- As the principles of value-based care become increasingly important for success in the healthcare delivery environment, provider organizations are facing some tough choices around how much financial risk they are prepared to take.
A commitment to a value-based contract, even a limited one, requires organizations to address a host of challenges related to clinical quality, utilization costs, benchmarking and measurement, preventive care, and patient engagement.
If the organization fails to develop the right population health management skills, they could be exposing themselves to financial losses that may breed skepticism about long-term plans to expand participation in risk-based reimbursements.
Cultivating a culture of success requires providers to identify their financial and clinical goals, plan carefully for how to achieve them, and leverage a set of foundational data analytics strategies that can help get ahead of the cost curve without breaking the bank.
Setting the stage for organizational change
“The first important step is to identify where your group resides on the continuum from free-for-service to full value-based reimbursement,” advised Dr. Betty Rabinowitz, Senior Vice President of Solutions at NextGen Healthcare during the Value-Based Care Summit: Population Health event hosted by Xtelligent Media in Dallas on April 5-6, 2018.
Most organizations are not progressing linearly through a well-defined transition towards alternative payment models, she said, and may not have well-defined plans to hit certain financial thresholds by a particular date.
Instead, many providers are beginning with a handful of pilots or preliminary contracts limited to certain high-risk populations.
While this strategy can help organizations develop best practices and identify opportunities for improvement, it can also make it difficult to understand how to invest in analytics tools and strategies that will scale to meet future needs.
“As soon as you start contemplating risk and taking on a value-based contract, you are going to find the need to identify patients who need enriched services or additional care management,” said Rabinowitz.
“The trick is figuring out how to use data analytics to support the stage you’re at while beginning to prepare for the next stage you’re going to. It takes time to get the data you need, and it takes time to learn how to use it. There is a little bit of finesse required so that you’re not boiling the ocean before you need a full ocean of boiling water.”
Finding that balance requires a provider group to understand its overarching vision before making investments in technologies or processes that might not help achieve the ultimate goal, added Dr. Mahek Shah, Senior Researcher at Harvard Business School.
“The value-based care train has left the station and it’s moving fast,” he cautioned. “We all know that the status quo is not sustainable. But before you start rushing to make changes, you need to identify where your leadership’s mindset is and you should work to find champions that are willing to own every part of that vision towards value.”
“Find opportunities that are both meaningful and manageable. And involve your physicians and other clinical experts as early in the process as possible – that can help to sustain a lot of the initial investments you’ll need to make.”
Those investments may not necessarily be technology-driven – at least not at first, said Joe Denney, RN, Chief Health Information Officer at Oklahoma Primary Care Association.
While it may be tempting to jump into a large-scale acquisition of the latest and greatest big data analytics tools, buying up infrastructure without a clear sense of how to apply it to a concrete problem can lead to trouble down the road.
“Don’t get wrapped up in the idea that you have to create a huge data lake right off the bat,” said Denney. “You really don’t need to drop hundreds of thousands of dollars on technology before you can start to do really important work that isn’t being done yet.”
“If you start by building a data lake or enterprise data warehouse and then try to figure out the problem you want to solve, then you’re coming from it from the opposite direction. Instead, I encourage folks to find a small problem, define that problem precisely, then figure out what data you need to make your solution happen.”
Defining the scope of a population health management problem
Many value-based care contracts require organizations to keep patients out of the emergency room or the inpatient setting by getting upstream of chronic diseases that can snowball into costly acute events.
There is a growing recognition that managing patients holistically is the best way to maintain their wellness – but taking a comprehensive approach to care management can leave organizations wondering where to start and how to define the parameters of their investments.
“There are as many care management models as there are healthcare organizations,” said Rabinowitz. “But they all face the problem of identifying the patients who will benefit most from interventions.”
Rabinowitz, an internal medicine physician, faced that exact conundrum during her time in practice at the University of Rochester, she explained. The organization wanted to improve care management for diabetics, but initially struggled to target effective interventions to the right group of patients within that category.
“We had these excellent, high-powered RNs who were assigned to a group of diabetic patients some with an A1C of around 7 or 7.2,” she said. “We set up cooking classes and grocery shopping tutorials, and the patients were delighted. They would come to the classes and they enjoyed the social connection and engagement – some were patients living with loneliness, and the programs filled an important social need for them.”
“That is certainly important…but it didn’t change their diabetic outcomes. And it didn’t improve our performance in our first value-based contract. We quickly realized that if we wanted to see more specific improvements in certain measures, we would have to identify narrower targets among the patients most at risk for the events that were included in the contract and we were being measured on.”
Defining the problem clearly can help to narrow down a pool of patients who are most likely to benefit from available interventions, added Denney.
“A lot of organizations want to reduce no-shows, for example,” he said. “That crosses both clinical and operational lines, and it can be difficult to get your arms around a problem like that.”
