ADAM: Awesome. Hey, Jason. Hello, everyone. Thank you so much for joining me and Jason Nardella again for another Prescription for Growth video.
Jason, we've been getting a ton of questions, particularly about and mostly from our real estate friends, but some of them are from a lot of other customers and partners. And that question is really how are we supposed to look into the future when the current state of the union is still so uncertain? So, particularly thinking about the impact that COVID's having, particularly thinking about the number of uninsured patients that are increasing into that market and migrating into that market.
How are hospitals needing to use their demand forecast tools to better predict where they're going in their market? They're making millions, if not hundreds of millions of dollars worth of investments over the next five, 10, 15 years in capital expenses. How do hospitals need to be thinking about that as they're really struggling to really understand what's happening in their market today? Do you have anything that you could lend some insight onto?
JASON: Yeah, for sure. We've been working pretty actually hard with the data scientists here at Trilliant to really come up I think with a really solid and understandable way of demanding, of forecasting the demand in your market, right? And I mean, Adam, to your point, it's not just real estate. I mean, anybody knows if you're going to put a new doctor in anywhere, that ramp up to get them up is three to five years. So, are you putting them in the right place? Are you going after the right patients? How are you reaching those patients? All of these are really, really important, right? Understanding the supply and demand in your market, so how many providers are there currently versus how much of the demand for the population is really important.
But a couple of things that I want to talk about sort of, that you want to make sure that any kind of forecast does have... as you start to understand just the validity of it, because any projection is wrong as soon as you do it right. But there are ones that are better than others, and I think there's a lot of variables that can be taken into account that will certainly help inform or at least get a lot closer to what that volume should be, right?
JASON: One of them is current trends, right? So, what is your base here? What are you basing this off of, right? Now, if you were to run a demand forecast just based off of 2020, it would look really funky, right? You'd have a lot less procedures, because we took two whole months off and the whole country from doing procedures. You'd have a ton of E&M visits, a ton of virtual E&M visits at that, because everything has been going more virtual, especially this year.
You'd have a lot less chronic disease, you'd have a lot less sort of procedures and labs being done on a regular basis, so my question there to you is: is 2020 the right base year to run a demand forecast off the next five, 10 years?
ADAM: Probably not.
JASON: Maybe. I mean, is the pandemic going to last forever? I sure as heck hope not, but the point there being you actually want to broaden out your base, right? You want to have multiple years within your base analysis, and if you can, if you've got a lot of high-powered, very specialized data scientists like we do, you want to actually start to get them to use some of these hierarchical regressions, right? You want to understand what the trends are year over year. Are things moving from inpatient to outpatient? Are certain procedures actually reducing in frequency, and are we seeing an uptick in other procedures?
Those are all these very small nuances that I think for the human brain, if we think about an individual service line, yeah, of course. We can kind of isolate that and we can sort of play that out. To do it across all of medicine and to do it across all of healthcare is very, very difficult and that's what we rely on the machines to do, right? So, using the machine learning techniques, doing a hierarchical regression, understanding what those trends are now, then you can take those trends, you can bump it up against: What's our current population? What actually makes up the community? What demographic information is really important for us and for this type of volume, right? So, so far, we've got sort of what your trends are, what are your base years? What does the population look like right now? Now we take all three of those, we kind of put them together.
Put them in a little bit of a blender, and we start to look at them, right? So, if we know what the population or what the demographics of an area might look like in five years or 10 years, then luckily for us, we've got a ton of great vendors out there – ESRI, Nielsen – these guys that have been doing it for dozens and dozens of years, these prediction models, if we can know what your population and know what your patient cohort will look like, we've got the trends of what is going on now.
We've got an incident rate that looks like that, and then we can actually start to plug that in and look out into the future and say, "Okay, so what does it look like if this population is generally getting older?" Right? Or what does this look like if everybody is now trending into a child-bearing age and starting families. What does your population look like, right? That is going to be the basis of a strong demand forecast that you can really start to plan around, right?
JASON: And it's those types of factors that taken in isolation really are helpful. Putting them altogether and actually seeing how they react to one another is really, really important. So, that's sort of stuff that we've been looking at and we really want to take into account, so we really wanted to share that with the folks. It's not just sort of that linear, "Take this, and multiply by 3% every year," because it's just not going to happen.
ADAM: Yeah, no. I heard a quote from one of our data scientists, I'm sure they heard it from someone else. But it was, "All demand forecasts are wrong, some are just actually helpful," and so when you're projecting on anything, it's not going to be exactly right. But let's make sure that the inputs are at least justifiable and you can start to stitch that story together. And so, the last thing that I'll add on, you briefly mentioned it, is that there's been this notion of taking national trends and applying them to local level markets.
And I think that what we've seen in the data is that it doesn't materialize. Right? How patients in Southern California and how policy is handled in Southern California is wildly different than in the Northeast, or in the Midwest, or in a more rural city. And so, taking these incidents per thousands, and applying that same logic for a patient that's in New Jersey, to a patient that's in Montana, it's not going to work. And so, or it's going to work, it's just going to be incorrect. And so, what we're really encouraging and what I would encourage everyone that's watching to do is to make sure that they're thinking about their specific patient population.
And are they of a younger demographic? Are they a more affluent demographic? What are your patients like? What do they prefer? How do they like to seek out care? What are those trends coming? And yes, you absolutely have to have a really good baseline of current, existing data to tell you what's happening today. We don't know what's going to happen tomorrow, nobody does. We can start to predict. Now the further out you predict, the less of a confidence interval you're going to have, but there's certainly a lot of information out there you can apply and I think you hit the nail on the head, Jason. So, I just want to add that, but that's all I wanted to hit on today.
I think that it certainly is a scary time to be making really big investments. It certainly is a scary time to be trusting certain aspects of data without fully understanding those inputs that are going into that. And so, I think making sure you fully understand what assumptions are being made before trusting a model is critical to any success. With that, if you do have questions about forecasting, if you do have questions about what's happening in your market, please, let me and Jason know. We'd love to jump on a call. We'd love to answer in our video. We get a ton of questions from people direct messaging us in LinkedIn or dropping us an email, let us know and we'll be happy to address that with you and certainly hope that this was helpful. But yeah, stay safe out there. Continue to grow that hospital volume, and I look forward to talking to you next week. Thanks, guys!