{{item.videoDuration}}
{{item.title}}
{{item.text}}
{{item.videoDuration}}
{{item.title}}
{{item.text}}
How payers are mastering advanced analytics to enhance patient outcomes
To listen to all PwC Next in Health podcasts, click here. Subscribe and listen to all episodes at your convenience via any device at Apple Podcasts and Spotify.
Overview
Tune in to hear PwC specialists discuss how advanced analytics are helping health payers use new tools to forecast and improve outcomes, experience and costs. Topics include:
- Improving affordability, quality and experience through more timely and actionable forecasting
- Data science and technology to make it possible
- Moving to implementation
Topics: analytics, machine learning, AI, payers, health, health care, affordability, insurers, technology, healthcare, patients, predictive analytics, transformation, upskilling, infrastructure, health insurance companies, providers
Episode transcript
Find episode transcript below.
00:00:04:20 Welcome to Next in Health podcast, I'm Jenny Colapietro PwC’s Vice Chair for Health Industries working across Pharmaceuticals, MedTech, Payers and Providers.
IGOR BELOKRINITSKY:
00:00:14:11 And I'm Igor Belokrinitsky, a Principal with PwC Strategy&, where I help leading health organizations with their strategies and operating models.
00:00:23:24 And Jenny, when I look back to the beginning of this year, which feels like a long time ago, we talked about some of the major issues that our health organization is going to be dealing with, and those included things like confronting affordability and disrupting costs of health care, attracting and retaining new customers and rethinking risk.
00:00:45:16 And as I think about the full range of these issues, they all have one thing in common. That one thing is mastering analytics really helps you make progress in addressing those issues, helps you get a handle on those issues and compete and differentiate and help address those issues. And that's why we wanted to spend more time talking about analytics.
00:01:07:10 And today our topic is advanced analytics, specifically for payers, for health insurance companies. To cover this topic, we have two great guests with us today.
00:01:17:10 We have Sachin Bajpai, who is a Director in Healthcare Data and Analytics. And we also have John Kiley, who's a Senior Manager and a Data Scientist. So we're excited to have him. And so Sachin and John, welcome to the podcast.
SACHIN BAJPAI:
00:01:31:06 Thanks, Igor.
IGOR BELOKRINITSKY:
00:01:32:07 It is wonderful to have you both. And so let's dive in and maybe let's start with why. So Sachin, go to you first, as you develop advanced analytics as health organizations out there building their analytics capability, what problem are they trying to solve with these analytics capabilities?
SACHIN BAJPAI:
00:01:51:13 I've been doing this for more than 15 years now, and the needle has really shifted in the last two or three years the technology has matured and the organizations are genuinely going data first. From the payer analytics space especially, we are seeing action in three areas. First one, improving healthcare experience.
00:02:10:13 You talk to any healthcare executive and this would sound very, very trivial, but research now shows that the patients want to be treated where they are most comfortable, which is their home, right, versus these acquired facilities where they may be left by themselves with machines attached to them.
00:02:27:01 So that's where we have been developing some algorithms and advanced analytics solutions in the LTSS market that can predict patients that are the risk of getting transferred to acute facility. So that's one. Second one, complementing to that is the cost of care and affordability, right? There have been so many algorithms in the market that do the admission for you, right?
00:02:48:23 But they you know, we have taken a more holistic angle to it and we think that is where the market should go, is creating a complementing package of readmissions and potentially preventable admissions by disease conditions, right? So that's what we are seeing more and more required in the market.
00:03:07:23 And then the third one is the shifting Medicare and Medicaid landscape and redetermination. All the subsidies and government benefit programs that were extended during COVID are about to expire. In the Medicaid world alone, there are like 10 to 15% of members who are at the risk of being excluded from some of these programs.
00:03:25:19 So developing a foolproof algorithm that can find the right population for your determination so that the payer companies can rightsize their campaigns. So those three have been the primary three things that I've been seeing in market. It's a lot of action on advanced analytics.
JENNY COLAPIETRO:
00:03:41:25 Thanks, Sachin. Yeah, that's helpful to understand the problems that we're trying to solve for. John, can you just share a little bit what is the ‘advanced’ part of ‘advanced analytics’ referring to?
JOHN KILEY:
00:03:53:25 Yeah, absolutely. Thanks for having me on here today, everybody, by the way. So advanced analytics in this context is referring to more I want to maybe say, sophisticated techniques than some of the traditional reporting that's been done in the past. A lot of these companies, like Sachin was saying, have gone through the data journey, right?
