Eric Widen is co-founder and CEO of HBI Solutions of Palo Alto, CA.
Tell me about yourself and the company.
I’ve worked in healthcare my entire career. I’ve had an eclectic mix of experience working for consultancies, electronic health record vendors, for myself for a period of time, and for providers. All with a focus on implementing technology to drive improvement, from a health system standpoint and now more so from a population standpoint.
The theme has always been around using data that’s inherent in these systems to help drive performance improvement. We founded the company on that concept of helping health systems and organizations take advantage of data to improve their performance, Specifically to improve population health approaches by leveraging data that’s mostly residing in electronic health records, which have become more ubiquitous over the last 10 or 15 years.
How do you position the company among the many that offer analytics and population health management technology?
Population health, analytics, and even predictive modeling are broad-based terms and topics. Many vendors are saying similar things.
Where we differentiate is that we’re not a platform company. We’re very much a focused solution that we term a precision health solution or precision medicine solution that’s leveraging real-time predictive models that are proprietary intellectual property that we’ve developed. These are our own real-time predictive models that we provide that drive our precision health solution. That’s a niche focus.
We’re technology platform agnostic. We see this as an important piece to identify people at risk for untoward events before those events happen. In real time, meaning leveraging electronic health record data to do that in order to keep people healthy and from creeping up the disease and cost curve over time. That engine that we built can be installed in many different types of platforms. We think it’s an important piece of the puzzle.
Population health includes analytics, care management to take care of the patients, and the interventions that are going to be applied to patients. Our focus is in real time identifying people at risk for poor outcomes before they happen and then identifying the interventions to apply to those patients in order to mitigate the risk from ever happening.
That engine is what we provide. It can be deployed on many different types of platforms, including interoperability system platforms or EMR platforms. Those two examples of interoperability solution vendors and electronic health record vendors also pitch that they do population health. They provide the platform to do that. Very few organizations are providing the specific engine that we provide.
Are providers becoming willing and able to intervene when their patients are flagged as high risk?
What happens on the provider side today is that they’re balancing multiple incentive structures. They’ve dipped their toe in the water. What we’re seeing is 10, maybe 20 percent of the health system’s population is under a new payment mechanism, meaning at risk and/or upside gain for populations. But they’re still balancing the fee-for-service methodology at the same time. These are schizophrenic conversations. Everyone agrees that future is coming and that taking care of patients and keeping them healthy is going to be the new care model going forward, but they’re not there yet.
Organizations are confused about the speed of when that’s going to happen and it freezes decision making a little bit. Organizations are being successful with the experiments in taking care of patients proactively to keep them healthy in order to make financial gain under these new payment mechanisms. Where they can carve out those patient populations and apply these methods, they’ll restructure their care management processes to do that.
They’re really struggling with that decision when and how to do that. We see them doing it well where the incentives are aligned and there is a service component to that to help them rewire their care management processes to think differently about taking care of patients pre-disease or taking them from an at-risk standpoint as opposed to post-disease, which has been the old care model.
Is it an ethical struggle for providers who are beginning to see the value of providing population health management but realize that it could cannibalize their incomes if they do it or everyone, including those for whom they’re being paid fee-for-service?
I don’t think it’s an ethical struggle. It’s a clear problem to solve. It gets back to the acceleration of when are these going to come in full force.
We have clients that have done exactly that. They’ve done such a good job at using our solution to target patients at risk, keep them out of the emergency room, keep them out of the inpatient setting, keep them on the right care programs to mitigate disease progression, whereby they have reduced admissions and volume to their hospital. They’ve had a struggle with that.
What they’ve said is that this is the right thing to do for the patients at the end of the day, to keep them healthy and out of the acute care settings. What they’re looking to do is figure out how to accelerate taking on more incentive-based contracts and risk-based contracts in order to keep this going.
I don’t think it’s unethical. They had upfront conversations about it and they’re trying to figure out strategically how to continue to navigate this process. All of the organizations we’ve talked realize it’s coming and they’re willing to prepare for it. It’s just a matter of speed.
Providers can’t just unilaterally reach out to a high-risk patient and tell them what to do. Is it a marketing challenge as well as a clinical challenge to get patients engaged in this process that’s new to them?
