Listen to the podcast:
Adelphoi Village’s President, Mark Mortimer, joins Coruzant Technologies for the Digital Executive podcast. He talks about an application called FirstMatch, which leverages machine learning and predictive analytics. This algorithm takes the client’s assessment data and matches them with a program or service that has the highest likelihood of achieving the desired outcome.
Mark Mortimer is the President of Adelphoi Village. Mark oversees planning, design and implementation of business operations at FirstMatch.
He is a graduate of Indiana University of Pennsylvania where he earned a Bachelor of Science in Criminology and an MBA. He began his career at Adelphoi as a counselor and in 2002, he was named Supervisor of Adelphoi’s Middle Creek Secure Group Home, which was named PA Juvenile Court Judges’ Commission Program of the Year in 2005. He has also served as Director of Residential Services and Vice President of Adelphoi Village. Mark is an expert trainer in Crisis Management and Situational Leadership with Adelphoi’s Professional Development Program.
Welcome. I’m here with the staff of Adelphoi Innovative Solutions. My name is Karyn Pratt. I’m the Vice President of marketing and strategy development. And I’m here with Mark Mortimer, who is the President, and Shawn Peck, who is the Vice President. And we’re here today to talk about FirstMatch. So, Mark, tell me a little bit about FirstMatch and why it was developed.
Sure, Karyn. So, FirstMatch is an artificial intelligence, machine learning software application that matches you with the most appropriate program position to meet their needs.
Sure. So less time away from the family. Clients have reduced trauma, higher likelihood of a positive outcome earlier in their system involvement, and a reduction in system penetration.
So Shawn talk to me a little bit about the typical implementation process.
Sure, Karyn. So typically it begins with engagement and readiness. So we have different tools that we use to identify where a service provider is in readiness process and also a referring agency. Once we know how many programs we’re working with and what type of data they’re collecting, we start to build a tool for them. We enter data into the tool and train the tool. So once it’s trained that we can start making predictions, which leads to dashboards and other performance analytics. And then after that, we have to validate the tool to continue to monitor it. So there’s a maintenance process as well.
Great. So as you’ve worked through this implementation process, tell me about some of the challenges that you faced.
I think the biggest challenge be Fidelity, also the availability of data at the time of referral, and also access to outcome data and remain out of care data.
Thank you. So, Mark, tell me a little bit about the platform and how the data is entered and how it’s managed.
So it’s a cloud based platform. And as Shawn mentioned, we work with each provider to build out a tool that’s customized for the information that they commonly receive. When they get a referral from a placing agency, they have complete control over entry of data into their tool. Typically, an average referral will take three to five minutes to enter into the tool. As I mentioned, the provider is the only one that has access to their own data. They control the credentialing, they control the permissions and their staff.
It’s a HIPAA compliant tool, so we can’t see that providers information in the at all. So we’ve set up this platform in a way, it’s very user friendly. And we’ve used conditional formatting to help speed up the process of a entry into the platform.
So, Mark, if I’m a FirstMatch user, tell me how I know that the model is functioning correctly.
Yeah. As Shawn mentioned earlier we have an integrative dashboard function. As the platform and as the application is making predictions, and we record those outcomes, we can tell in real tireal-timeme as those outcomes occur if they align back to the prediction that the application originally made. So the providers getting real treal-timeime feedback about how well they’re predicting the outcomes that they’re experiencing.
So, Mark, talk to me about the prediction report. Who sees those likelihood scores and how are they used?
So it depends on where the tools deployed. If the tools deployed by the provider organization, then they’ll be the only ones to see those scores, and they’ll decide who they share those scores with. Oftentimes, we encourage the providers to share that score with the placement decision maker at the Juvenile Probation Department, Child Welfare Organization, Behavioral Health Organization. We want them to share the scores with the placement decision maker. That decision maker can make the best decision for that particular kid. If the tools deployed by the placement decision maker, then of course, they have that information, and they don’t need to share it with anyone because they are the decision maker.
So typically, the way we’re seeing this rollout is residential providers oftentimes have to control their own tool and share that information with the placement decision maker. On the community based side, we’re seeing counties implement their own tools.
So what other systems do you think might find value in the FirstMatch platform?
Yeah, it’s a good question. So I would say there are a broad range of programs that would find value in this. We know that educational programs there’s certainly an application there. Drug and alcohol providers have expressed interest and value. And then, of course, the adult system, the system that’s rehabilitating adults.
All right. So thank you both for joining me today. If you’d like to learn more about FirstMatch, you can call us for a demo, or you can visit our website at www.FirstMatch.Com.