Code for America is working on reducing ER calls in Long Beach using big data

Updated at 11:56 a.m. PST to show the Long Beach project is in progress, not complete.

Code for America is working on a project that uses big data to reduce emergency room calls in the California city of Long Beach.

[aditude-amp id="flyingcarpet" targeting='{"env":"staging","page_type":"article","post_id":1495720,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"big-data,business,","session":"D"}']

Rhys Fureigh, Molly McLeod, and Dan Getelman are the three Code for America 2014 fellows working on the Long Beach project. Fureigh is a web developer. McLeod is a graphic designer. Getelman was the co-founder and chief technology officer of Founders Fund-backed education startup Lore, which Noodle Education acquired last year.

Long Beach is one of Code for America’s 10 partner cities this year. The city wants Code for America to help it improve health care there.

So the team went to Long Beach in February and researched the situation. “We talked to paramedics, we talked to people from insurance plans, we talked to people from hospitals, we talked to police officers,” Getelman said in an interview with VentureBeat. “One of te trends that struck us is that all of them have this part of population they see all the time.”

The problem was not the number of ER calls, but that the same people were making ER calls again and again, which meant there were deeper issues the ER responders couldn’t handle on their own.

One example, not from Long Beach though, was that one woman was sent to the ER multiple times a month because of her seizure. No one knew why. Someone was sent to talk to her. It turned out she had medicine for her seizures, but her family members were stealing it to sell in the street. The solution was simple. They bought her a $19 drug safe, and she stopped having seizures.

So, how to identify those people who are making ER calls repeatedly? The team aggregated ER call data from the fire department and the police department — address, time, and the type of call it was.

To make it actionable, they also poured in business-license data from the city’s Business License Division for determining what the addresses are. If a single-family home made 40 ER calls last year, you probably want to send a nurse there. If a restaurant made 40 ER calls last year, you might want to send an inspector.

The team visualized its data to identify trends for the city’s departments to look at and to assemble the right team to take action.

[aditude-amp id="medium1" targeting='{"env":"staging","page_type":"article","post_id":1495720,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"big-data,business,","session":"D"}']

“It has human impact, if someone is going to ER all the time, and you can fix the reason why, they can live much better lives,” said Getelman. “But it also has big financial impact for the city, because a lot of people who are going frequently are also uninsured.”

Although this Long Beach project is still in progress, the ER has attracted quite a few big data efforts in the past.

University of Pittsburgh’s Schools of the Health Sciences analyzed more than 50,000 phone calls from more than over 3,000 patients and found out that “telephone calls are predictors of how likely patients are to enter the emergency room,” according to a statement from the university. Researchers sought to identify at-risk patients early and prevent their hospitalization.

Symcat is a symptom checker that has analyzed 60 million patient records to come up with machine-learning algorithms to predict your conditions, according to Flip the Clinic. It might help patients better decide if they should go to ER and cut down unnecessary ER visits.

[aditude-amp id="medium2" targeting='{"env":"staging","page_type":"article","post_id":1495720,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"big-data,business,","session":"D"}']

And UC Berkeley biostatistics Professor Alan Hubbard teamed up with San Francisco General Hospital trauma surgeon Dr. Mitchell Cohen to “develop a predictive computer model for the prognosis of trauma patients,” according to Berkeley Health Online. The model tries to better inform clinicians who are making life-saving decisions in the Emergency Department.