00:04:17:00 They've gone through, they've collected all of this useful information and they're trying to use it to help improve health outcomes, improve cost, care, affordability, so on and so forth.
00:04:25:23 But after you've collected all that data and you've got all this information, what can you do with it? There is a couple options, right? You can take a look at it, do some reporting, do some descriptive analytics and say, Here's what I have seen happen in the past where it gets really more powerful or what makes it advanced is when you're able to take that information and convert it into predictive models or machine learning in a lot of cases here.
00:04:47:05 So when you're using machine learning and predictive analytics, you can take the ability to use all that information that you've collected and manage your population at scale, your populations, you know, 5 million, 10 million, 20 million people, sometimes smaller, sometimes bigger, right? And you don't want to bring in 100,000 people to help manage that population, right?
00:05:07:05 You want to keep your team lean. You want to keep your team efficient, and you want to be able to take advantage of all of the information that you've collected. And that's where advanced analytics can slide in and help you improve that.
IGOR BELOKRINITSKY:
00:05:16:19 John, this is really helpful and I'm kind of picturing right now this box that says advanced analytics on it. And what you're telling me is amazing because it can solve all of these challenges and it can be more predictive and more useful and more timely. But can we just kind of open the box and see what's in it?
00:05:36:07 What are some of those advanced analytics if it is tools, algorithms, models, like how does it work, what kind of technologies and I guess methodologies are you able to use to produce this much greater performance?
JOHN KILEY:
00:05:49:19 Absolutely. So the way that I like to think about it is we're standing on the shoulders of giants in a way. We're taking advantage of all of the research that's been performed in academia and industry over the years and converting all of that collective human knowledge into applications that could be used to help support your peers. It all starts with your use case.
00:06:12:12 So when you take a look at your business and you say, I've got this problem that doesn't seem to scale, right? Or it seems like it could be enhanced or improved by data and analytics, this is where we take a look and say what algorithm or what kind of data science technique can we use to rightsize the solution to the problem.
00:06:30:27 So when we open up the box, that is the predictive analytics field and predictive analytics space, we then take a look first at the problem and say: we want to use deep learning to address something that's maybe this complex or at this scale. deep learning all will be really, really powerful for when you have a ton of data available to you or you're going into something like computer vision or natural language processing, or you're doing something very complex, heavy-duty.
00:06:57:27 Traditional machine learning is going to be really good when you want to try to do something like a classification problem where you say, “In my population, I have some folks that I need to label as in need of support, or this person does not need support.”
00:07:09:11 And so we can use machine learning methods to parse through the entire population and label the individual as someone who may need some additional services to help them get the care that they need to stay happy, healthy, and maybe in their home if they're at risk of transitioning to a more acute care setting.
00:07:26:25 Like Sachin was saying earlier, folks like being in their home. On the regression side, maybe you have something with payment integrity or let's say you want to try to estimate what is the expected cost of a particular service. While a regression problem will allow you to estimate the numerical value for a particular problem. So in the event of like the payment integrity, you can say I expect this service to cost $2,000.
00:07:52:23 The actual request is come in for five. This is a huge variance. What can I do about it? And then as the business you're able to take those insights and transform, alter outcomes, and those are just a couple of the more established practices as your traditional machine learning, a deep learning. But now we're seeing all of this space open up into generative AI.
00:08:11:06 So I would say like generative AI's kind of a specific application of deep learning. It's a little bit more focused. It's a slightly different family of techniques that are being used within the deep learning umbrella. But it's transforming and it's a hot-button item, right? It's transforming a lot of industries these days. You're hearing everywhere, right? All these other LLMs that are out there, there's different opportunities to plug these types of advanced analytics and extensive capabilities into your workflows as well.
JENNY COLAPIETRO:
00:08:39:08 John, this is helpful. To put this into more context for our audience. How do we make this actionable and actually put this advanced analytics into play?
JOHN KILEY:
00:08:47:25 Yeah, and this is the secret sauce, right? So we follow a nine-step process when we're developing models and solving advanced analytics, use cases where we'll go through and break things down from. It starts with always the business understanding and breaking down the use case into its component parts. Some people might call this requirement gathering, scoping, but really that front end of the workflow is always the most important thing to make sure that you're going to get the business value that you want when you ultimately push this model all the way to production.
00:09:18:16 The next couple of steps in our nine-step process are around actually creating and developing the model and training and testing and going through that traditional data science workflow that all of us you know get to learn about in school. That's where the bulk of the true data science work happens. And then probably one of the most important parts comes after that modeling section occurs and that's going to be where you're taking your models out to production.
00:09:44:26 And this is where you're again, you've got a working model. It's gone through all of your training, your testing cycles, you've validated it with the business. Now you want to take it and you want to integrate it into your workflows and make sure that it's in the hands of the professionals that are going to ultimately take the outputs and apply it to the actual business problem.
00:10:00:25 So you can actually generate your business value and start seeing results in action. And a lot of organizations will struggle with that too. So it's a challenge, but it's the most important one. I would say.
IGOR BELOKRINITSKY:
00:10:12:25 Thanks, John. And maybe on that note, we'll turn it to Sachin to follow up on the comments that you're making around the challenges of implementing this in an organization. What kind of infrastructure would an organization put in place to make sure that it's taking advantage of all these newest tools and getting information that is useful, that is actionable, that is complete, while also managing all the risks that I'm sure could arise while all doing this. So Sachin, any guidance and advice to an organization that is looking to take advantage of advanced analytics.
SACHIN BAJPAI:
00:10:49:01 Absolutely. And upskilling never stops. Given how much action there is in the market right now? I mean, you know, there is a huge, huge potential, you know, establish these technologies and benefit from them. First and foremost, I would say like, companies out there should stop treating this as a lab experiment. Many times we have seen that, you know, there would be a couple of data scientists with a couple of engineers will be working on these experiments and they never scale, right?
00:11:16:02 So first one is that discipline, that rigor of making it industrial with ML Ops, that should be the first focus. And then the second one is have an API-based integration with these central business applications that score because data science world is not a fancy word. At the end of the day, people like John and I, we produce a value.
00:11:39:12 It's a propensity of something, it's a prediction or something, right? It is not fancy at all, but the value is like how soon this information goes back into the primary application where the agents, the field agents, they can benefit from it. So the two parts, making the ML journey more industrial with the adoption of ML OPs and second is disseminating these insights via APIs back into the primary application.
00:12:06:17 That is the taking the journey forward piece, Igor, like you are saying. And then in terms of upskilling, I think it's we live in a world where the knowledge has to be seamless and it is no more like, okay business holds this information, IT holds this information, and data scientists only need to know that much.
00:12:24:13 Those times are gone. We need a world where the frequent flow of information needs to be there through lunch and learns, through webinars, through town halls. This is the time where the frequent flow of information will benefit everybody.
IGOR BELOKRINITSKY:
00:12:39:02 Sachin and John, this is a really helpful look into the world of advanced analytics. And just to recap what you've told us, it sounds like for health payers, for health insurance companies, because of all of these new tools, there's an ability to forecast more accurately, forecast clinical events, forecast budgets, forecast spending, maybe forecast member behavior, and potentially to prevent bad things from happening bad health outcomes, unnecessary spending, things that negatively impact member experience.
00:13:15:01 So it's worth doing for the members, for the providers, for the insurance company itself, but it's worth doing right, which means continuing to learn about the potential of these technologies, continuing to take advantage of that potential and continuing to integrate these technologies into other systems so that when these insights and forecasts are generated, they can be made actionable and turn into specific steps that a health insurer would do to create a better outcome and prevent the bad ones.
00:13:47:17 So quite a powerful set of tools if used right. And really appreciate you talking us through it. So thanks for joining us and giving us your insights.
SACHIN BAJPAIN:
00:13:56:29 Thanks, Igor.
JOHN KILEY:
00:13:58:06 Thanks for having us on.
IGOR BELOKRINITSKY:
00:14:00:25 For more on these topics and other health industry insights driven by policy, innovation and care delivery changes, please subscribe to our podcast and be sure to listen to prior episodes as well. Until next time, this has been Next in Health.
ANNOUNCER:
00:14:19:22 This podcast is brought to you by PwC. All rights reserved. PwC refers to the U.S. member firm or one of its subsidiaries or affiliates and may sometimes refer to the PwC Network. Each member firm is a separate legal entity.
00:14:32:05 Please see www.pwc.com/structure for further details. This podcast is for general information purposes only and should not be used as a substitute for consultation with professional advisors.
{{filterContent.facetedTitle}}
{{contentList.loadingText}}
{{contentList.loadingText}}
Contact us
Follow us