“You need to drill down and be very specific. Instead of saying ‘let’s reduce no-shows,’ try something like ‘let’s reduce no-shows for our pediatric, non-English speaking patients from 30 percent to 20 percent within 90 days.’ That’s a well-crafted goal that gives you something to work with.”
Just like Shah suggested, Denney believes that involving front-line staff is a key component of any quality improvement goal.
“Talk to the people who are actually interacting with the patients who aren’t showing up,” he urged, such as a front desk staff, the nurses, and the medical assistants who may understand the patient and family histories.
“They don’t need to be experts in data science, but they’re the ones who hear the reasons why the appointments are being cancelled, and chances are they have a very good idea about what’s going on. They will add a ton of value to your project at almost no cost.”
After using the input to identify possible areas of focus, project leaders can start looking at the data behind those regions of concern.
“There are some things that are going to jump out as obvious,” said Denney.
“I’m going to give you $10,000 worth of free data science consulting right now: the people who don’t show up previously are more likely to not show up in the future,” he said, drawing a laugh from the audience. “There you go – it’s that simple.”
“The point of using data analytics is really to find those things that are not obvious and maybe not intuitive,” he continued. “Look and see if there’s a day of the week or a time of the appointment – maybe there’s a length of time between when the appointment is made and when it’s scheduled to occur that shows some spike in no-shows. Then you can work with your team to develop some interventions.”
Segmenting the population allows organizations to tailor interventions to smaller groups of patients with commonalities, explained Denney – and also allows project leaders to quantify the results of their efforts.
“That’s how you can build credibility with stakeholders, including physicians and people in the C-suite,” he pointed out. “If you can have a pilot that shows a little bit of value, that’s a whole lot more impressive than saying, ‘Oh look, we dropped a quarter of a million on a data lake, but now we don’t know what to do with it.’ You can’t be successful until you identify the problem.”
Ensuring a scalable, sustainable health IT strategy
All three panelists agreed that organizations need to clearly define their problem statements when attacking a population health management problem, but organizations have innumerable options for how to apply health IT tools to the questions of value-based care.
Denney, who works primarily with smaller providers in rural settings and federally qualified health centers (FQHCs) with limited resources, encouraged organizations to “start anywhere,” even if that means relying on Microsoft Excel, that ubiquitous office tool.
“For a lot of organizations, especially if you only have one or two experts to handle all your data analytics needs, you need to get those small-scale pilots going so you can demonstrate the value of future investment,” he said.
“If you can start simple, even with an Excel spreadsheet, then iterate quickly and prove the ROI of what you’re doing, that gives you a strong sales pitch to take to the board. It can be very powerful to say, ‘Look at what we did using Excel and not much more – we saved $100,000. We think that if we scaled this over 50 projects and maybe added some new tools, we could save $15 million.’”
“That’s when your executive board can start considering whether or not to drop a million dollars on an infrastructure build-out. It doesn’t always make sense for smaller organizations to do that infrastructure development first. There often isn’t the budget for it.”
Rabinowitz agreed that starting somewhere with a proof of concept project is important, but urged organizations not to rely too heavily on piecemeal technical development.
“I’ve lived with an Excel spreadsheet – that’s where we started as a clinical group at Rochester. The disadvantage of that approach is that you can very easily recreate siloes in your data that make it almost impossible to benchmark, compare, and grow. It’s a difficult approach to scale.”
“As the data becomes more varied, complex, voluminous, and challenging in terms of its veracity and accuracy, you will quickly outgrow the spreadsheet.”
As an analytics developer herself and a representative of a well-known health IT vendor, Rabinowitz has seen a number of provider groups fail to scale their homegrown technology tools appropriately.
“You need to create the space and the scaffolding to structure your data in a way that is scalable,” she said. “You can answer some questions with simple tools, but the tools that are required for more advanced analytics should be delegated to pros.”
Dr. Shah sees merits in both approaches. “As a former investment banker, I can say that I love Excel,” he joked. “That industry seems to be operating just fine with it.”
“In healthcare, however, you need to make sure that whatever tools you’re using are part of a larger strategic plan. You also need to be certain that you still maintain the capacity to innovate even when you’re implementing a new infrastructure tool kit.”
“I’ve seen big organizations on the verge of bringing in a brand new major EHR system, and all innovation just freezes up while the whole team is devoted to that project. A clear organizational vision of how to continue delivering value to patients will prevent that.”
No matter what tools are brought on board, organizations must remember to keep their value-based care goals front and center, he continued.
“You can spend a lot of time in the weeds of how to slice and dice your big data, but you have to remember to make it meaningful and actionable,” he asserted.
“If you don’t do those things, then no amount of investment is going to give you a return. It will just sit in a corner and collect dust, and you won’t be able to scale it. Make sure your analytics technologies are part of a comprehensive plan that is both detailed and forward-thinking.”