Physician Health Partners: Data Quality and Validation Improves Care

Solving the challenges of working with a large, heterogeneous network

PHP has 43 individual clinical connections that span across 12 different EHRs and eight health plan payer feeds coming into its Arcadia platform, which unsurprisingly, leads to a few challenges. The biggest issues include technological savviness, adopting new technology, and getting providers on board with large initiatives.

Acknowledging it can be a tough process to get physicians on board with entering data, Smith believes you have to show, not tell, to create change.

“The main way that you drive those conversations — you have to have the data in front of the providers to show them, ‘Here are the actual numbers. Here’s how many patients that you have. Here are how many we’re actually capturing.’ Without that data, it’s a real nonstarter to have those conversations with providers,” Smith said.

The next step of the conversation is for Smith and his team to tie the effort to incentives because some providers don’t fully understand how doing this extra work benefits their practice in addition to their patients.

Standardization is also important. Without it, the work being done cannot be measured.

When we first look at an EHR system and there’s five physicians and a couple APPs entering data into an EHR system, we will pull the data out — mammograms, for instance — and there will be 10 different ways that those have been put into the EHR system. We’ll try to create, for instance, dropdowns that have a specific value that they select, instead of having seven different people typing in seven different things.

Glenn Smith
Director of Technical Program Management, Physician Health Partners

Technology has reshaped the data quality process

Less than five years ago, healthcare organizations like PHP sampled a few dozen random patients to get a pulse on quality measurements, sending people out to practices to collect these quality measures. They also relied on hospitals to send admission, discharge, and transfer (ADT) data.

Today, by embracing technology to augment these processes, PHP has one place for quality coaches, for instance, to go to see how their practices are doing.

“In the past, we had to try to run separate reports out of 10 different EHR systems and figure out where we get to the patient registry in one EHR system and get a good quality report out of another,” Smith said. “We’ve been able to eliminate all of that work and streamline our processes.”

As a result of these changes, patients are getting better care, quicker.

“We see when they get discharged in real time and providers are able to understand the whole picture of what’s happening with the patient, because we have all of the claims data and the EHR data, and potentially EHR data from other providers that have seen that patient,” Smith said. “Anytime a physician can know what has happened outside of their four walls, the better care that patient is going to get.”

Part of the change means PHP is no longer reactive to managing whatever data comes in. Instead, it intentionally provides its health information exchange with a list of members every month, and they send the ADTs for those members. Now Smith and his team can push those into Arcadia and filter the information to provide focused data to other PHP teams.

“One team is looking at emergency discharges in the past three days, so we set up a report to capture that,” Smith explained. “They come in every 25 minutes throughout the day. Prior to that, we were on a three-day lag between the time we got the paper report, entered it in, triaged it, entered into a system, and then put it out on a separate portal for a provider to pick up. We went from a three-day lag to same-day.” 

In the past, we had to try to run separate reports out of 10 different EHR systems and figure out where we get to the patient registry in one EHR system and get a good quality report out of another. We’ve been able to eliminate all of that work and streamline our processes.

Glenn Smith
Director of Technical Program Management, Physician Health Partners

Creating an effective data validation process

When PHP implemented its Arcadia platform, it learned that it needed to validate the data before it went to physicians, so Smith and his team developed four stages of data validation:

• Stage 1: The data integration team gets the initial connection to the data and does an initial analysis to compare what is in the EHR and what the data shows to ensure there are no major gaps.
• Stage 2: The clinical quality improvement team does an initial validation internally on the data. They look at the EHR, 25 Medicare patients, and 25 Medicare Advantage patients to compare across the measures.
• Stage 3: The practice will do its own validation on their own data.
• Stage 4: The data goes to the physicians.

“We’ve got it down to where, by the time it gets in front of a physician practice, most of the major glaring issues with the data have been resolved,” Smith said.

Data validation helps PHP and doctors act when it will impact patients most.

“We have teams of care managers and social workers that we deploy when a patient is discharged or admitted,” Smith said. “It allows those team members to offer services to patients in more real time, and potentially prevent readmissions that happen in the first couple of days after a patient gets discharged.”