Detecting potential Coronavirus cases in real-time, based on 911 call data

chest pain - heart history
A 911 call flagged as a potential Coronavirus case due to symptoms listed by the caller.

As we adapt to life in a post-coronavirus world, a central challenge and opportunity is to convert an overload of raw data into actionable information, in real-time, so that emergency managers and operators can make decisions and take action quickly and efficiently.

If you work for or are connected to a Public Safety agency, Emergency Operations Center (EOC), or Health Department looking to integrate and analyze dispatch data, we can help, at no charge. Schedule time with our CEO here.

Problem #1: Siloed Data

In the case of the first County where we went live in May of 2019, the primary County 911 call center does not have a direct data link to the primary ambulance service in the County, that runs their own Computer Aided Dispatch (CAD) system. To complicate matters further, the largest Fortune 500 employer in the County runs its own, separate incident management software and process.

A patient may report COVID-19 symptoms to one of the two public safety agencies, but the other won’t have access to that data. This siloed information prevents the emergency operations center from having a complete picture of all available data.

In the case of a military base, the problem can be even more significant On our first Air Force contract, we studied a base outside of Boston that could have up to 19 agencies, state, local, federal, university, enterprise, and military, responding to a major incident. Each of these agencies has its own software and databases, siloed from the rest.

The 19 public safety agencies located around Hanscom Air Force Base, outside of Boston.

By securely collecting and fusing incident data from all agencies into a standard format, SimpleSense solves this interoperability problem.

Problem #2: Converting Raw Data into an Actionable Information

The bigger problem is converting all this raw data into information that a user can act upon immediately.

Over the past week, we developed a method to flag 911 calls as potential COVID-19 cases. By comparing 911 call data to a list of keywords based on symptoms of the virus, we can identify potential cases and flag them for the user.

SimpleSense interface
A screenshot of the SimpleSense interface, with a 911 call flagged as a potential Coronavirus case.

However, there are a few key issues to solve. How is this information best displayed? Who has access? And, when will it be shared? Emergency personnel are likely already inundated with information and sources of information, so any new information must be simple to digest, useful, and actionable.

User #1: The Emergency Operations Manager

The EOC manager is looking for high-level trendlines, hotspots, and situational changes and may check the data one or two times a day.

A heat map could be a good way to visualize new potential hotspots of the virus. Plus, it could give managers a heads up of new clusters before they get out of control. This info is especially helpful if managers could view the map based on time, i.e., new cases over the last 24 hours compared to the past week.

live map of covid-19 in Hong Kong
A live map showing cases of COVID-19 in Hong Kong.

Trendline analysis can also be useful, showing if measures to control the spread of the virus are slowing its spread. We ran an analysis of over 66,000 911 dispatches from a County of ~275K people over the past 120 days. Note that this analysis will also pick up symptoms for the flu, which was bad this winter.

911 calls involving potential COVID-19 symptoms.

User #2: The ESF

Emergency Managers structure their EOC into defined roles, called Emergency Support Functions (ESFs), to monitor specific areas. ESFs view streams of information on an ongoing basis, looking for meaningful signals in the data that inform recommendations, preventative actions, and response operations.

The NYC EOC, one of the largest in the world, has over 130 workstations.

SimpleSense can provide real-time alerts via email, SMS, or to a web interface on a tablet, mobile device, or desktop computer, of potential COVID-19 cases, along with additional information on the case, such as address, neighborhood, structure type, etc.

For example, this enables an ESF to immediately communicate with a large housing complex to initiate disinfecting and quarantine procedures if there’s a suspected case of COVID-19 at that location, without delay.

We’re here to help

We have engineers ready and willing to volunteer their time to integrate dispatch data into alerting tools, data visualizations, dashboards, and whatever is most helpful. Contact our CEO here if you think we can help.

Share this post: