Predictive analytics

strategies? It's a million dollar question.

We always talk about it from the lens of predictive versus proactive.

Predictive is you're trying to predict the next behavior.

As a marketer or an agency supporting a marketer, you're trying to figure out, okay, what is the next best action? What is the next best offer? What is the next best product, for example? Proactive is a little bit differently.

It's actually engaging with that potential consumer in real time.

Get a front row preview of the latest innovations, expert insights, and what's shaping the future of pharmacy practice.

I'm Cindy von Herringen, senior education director on the national meetings education team at ASHP, and today we're sitting down with Jennifer Sternbach, assistant vice president of clinical pharmacy services at RWJBarnabas Health, to discuss her upcoming session, Data with a Dose of Insight, Predictive Analytics in Health System Pharmacy.

Welcome, Jen.

Thanks for joining me.

Obviously, in an hour and a half or three hours, you can't change the world, but you can change the questions you're asking and what you're going to look into when you go home, so thank you for that.

That is all the time we have today.

I wanna thank Jennifer for joining us to discuss her session, Data with a Dose of Insight, Predictive Analytics in Health System Pharmacy.

If you haven't already, I encourage you all to visit leaders.ashp.org to register for the meeting.

love that.

Before we ask you kind of how do they get in touch and for U.S. Foods existing clients, where do you see it going, James? You know, you guys have taken some of the new technology, allowed people to build out recipes in ways that we never could do before, allowing them to do some predictive analytics on what they're going to need for tomorrow and their next delivery.

All of these things 10 years ago, 15 years ago were next to impossible to do.

They took schools and schools of people to kind of figure these things out.

And where I see this going is more looking further in the future to help them make decisions around ingredients, food choices, alternatives, to make sure that they are profitable and stay in the mainstream there.

So I think that's one area.

Another area that I see this going is just being able to use predictive analytics as we link tools together.

I think, as I said, we do inventory, we do menu profitability, we help build menus, but I think as you see more and more of these tools connect to point of sale systems and other things, and you're bringing all those things together, You're going to be able to look at analytics and start messaging them in the future like, hey, we see the profitability of this slowly eroding.

My pleasure.

If we look at things from the big picture, what do you think companies usually get wrong when they look at a topic like predictive analytics in HR? Can you start there maybe for us?

Sure.

I usually start at the beginning, which is the data that underlies all these statistics and analytics.

So we may not enter the new dependent that we have.

or we may not enter the loss of a loved one, or we may not put in for that time off that we've scheduled to take.

And so measurement of the data that's in our HR technology systems typically is flawed because even if we have the best predictive analytics in the world and the best descriptive analytics in the world, they're going to be wrong because the data underlying them is wrong.

And That gets into the whole concept of artificial intelligence utilizing the data that's in our HR systems.

Okay, so these are all very insightful.

Paul, can I ask a question? The predictive analytics that you talked about, maybe you didn't use that exact terminology, but the predictive ability of AI in the future, is that something that can really look at all this telemetry and help make predictions about where vulnerabilities lie?

Yeah, it's actually, I mean, we've been evolving this capability for quite some time now, and some of our competitors do too, right? Like this is absolutely a use case.

Now, the thing that we do differently now, and I think you'll see an acceleration of this across products in the industry, is we have enough corpus of data now to look at and to analyze and say, oh, we can say with 99.9% accuracy.

As a security leader, now you're actively defending against AI-driven malware that can observe the network environment in real time, understand what's deployed as far as defense goes, and dynamically rewrite its own code to evade capture.

So that's, I think, really what I hear from people on their mind.

When we've been at some shows, you know, being able to show them some of the benefits of having that AI, fighting AI technology, right, showing them how quick I can get an incident summary done, So I can, you know, from a SOC threat hunter perspective, move on, get leadership the update they need, as well as getting those predictive analytics that are going to say, hey, this is the likely next step.

Go ahead and stop this right now.

We were the first influencer marketing platform in the world.

And, you know, Predictive is really a continuation of that goal.

And, you know, how do we, how do we make this the biggest part of advertising? Because I believe, you know, if brands can choose to spend their money with Meta or TikTok or with individual people, uh, I'd rather go to people.

I think they're better stories.

Another area you mentioned is about sort of this care management and care standardization.

What has been the general approach to using AI to implement clinical protocols or predictive analytics? What are some of the more successful areas that you've been able to achieve in that area?

I think from a predictive analytics perspective in the population health space, we've found a lot of benefit from leveraging this in regards to our cohorting and how we do outreach to our patients.

We've had a lot of work historically in trying to decide, okay, and I alluded to this a little bit before, when you have large populations, how do you slice and dice them? What's the Venn diagram look like of population identification so that you can get upstream and create what we call non-events, meaning you want to prevent that DKA event.

You want to prevent that MI.

You want to prevent the stroke.

In order to do that, you have to get upstream and try to find patients that are at increased risk and have modifiable things that you can intervene on.

And so predictive analytics, which oftentimes, you know, leverages these AI tools has been something we've leaned into.