In the healthcare industry, “interoperability” is more than just a technical term for the way electronic health record systems shuffle information back and forth.
It’s a concept that can – and should – be applied to the connections between organizations across the care continuum, from primary care and specialty services to inpatient admissions, long-term care, and especially behavioral healthcare.
Historically, many of these disciplines have operated with a certain sense of isolation. Hospitals and specialists treated acute problems and performed surgeries; family doctors dealt with everything else.
Communication between them was limited to the tedious faxing of hundreds of pages of records, which may or may not contain pertinent information in an easily accessible way.
Behavioral healthcare existed somewhere off to the side as an option for those who specifically sought it out.
Access was relatively limited to the lucky few fortunate enough to have the right health insurance or independent means to pay for it, and primary care providers did not generally focus on how a patient’s mental health may be contributing to their physical ailments.
The delineation between physical and mental healthcare was clear and, due to the nature of fee-for-service reimbursement, perhaps unavoidable.
But with the advent of technical interoperability, the seismic shift in payment mechanisms that now reward care coordination, and a growing body of research that reinforces the notion that a patient’s mental health does not exist in a vacuum, the landscape is quickly changing.
The provider community is now embracing the notion that there is, in fact, little distinction between the realms of mental health and physical wellbeing, and that the delivery of effective population health management and truly accountable care must be predicated upon the integration and interoperability of both types of services.
In New York City, the nation’s largest public health information exchange (HIE) is helping behavioral healthcare providers transform this new perspective from theory into action. The first step is bringing clinical and behavioral health data together into the same patient record.
Healthix, one of New York State’s eight health information exchange organizations, covers 16 million patient lives across the five boroughs and Long Island, with 61 hospitals among the 1300 facilities connected to its services.
Using a combination of predictive analytics, standardized alerts, risk stratification, and other important population health management techniques, Healthix is connecting primary care providers, hospitals, and other clinical organizations with behavioral health providers such as ICL, a non-profit network of clinics, shelters, and residential care facilities that serve a uniquely urban population.
“The interest around behavioral health is a wonderful, unique opportunity to dig deeper into data analytics and population health management, said Todd Rogow, Senior Vice President and Chief Information Officer at Healthix.
“A lot of organizations have already invested money in analytics, especially the large integrated delivery networks and hospital systems. But our focus is to provide predictive analytics, risk stratification, and population health management to the organizations that haven’t taken on these endeavors yet, whether because they’re too small or they simply don’t have the financial means to do it.”
Many behavioral health providers fall into one or both of those categories, and face significant challenges when it comes to bulking up their health IT infrastructure.
These organizations were largely excluded from collecting meaningful use incentive dollars for EHR adoption in the early years of the program, which left them without the basic tools required for participation in many data-driven care coordination activities.
Health information exchanges like Healthix are looking to give behavioral health providers a boost in that department, Rogow said.
“We want all our participants to be able to get the 360-degree view of the patient that is required to manage their populations. It’s our job to collect data across the continuum of care – and across our communities – and turn that into a service that they can use to drive value and better outcomes as they tackle the challenges of healthcare reform.”
“Organizations have to focus on cost reduction and care quality, not on expanding their networks to gather more data. The value of a health information exchange is that we take on the problem of collecting the data and making it actionable for them.”
Defining the value proposition for health information exchange
Public and private health information exchange organizations haven’t exactly had an easy journey towards sustainability and success over the past decade or so.
While many innovative healthcare leaders clearly saw the need for some sort of infrastructure that would bridge the information gaps between disparate EHRs based on fragmented, proprietary technologies, the traditional fee-for-service reimbursement environment provided little incentive for organizations to invest their own hard-earned funds in an HIE membership.
Providers who weren’t getting paid based on how well they communicate with their colleagues across town watched from the sidelines as wave after wave of health information exchanges sputtered into life only to collapse a few years later under the weight of immature technologies, fierce competition, unsuccessful recruitment efforts, and financial mismanagement.
Yet at the same time, CMS and the Office of the National Coordinator was introducing some key new terms to the healthcare industry, including “interoperability,” “value-based reimbursement,” “population health management,” and “accountable care.”
Suddenly, health information exchange wasn’t just a nice-to-have extra that made it easier to taper off reliance on the fax machine, but a financial imperative that was quickly raising expectations about what health IT systems could and should do.