Disengagement from a patient standpoint is a continuous problem for care managers. The ability to engage the non-engageable is a never-ending problem for the care management folks.
What we’re seeing and what we think is important is that the applying the same interventions to the whole population is inefficient. Applying risk stratification information to your patient population allows you to target both resources and the right interventions to the right patients in order to focus. It’s a much more efficient deployment of resources in order to be successful in this game so you’re not wasting time on patients who are otherwise low risk.
The non-engaged patient population, there’s always a sub-cohort of those patients that are always there. It just requires different skills to engage them from a care management standpoint. It’s very much an approach and a methodology that these organizations need to think about to solve that problem.
We will probably look back years from now and see the readmission focus as tactical, with an uncertain impact on outcomes and maybe even on overall cost. Will this push to identify high-risk patients extend further than just keeping them out of the ED and inpatient beds?
I think that’s right. CMS has been thoughtful about their approach for aligning incentives. They’ve gotten better over time for doing this. You see the commercial insurers following CMS’s lead.
The one metric of focusing on readmissions post-discharge, you do have to apply advanced proactive and thoughtful discharge planning to mitigate a patient from coming back, which includes understanding the local and outpatient ambulatory resources that are available in order to mitigate the acute readmission from happening. Even though it was focused on an inpatient metric, the ability to affect that measure required them to think pretty broadly about systems that are potentially external to their four walls to put these programs into place.
I thought it was a good exercise to being able to mitigate that measure or outcome on patient population against a broader portfolio of measures that they’re going to put into place, which is going to inevitably head to capitation 2.0, payment to keep patients otherwise healthy and not using unnecessary resources to stay healthy.
Couldn’t hospitals dig through their EHR data themselves without additional technology? Also, is it enough to use that inpatient data snapshot alone vs. what might have happened to that patient in the 30-day readmission window?
The philosophy is to use any and all available data on the patient in order to understand what’s going to happen in the future. EHR has provided a great, rich resource for that data set. They are real time and they’re clinically based. But you can augment that with claims data, billing data, and things like natural language processing, which is extracting information from the notes and also connecting that to publicly available data from things like the CDC or census information to understand average income levels or average education levels per ZIP code. All the information that is becoming more and more available on patients is very helpful in predicting the future that’s going to happen.
You want as much information as you can possibly get on a patient to predict the future. That includes not just the inpatient data, but the full gamut of inpatient, outpatient. You’ve got public HIEs, which can provide a rich resource if they’re structured correctly in capturing data centrally to have a longitudinal health record across the geographic area. But what you’re seeing health systems do more and more now is deploying more private HIE infrastructure to tap into that ambulatory information that’s extending beyond their four walls and at least setting up agreements with ambulatory providers to capture as much information to provide a comprehensive view on the patient.
Where solutions like what we provide come into play is allowing the machine to do as much as work as possible to help augment clinical cognitive thinking on the patient population. Computers and computer machine learning and so forth can automate a lot of information that a physician and or care manager wouldn’t otherwise be able to do. It can help them augment their clinical education and background in order to take care of patients by providing more information that they otherwise wouldn’t have.
Another component is the ability to integrate into the workflow. Risk information is helpful in providing the content to understand which interventions to apply to mitigate the risk. Automating that into the clinical workflow so that it becomes part and parcel of what a clinician and or care manager is doing on a day-to-day basis is a necessary component in order to not have bifurcated systems and make the workflow as efficient as possible.
What this gets down to is identifying proactively patients at risk with the interventions that apply to that and automating suggested care plans and orders on the patient that a physician or care manager can quickly think through in order to provide the right intervention to the patient.
Where do you see the concept of predictively identifying patients at risk playing out over the next five years?
When we first started this, there weren’t too many players in the game. What we saw mostly in the market were legacy, claims-based risk vendors who were focused on the insurance market or health plan market. What we’re seeing now are more companies like us using clinical information to provide real-time risk stratification information.
Over time, these will become more of a commodity and part and parcel of doing work because it’s necessary for organizations to think this way proactively about their patients and patient population and keep them healthy at home. All the right incentives are aligning to make this a necessary core component of taking care of patients while they’re healthy, while they’re in a pre-disease state, forever escalating up the risk curve.