Season 2 Episode 2: Minutes to Feasibility in Long-Term Care Research

Andie Cartwright

Welcome back to the Better Living Through Data podcast! I'm Andie here with Anthony, and I want to start with what feels backwards: maybe the biggest innovation in long-term care research is not MORE data. It's less waiting. Less wrangling. Less of that horrible stretch where a smart research question disappears into analyst review for, what, weeks... sometimes months.

Anthony Pero

Months is the right word. And the painful part is, long-term care has never really lacked richness. It's lacked speed to insight. You can have clinically dense data, deeply longitudinal data, but if every feasibility check becomes a custom project, then the bottleneck isn't the science. It's the workflow.

Andie Cartwright

Ok so "feasibility check" is one of those phrases people nod at and maybe don't fully unpack. When you say that, you mean the very first, very practical question: do we even have enough of the right residents, with the right characteristics, to study this properly?

Anthony Pero

Exactly. Can I identify the population? Can I define the cohort? Are the variables there? Do they exist consistently enough to support the study? Before a researcher gets to insight, there's usually this long pregame of data translation. And that's where so much time burns off.

Andie Cartwright

And that's the part people outside research don't see. They imagine access as the hurdle -- like once the dataset is "available," you're off to the races. But really, access is just the front door. Then you've got all the unpacking: who belongs in the cohort, what counts as an exposure, which variables are useful versus technically present but messy.

Anthony Pero

Right, and in long-term care, those choices matter a lot because the residents are medically complex, the stays can be extended, and the questions are nuanced. So if you're studying medication use in dementia, or treatment patterns in frail older adults, the manual work is not trivial. It's a lot of back-and-forth just to get to a clean first pass.

Andie Cartwright

Which, from my marketing brain, sounds like the classic hidden cost. Not the dataset itself -- the labor around the dataset. The meetings, the revisions, the "actually can we redefine this cohort?" email chains. That's the stuff that quietly turns one research idea into a six-week calendar problem.

Anthony Pero

And six weeks is generous sometimes. The point here isn't that analysts are the problem. They're essential. It's that highly trained people end up spending too much time on repetitive setup work instead of the higher-order work: interpretation, challenge-testing, study design, academic rigor.

Andie Cartwright

So the real headline is almost boring on paper but huge in practice: if you can compress feasibility assessment, cohort building, and variable clarification from months to minutes, you've changed the economics of asking the question in the first place.

Anthony Pero

Yes. And in academia, changing the economics of a question changes which questions ever get asked.

Anthony Pero

That brings us to Study Buddy. What stands out to me is not just that it's AI-powered. Plenty of things claim that now. It's that it's designed to turn natural-language research questions into credible outputs in minutes. So instead of starting with the data model, a researcher starts with the actual question they care about.

Andie Cartwright

And the phrase "in minutes" is what we keep circling. Because if the first draft arrives in minutes, then academic exploration gets looser in a good way. You can test an idea, kill it quickly, refine it, come back with a better version -- without treating every exploratory question like a major production.

Anthony Pero

Exactly. Faster feasibility assessment means you learn sooner whether a study is viable. Faster hypothesis generation means you can explore more intelligently. And for grants, this is the piece I think people will really feel: getting to grant-submission-ready evidence faster changes how competitive and how confident a submission can be.

Andie Cartwright

Grant-submission-ready evidence. That's a very specific promise. Not "here's a cool dashboard," not "here's some AI sparkle." You're talking about outputs that can actually support the work academics already have to produce.

Anthony Pero

Yes, and that's where the purpose-built part matters. Study Buddy isn't a generic AI research assistant. It's built for long-term care, and the sample queries are designed and validated by PointClickCare Life Sciences experts using structured prompt engineering. That matters because consistency matters. Researchers need outputs they can trust enough to work from.

Andie Cartwright

And one of those outputs is table shells, which I love because it's so gloriously practical. Table shells are not flashy. They're useful. They're the kind of thing you can put straight into a manuscript draft, a poster outline, or a grant application and say, okay, now we're actually moving.

Anthony Pero

"Gloriously practical" is exactly right. And that's also the right place to push back on the usual AI anxiety. This is not replacing academic rigor. It's not replacing epidemiology, biostatistics, or peer review. It's compressing the low-value, repetitive setup work so the humans can spend more time where the value really is.

Andie Cartwright

Let me try to say that back. So the AI isn't writing the science for you. It's shortening the distance between "I have a question" and "I have something structured enough to critique." Did I get that right?

Anthony Pero

That's exactly it. It augments judgment; it doesn't substitute for judgment. Researchers still have to interpret findings, pressure-test assumptions, define endpoints carefully, and meet academic and regulatory standards. But if the first usable draft shows up in ten minutes instead of ten weeks... the whole rhythm of research changes.

Andie Cartwright

And honestly, from a creative strategy angle, first drafts matter. A bad blank page is worse than an imperfect starting point. Once there's something on the page -- a cohort outline, a variable set, a table shell -- smart people can make it better. But they need something to react to.

Andie Cartwright

Okay, but this only works if the underlying data is actually deep enough to support serious questions. And this is where the numbers get sticky in the best way: 50-plus structured data points per resident per day. Per day. Across extended stays. That's not a thin claims snapshot; that's a much more textured picture.

Anthony Pero

And that texture is backed by scale. We're talking about a connected ecosystem of more than 30,000 providers, more than 400 integration partners, and every major U.S. health plan. So when researchers use this long-term care dataset, they're not looking through a keyhole. They're working with one of healthcare's most underused but clinically rich sources of real-world evidence.

Andie Cartwright

Thirty thousand providers is the token I'm stuck on. That's not just "big." That's the difference between hoping your question is measurable and having real confidence you can find the population, the timing, the patterns.

Anthony Pero

Yes, and it lowers another barrier people underestimate: specialized technical or regulatory expertise. The data is deidentified in accordance with HIPAA, and by designing the platform around the research question instead of forcing people to wrestle the underlying data model, it becomes more accessible to teams across academia, government, biopharma, health economics.

Andie Cartwright

Which means the practical payoff is kind of beautiful in its simplicity. Less time waiting on analysts. More time refining the question. More time validating whether your assumptions are nonsense. More time figuring out if the story your study wants to tell is actually true.

Anthony Pero

And more time moving toward publication or funding. That's the part I don't want to understate. When review shrinks to a fraction of what it was, academic teams can iterate more. They can discard weak ideas sooner and strengthen good ones sooner. That accelerates discovery without lowering the bar.

Andie Cartwright

I keep thinking about all the questions that die quietly because they're too slow or too expensive to explore. Not bad questions -- just inconvenient ones. Questions that need one more pass, one more feasibility run, one more cohort definition... and then the semester ends, or the grant window closes.

Anthony Pero

That's the bigger implication for me too. If the first pass of long-term care research becomes nearly instant, then researchers aren't forced to be quite so conservative. They can pursue narrower questions, more novel questions, maybe even riskier questions, because the cost of finding out is lower.

Andie Cartwright

Hey Anthony, one cool thing I wanted to mention is that we’re offering free access to Study Buddy! Academic researchers and the like can sign up now!

Anthony Pero

That’s right Andie! Listeners can click the link in the description for today’s episode and learn more about Study Buddy and sign up for free access!

Andie Cartwright

Very cool, we hope to see some sign ups! We’ll see you next time on the Better Living Through Data podcast!

Anthony Pero

Thanks for listening! Bye everyone!