Clinicians have started to rely on HIE capabilities to make informed decisions about patient care, not only because better outcomes are starting to have a direct impact on revenue, but because they are committed to delivering the best possible results.
Health information exchange, whether through a public or private HIE organization, Direct secure messaging, EHR-to-EHR interoperability, or APIs utilizing emerging data standards such as FHIR, has been shown to reduce gaps in care, prevent duplicated services, enhance care coordination, and smooth transitions from the hospital to the home.
Making information available across the whole of the care continuum is so important to the larger goals of healthcare reform that the ONC has even made it a goal to connect the United States’ network of state and regional HIEs into a nationwide superhighway by the end of 2016.
Behavioral health providers are likely to be among the greatest beneficiaries of this ambitious challenge, especially as health information exchange organizations like Healthix move beyond the basics of making clinical information more available to its participants.
“We know that the biggest burden for people with severe mental illnesses isn’t always their mental health, but their physical health,” said Chris Copeland, Chief Operating Officer of ICL.
“That means we have to do a better job of understanding the impact that the two issues have on one another, especially when it comes to people with chronic physical health conditions that have been undiagnosed or have unrecognized psychological underpinnings to them.”
Because many behavioral health organizations have not been integrated into data exchange systems on the clinical side, much of the physical health data they collect is provided by their patients, who may not be able to accurately recall every diagnosis, treatment, or medication that may be relevant to their care.
“Much of our data is generated internally, and a lot of it is self-reported by patients,” said Copeland.
“Over the last few years, we’ve been able to look at data around emergency department use and hospitalizations, as well as physical health indicators like BMI and metrics related to diabetes. We also look at connections to primary care, medication adherence, and additional indicators that function as proxies for how well patients are doing with their healthcare in general, and if they are improving with us.”
“But what we don’t have right now is actual claims data,” he said. “We recognize the need to access real-time clinical data, because that is going to help us make sure that we are a data-driven organization and that we are maintaining a very high quality of service.”
“Partnering with Healthix is a very important step for that, because we can see who is going to the emergency room or who is in the hospital. That immediately helps us better coordinate care for our patients.”
Healthix distributes more than 245,000 continuity of care documents (CCDs) to providers and sends out 115,000 real-time admission, discharge and transfer (ADT) alerts per month. The HIE also generates notifications when a patient is incarcerated, which is particularly helpful for the behavioral health population.
The HIE is starting to fill the analytics gap that will enable ICL to take a more active role in population health management from their position as bridge between behavioral healthcare and the patient’s community.
“It is a change for us to begin looking at predictive analytics for our population,” Copeland said. “Healthix can already do some predictive modeling based on historical clinical data. What we’re really interested in – and where the synergy with Healthix begins to get exciting – is the ability to introduce behavioral health information into those algorithms.”
Moving beyond clinical data for actionable insights
“There is just a humungous amount of data out there,” stated Tim Trapanotto, Chief Information Officer of ICL. “One of the challenges we have faced is how to get timely data that wasn’t three or six months old by the time it got to us. Before we started working with Healthix, we would think about what we could do with hospital data that is accurate and up to date.”
“There are certainly a lot of options, but for a non-profit agency to have to build all of it on our own – well, that’s a big mountain to climb. Healthix not only provides us with excellent data, but they have the tools and resources to start looking into the social determinates of health, the community, and the support structures available to our patients.”
Incorporating these factors into the patient’s history can deliver more accurate risk scoring and give behavioral health providers a better idea of the comorbidities, socioeconomic situations, and physical health challenges that may impact adherence to a mental or behavioral healthcare program.
"If we can get in front of the other challenges that these patients are facing, then the cost of care is going to drop immensely.”
Sometimes the obstacles lie outside the traditional realm of the care continuum, and may not be apparent from the clinical record. But a little creativity on the analytics front can reveal some interesting opportunities for providers.
“We did some basic analytics with some appointment data that we had,” Trapanotto recalled. “When we looked at cancelled appointments, we noticed that the cancellation rates were similar for males and females until we got into the summer months. In June, July, and August, female patients showed a higher rate of cancelled appointments. Then when September rolled around again, the cancellation rates evened out.”
