile and influence. As well as being the founder of the university’s Innovative Genomics Institute (IGI), she is the co-developer of the genome engineering technology CRISPR, for which she shared this year’s Nobel Prize in Chemistry with Professor Emmanuelle Charpentier.

At the meeting, Professor Doudna asked attendees for their ideas on how they might help. The overwhelming consensus was clear. What was needed in California’s East Bay and beyond, they said, was testing - a lot more testing, and in particular, testing for members of underserved and at-risk communities. These included homeless people and those living in crowded, multigenerational housing, as well as first responders and essential workers.

And so the idea to launch a COVID-19 diagnostic testing lab at the IGI was born.

Big project, tight deadlines

There was a huge amount of work to be done, in a very short space of time. The plan was to get the new lab up and running in just a matter of weeks, because the unfolding crisis meant that the testing effort simply couldn’t wait.

Physical space at the IGI had to be given over to the initiative and equipped accordingly. Volunteers had to be trained on how to work with patient samples and protect private information, in safe, socially distanced learning sessions. And the new lab needed new data management capabilities for securely processing and tracking information relating to patients and their test results.

On that last point, Shana McDevitt was one of the many UCB scientists to attend the meeting and subsequently took the lead on the data management effort. Normally, she explains, she’s the director of the QB3 Genomics Sequencing Lab at UC Berkeley, but the work to get the Covid testing lab at the IGI up and running has consumed some 10 to 12 hours a day, seven days a week throughout much of this year. She says:

It’s been exhausting, insane. I started work on the basic set-up of the lab, because that’s what I do, professionally - I run laboratory facilities. But quite quickly, I moved off into the data infrastructure layer, seeing very clearly that we needed some kind of laboratory information system to deal with all of the data we’d be processing.

One of the first calls she made was to Third Wave Analytics, a San Francisco-based specialist provider of laboratory information systems (LIMS), with whom she was already working on a system to tie together her own QB3 facility with the IGI. Could the company help her build a LIMS for the new Covid testing lab at IGI, she asked? Could they do it very, very quickly? And could it also be easy for time-pressed physicians to learn and use, not to mention HIPAA-compliant?

A robust set-up

In a matter of weeks, Third Wave had built a robust LIMS based on its LockBox platform to manage the technical data flow of clinical samples in and out of the lab.

Subsequently, with pro bono support from Salesforce and MuleSoft, it also built and implemented a secure physician portal, based on Salesforce, and integrated with the LIMS, in order to accept requisitions for testing and return clinical-grade results to healthcare providers using the lab.

This set-up was then integrated with the electronic medical records (EMR) systems of healthcare providers using Mulesoft, saving a huge amount of time and manual data-input effort involved in the exchange data between the lab and these organizations. The level of automation provided by this set-up is critical, says McDevitt:

If you can’t get various tasks done automatically, you have to have a human involved. And my experience in running this facility is that my most critical resource is people. I just can’t find enough people to get the work done. We’ve got resources in spades: you know, machinery and testing equipment and swabs. But we couldn’t possibly run at the scale we do now without getting crafty with technology, because I absolutely can’t be relying on humans just to move data around.

Integration using Mulesoft has not only had a big impact on the number of tests the lab is able to process today, she says, with averages rising from hundreds of tests per day to over 1,000, but has also resulted in the lab being able to return results to patients and their physicians far faster, typically between 24 and 48 hours:

These are difficult populations to serve, but it’s so important that they ARE served. It’s a bit like a game of Whack-a-Mole, where you’re helping and making progress in one area of the community, but then another hotspot flares up. It’s really, really hard to make a difference. But on the whole, I feel like we’ve worked well with the California Department of Public Health and with healthcare providers on the absolute best effort we can to get as many people tested as possible, as frequently as possible, with the quickest turnaround times, so that contact tracing and quarantining can work.