A hallmark of the current pandemic is the overwhelming volume of real-time data. It’s easy to be overwhelmed by the stream of dashboards, testing counts, reopening metrics, and excessive death estimates.
The Hill, 5/15/20
An essential requirement for states in the process of reopening is outbreak surveillance: identifying community spread of the virus in real-time, without delay. Undetected outbreaks can exponentially grow and quickly re-infect other areas, especially in under-resourced and rural areas which are now the fastest growing areas of infection in the U.S.
The most vulnerable counties in dark blue, as identified by the COVID County Vulnerability Index.
The challenge for emergency planners and health managers is finding meaningful signals in the the volumes of data currently available. In this post, we explore a critical, overlooked source of medical intelligence: 911 dispatch data.
An Overlooked, Critical Dataset: 911 Dispatch Data
Hospital and testing records are the current best, near-real-time datasets available to decision-makers. However, there is another overlooked, front-line dataset that can provide rapid, accurate measure of infection: 911 dispatch data.
Could he be infected with the coronavirus? Possibly. But he was neither wheezing nor gulping for breath. The chief explained to the man that he could become far sicker by being in the hospital. “So do you just want to stay here?” he coaxed. Looking bewildered, the man nodded and shuffled back inside. The ambulance roared off to answer another call.
12 Fraught Hours with EMT’s in a City Under Siege (New York Times, 4/1/20)
Many cases of COVID-19 go undetected, and patients do not make it to the emergency room. Early estimates show an excess mortality rate (undetected COVID-19 cases) of 3-5X reported numbers.
Many of these patients do call 911 but either die before being tested or do not end up going to the ER. In fact, in some municipalities, first responders actively discourage patients from going to the hospital.
Data provided by FDNY. FDNY Sees Huge Uptick In DOA Ambulance Cases As COVID Ravages City, TPM on 4/10/20
911 call takers follow specific protocols to identify and prioritize callers with symptoms of COVID-19, documenting their findings in Computer Aided Dispatch (CAD) systems. Once shared, these symptom logs can be mined in real-time to identify increases and decreases in front line cases of COVID-19 that may never make it to the emergency room or hospital.
Pandemic Flu and Protocol 36, International Academies of Emergency Dispatch
Overcoming Historic Barriers to Information Sharing
The core challenge with 911 dispatch data is that over 6,000 Public Safety Answering Points (PSAPs) exist across the US, with most counties and cities running their own PSAP.
Adequate surveillance requires real-time access to data. PSAP managers have traditionally locked down access to these figures for a variety of privacy, budgetary, and management issues.
We must overcome past barriers to data sharing and make our best effort to securely aggregate, standardize, and deliver critical data to decision-makers in real-time. Fortunately, there’s little technical barrier in the way of real-time information sharing.
Working with Fortune 500s and the US Air Force over the last year, SimpleSense has developed, tested, and deployed the technology infrastructure to rapidly aggregate and normalize 911 dispatch data from thousands of PSAPs, using existing software. What’s more, this deployment requires minimal time effort by the PSAPs.
There are two primary impacts of successfully mining 911 dispatch data:
- Identify and respond to future infection outbreaks in real-time
- Build a rich dataset for researchers to measure policy effectiveness over time
We’re offering this service at no cost to state emergency managers. We need help reaching these teams now as this tool is critical during the reopening and monitoring phases of the next months and years.