The Los Angeles County Department of Public Health identified a strategy to manage the torrent of data that began arriving early last year as the pandemic took hold, said David Cardenas, the department’s acting deputy director of operations and chief information officer. He spoke Tuesday during a discussion of how “Automation is Changing the Game” at the Los Angeles Virtual Digital Government Summit. Among the takeaways:
- Charged with preventing disease and protecting the health of nearly 10 million residents, DPH is considered the lead response agency during health emergencies, including large outbreaks like the pandemic. It activated in earnest early last year to monitor COVID-19 cases and saw a significant upsurge in March 2020. The declaration of a local health emergency enabled further planning, and DPH scaled up response teams, identified additional staff from local entities, purchased supplies and coordinated with partners. An increase in cases, though, quickly translated to a rise in the influx of data coming in from those partners. DPH collects data on “thousands of cases and contacts per day for COVID,” Cardenas said during a conversation with Avik Batra, advanced customer engagement lead for public sector at Accenture. And it encountered difficulty managing the import of all that raw data into its system of record — dedicating “dozens and dozens of staff” to data entry just to keep up.
“It was just becoming too burdensome, a very, very resource-intensive process. So, we were really needing a solution to provide that and speed up our ability to publish data that the public was seeking on an active basis,” Cardenas said. - With that in mind, DPH worked with Accenture, utilizing robotic process automation from UiPath to automate some of its most time-intensive processes, then tracking the results of that automation. Specifically, work centered on three use cases for the bots. One targeted the creation of an “initial case record” in the county’s Integrated Reporting, Investigation, and Surveillance System (IRIS). A second use case focused on when potential COVID-19 victims were interviewed by investigators, uploading and entering that data into IRIS as well. And a third use case centered on uploading hospitalization data being collected into IRIS. To date, Cardenas said, more than 600,000 records have been entered into the system via the bots; he highlighted the “human resources” able to be diverted as a result.
- The work also enabled transparency in several different ways. The data collection itself was, the CIO said, “under intense scrutiny” that is ongoing as staffers “really want to make sure that we’re collecting all the right information,” focusing on the correct statistics, handling personal health information securely and being as open about the process as possible. Utilizing RPA confronted the chance of human error and sped the process up, while the tracking technology let officials view “every single step the bot was taking” and retrace its work if needed. Stakeholder buy-in was crucial, Cardenas said, and DPH found a group that was ready to pilot the technologies — which, as they spooled up, freed employees to focus on higher-value tasks. Establishing such a “beachhead” drove further interest in what the technologies could do — and once COVID-19 cases subside or become more manageable, the CIO said, plans are to look at extending the process to other use cases. This could include mundane tasks related to other disease types, clinical work, billing and claims, and financial transactions in need of quality assurance.
“Those are places where our next targets are. Certainly, as resources become available, that’s where we’re going to be headed,” Cardenas said. - He offered several potential takeaways for state and local governments contemplating a similar use case or application. DPH’s approach was “to start small” and build confidence in what the bots could do. Cardenas suggested starting “from a very targeted activity, a targeted use case” — again, enabling replicability if needed. Choosing a simple workflow is similarly valuable, as re-creating a complex workflow can be problematic. It’s also helpful to team with stakeholders who are more comfortable with the new work; DPH was able to do so, and found that activating the bots became “almost like a natural progression.” Solid documentation of the time expended pre-bot and saved post-bot is also vital, Cardenas said, noting DPH was able to show hours saved, tasks the bots accomplished and the higher-level work this all brought within reach of staff.
