Going Beyond EHR: Leveraging AI and Digital Transformation Strategies

When people hear “digital transformation,” they think of a forward-moving strategy with a modern approach to delivering care. It could mean converting data into action or refining an electronic health record (EHR). With the adoption of EHRs and increasing interoperability, it has become possible to, for example, reconcile medication lists across multiple locations of care, even with hospitals and practices being across state lines, which is saving lives and reducing admissions. Drug-to-drug interactions are also being reduced by having access to the complete list of medications of the patients being treated. Additionally, artificial intelligence (AI) algorithms are increasingly used in diagnostic equipment such as retina scanners and radiology imaging to examine clinical findings and detect early signs of diseases. These AI platforms can aggregate data at much deeper levels, including listening to conversations or detecting stress and pain in our vocal cords. With the increasing adoption of EHRs, the next phase appears to be the transformation of IT platforms from data collection tools to interactive automation, which would include replacing many tasks historically managed by people. This article will examine the key areas where digital transformation may be leveraged to improve operational decision-making and quality of care.

Before we dig into this topic, we first need to provide a disclaimer:

Coker (nor the Author) has any financial conflicts of interest and/or any affiliation with any AI platforms or AI solutions.

‍

The Data

All digital transformation strategies and AI strategies assume we are starting with data. And not just any data. The data must be high quality and formatted or organized into structures necessary for AI algorithms to leverage. Therefore, data is fundamental before significant efforts are made toward digital transformation initiatives or adopting an AI strategy. The problem most health systems quickly run into is the age-old “record of truth” dilemma. Most systems consider their EHRs to be their record of truth, but EHRs obtain data feed from a myriad of internal and external IT environments. Most health systems will deploy a data lake or enterprise repository in response.

To determine the quality of data, the following questions must be considered:

• Is the data complete? What data is missing or usable?
• Are there duplication issues?
• Is there a record of truth associated with a master patient account or a master patient ID?
• Will the data hold up to integrity security? Are there any misspellings of words? Are outdated terms or improper medical terms being used?
• Do the current inputs drive the expected outputs? If not, why?
• How is the data conforming across the IT environment? (Structured vs non-structured vs. non-standard?)
• Are users making up their own special codes to signify something only understood by an isolated department? (For example, adding an asterisk in a field as a reminder for a specific task to be populated into a larger database.)

Examine the available resources and take a role-based approach to improving data quality. For instance, a data steward will need support from a data architect, and the data architect will need support from a business manager or Chief Medical Information Officer (CMIO), depending on the desired outcome.

The Components

For digital transformation strategies to be successful, the following three components are essential: the people, the process, and the technology. These pillars have stood the test of time, and the following is a breakdown of each:

1. People (this can be any user, patient, or payer)
   • Business leaders
   • Analysts
   • Architects
   • Developers
   • Data minors
   • Data stewards
2. Process
   • Analytics
   • Application-to-application
   • Business-to-business
   • Data migration
   • Data quality
   • Big data
   • Master patient indexing
   • Data archiving
3. Technology
   • Departments
   • Enterprise
   • Cloud
   • Hadoop
   • Embedded systems

The Market Pressures

The pressure to modernize and drive more open access to data sharing comes from payers, consumers, and the United States Department of Health and Human Services (HHS). In 2016, Congress passed the 21st Century Cures Act, which included many new rules about sharing patient data. [1] The act calls for vendors to publish their application programming interface (API) to allow data to flow across multiple platforms and apps. This concept is identical to how Google and Apple allow anyone to build apps compatible with their operating systems, hence why there are thousands of apps accessible in the App Store leveraging the data stored on our phones.

The Cures Act has opened up a world of possibilities and allowed innovators to develop digital solutions outside the practice's four walls. Even a savvy health system can now build a patient-facing app or enlist the support of an app developer. Many EHR vendors have created their version of an App Store or Marketplace to find these digital solutions. In most cases, the EHR vendor has published downloadable documentation needed to build these apps. This documentation is usually free, but some connection fees may apply.

As with any new digital solution, trial and error will be involved as these platforms mature. There were over 600 EHR vendors not long ago, so there is also a risk in being an early adopter. For those reasons, this is still a “buyer-beware” market.

Therefore, the following is essential:

1. If protected health information (PHI) is involved, have a business associate agreement (BAA) and ensure the vendor takes responsibility for any breaches while the PHI is in their control, including cyber protection.
2. Require an acceptance period in the contract with payments starting at activation. This allows you to determine if the digital solution works as promised before you fully accept financial responsibility for the platform.
3. If the app is being developed from scratch, understand who owns the IP right and have this in writing, including that the app may not be disabled.
4. Set parameters around costs. Include the right to have data returned at no cost and cap future rate increases at 2% or CPI.
5. Build in the right to terminate should any of the following occur:
   • The practice closes
   • A doctor retires or leaves
   • The practice is acquired
   • The product defects are unresolved
   • Chronic outages or other performance issues persist

Originally published May 9, 2024. Reprinted with Permission from Healthcare Administration Leadership & Management Journal, Volume 2, Issue 3, pages 93-94, Copyright © 2024, American Association for Physician Leadership, (800) 562-8088; www.physicianleaders.org.