Provider Perspectives: Interoperability & Patient Data

Data interoperability is a hot topic in healthcare, especially in light of recent CMS mandates aimed at improving data exchange. But investing in interoperability is not just a means to achieving regulatory compliance; it also supports more effective care delivery and better clinical and financial outcomes.

As the Director of Risk Adjustment at the University of Pittsburgh Medical Center (UPMC) and Professor of Medicine for the school’s Division of Geriatric Medicine, Dr. Adele Towers has a unique understanding of the clinical and administrative benefits of interoperability. We recently sat down with Dr. Towers to discuss how data sharing in healthcare has evolved, what the industry can do better, and how the right solutions can support better outcomes for everyone.

—

As a clinician, what have you seen change in the last 10 years or so?

Dr. Towers: We used to do everything on paper. We called them the blue sheets, and providers would get information from the health plan on a piece of paper. If you were lucky, as a provider, you could look at that paper before the patient came and remember to put that documentation in your note. But most of the time the paper came after the patient came and you’d already done your note.

Now, fortunately, we’ve had a transition to electronic-based communications and documentation at the point of care.

How important is data reciprocity between trading partners?

It’s very frustrating as a provider to get information from a payer that you’re not aware of at all. The payers have a group of data, the providers have a group of data, and they don’t always speak. But now UPMC’s payers have access to a treasure trove of data from our clinicians through a large database that we call Alexandria Charts. This enables the clinicians to get data about their patients that ultimately—you hope—will improve patient care, because that’s our objective as providers.

[I can] speak to the frustration providers feel when they get notifications from payers about patient conditions that the providers were completely unaware of. It’s just a challenging environment to be in: we have 15 or 20 minutes to be with a patient and we have to assess all these different aspects.

The more that we can create that interoperability where the data is shared between the payer and the provider and it’s shown to the provider at the point of care, that will make our jobs—and the validation of conditions and risk adjustment and in quality—much easier.

What steps has UPMC taken to gain a more comprehensive view of patients?

We started back in 2018 with developing a suspecting engine that ingested payer claims data that was curated by the health plan. We also created a suspect list based on NLP (natural language processing) digestion of the EMR records that includes things like the problem list, the medications, the labs, all the clinical content of PFTs, chest X-ray, CT scans, all that documentation, [creating] those suspects, and then implementing that initially with a pilot group of physicians who didn’t mind clicking the little clickers.

At the same time, we also implemented [Edifecs’] Post-Visit Review tool, which is where—before the claim is submitted—the NLP engine reviews that note to ensure that the coding is accurate. And just as a background, 80% of our outpatient notes at UPMC are coded by providers. We don’t have someone who will digest the chart before you go in and tell you, “Do this, do that”; we have to leave it up to the providers.

UPMC has thousands of providers, so there’s no way to scale a CDI program for the number of patients that we see. So we implemented [Edifecs’] Point of Care Suspects tool, where the provider receives an alert within Epic to review the diagnoses. If they agree with them, it goes onto the claim and into the problem list, and if they disagree, then that information can be sent back to the health plan to be reviewed for removal of that diagnosis from the chase list.

The Post-Visit tool [showed us] that we often remove more diagnoses than we add. So you really need the two tools together, from my perspective, because I view the Post-Visit tool as a compliance tool that removes [outdated information]. With the two tools together we’ve been able to generate an incredible amount of revenue. Point of Care Suspects brings the conditions to the providers and gives them documentation, and then Post-Visit removes the ones that are inaccurate, and it also adds diagnoses.

We do three lines of business: we do Medicare Advantage, we do ACA and we do Medicaid, and you can’t do retrospective coding [with Medicaid]. So we put a big focus on our Medicaid population to ensure that those diagnoses on the claim are accurate. We’ve also used the Retrospective Review tool since 2013 with great success.

Any final thoughts you’d like to share?

In the time that we’ve been using Edifecs’ solutions, it’s been extraordinarily successful. I like to think of it as a team of the Retrospective, Point of Care, and Post-Visit. They have just been life-changing for us at UPMC in terms of ability to capture diagnoses and get that full picture of illness for our patients.

—

Our conversation with Dr. Towers underscored the importance of synthesizing payer and provider data, particularly when it comes to highlighting suspected conditions. Interoperability is the bridge that connects these two worlds and enables a more comprehensive picture of patients before, during, and after the encounter.