“Well, it didn’t take too much digging to find out that these ladies may have children who are out of school during the summer months, and they don’t have access to childcare that would allow them to keep their appointments,” he continued. “So maybe we need to partner up with a daycare center or provide childcare services at our location for patients when they’re going through their treatment.”
"Just something simple like that can keep these patients adherent to their programs without taking a break, which could set them back. If we can get in front of the other challenges that these patients are facing, then the cost of care is going to drop immensely.”
Healthix has the same goal for its services, Rogow said, and aims to enhance the analytics capabilities of providers like ICL, who have relied on their own resources in the past.
Like many analytics-as-a-service providers, the HIE is focused on making data actionable, relevant, and easy to access within a clinician’s often hectic workflow.
“Bringing value to providers as middleware is really important to us,” said Rogow. “As part of that, our job is to take all the data, including some analytical data and risk scores, and push it back into whatever EHR the organizations are using so it can get back into the clinical workflow.”
“The real-time nature of the data is generating a lot of positive feedback,” he added. “It’s very important to the organizations that we work with. And when it comes to analytics, having the ability to look at someone’s real-time risk score while they’re still in the hospital is so powerful, especially when it comes to transitions of care or preventing readmissions.”
Transitioning from information to intelligence
Clinicians may be thrilled to see a real-time risk score pop up on their laptop while treating a patient, but they may not know exactly how much technical back-end work goes into generating that information, especially when it comes to synthesizing, standardizing, and analyzing behavioral healthcare data.
Healthix has partnered with HBI Solutions to provide the backbone for its analytics capabilities, turning raw data into “clinical intelligence,” according to Eric Widen, the company’s CEO.
“As we intake the data, we normalize and standardize it so that we can use it to feed our clinical intelligence tools,” he explained.
“New York City and Long Island have a very unique population, relative to other parts of the country. It’s very dense, urban, and multiethnic, and there’s a significant number of patients with mental health needs. We help organizations hotspot those patients to prevent unnecessary disease development and reduce unnecessary utilization.”
"We may even find that we will have to reevaluate some of our hypotheses about how to best serve our communities.”
To accomplish that, HBI tailors its predictive modeling tools to the specific needs of the mental healthcare population, picking up on the characteristics that are most likely to drive up risk scores for emergency department visits or crises due to chronic diseases such as diabetes.
“These algorithms are very precise in terms of the local population, so ICL can use them to flag patients before they have these costly events, or before their diabetes gets out of control,” said Widen. The algorithms incorporate socioeconomic and community-based issues, such as housing insecurity or access to transportation.
These factors are not always captured in the clinical record – and when they are, it is usually as part of a free-text clinical note instead of a standardized data field in the EHR, Widen added.
“So as a part of the underpinning for our analytics, we use tools like natural language processing, which help identify data that is not necessarily captured discretely off the health information exchange transaction system. Natural language processing is key to helping us include those free-text indicators of need.”
“Using these types of tools, I think we will be able to demonstrate the different factors that are driving risk for specific populations, and we may even find that we will have to reevaluate some of our hypotheses about how to best serve our communities.”
From behavioral healthcare to better overall healthcare
Healthix and ICL are learning a great deal about how to deliver effective population health management services to behavioral healthcare patients, but the impact of these lessons is unlikely to stop there.
“This is a project that is truly going to help the mental health population in New York City, and will likely have wider applications for other patient groups,” Widen predicted.
“Understanding how mental and behavioral health characteristics drive the risk of disease, utilization, and costs is going to be applicable to the entire HIE. As we dig deeper into the data, we’re going to find new information that we haven’t found before.”
These new insights are likely to strengthen the notion that the type of care that will help providers tackle the value-based reimbursement revolution must be integrated, comprehensive, patient-centered, and community-driven, Copeland said.
"I hope that within a reasonable space of time, we won’t even be talking about those as separate entities. We will just be talking about healthcare.”
“With the massive amount of reorganization that’s going on with managed care, it’s important that we try to reduce costs. But we’re only going to be able to bend the cost curve when we realize the relevance of psychological and mental health issues to general clinical care,” he asserted.
“Proving that out is going require a major change in thinking about the place of behavioral healthcare in the larger care continuum. I hope that within a reasonable space of time, we won’t even be talking about those as separate entities. We will just be talking about healthcare.”
“Using health information exchanges and analytics will be a critical part of putting these pieces together to create truly interoperable, integrated care for patients.”