When people think of electronic health records (EHRs), they assume these digital systems for storing patient information were designed for clinicians. But that is a fallacy. EHRs initially were designed for billing and compliance—and it shows.

Given the original purpose of EHRs, it’s not surprising that clinical workflows are disjointed. The limitations of traditional EHRs have made it more difficult for clinicians to quickly access relevant patient information at the point of care. This not only challenges clinician decision-making but also threatens patient safety and outcomes when key data is buried or missing altogether.

As a physician, here are the recurring challenges I consistently hear from peers, and a snapshot of the progress made that can help forge a better path forward.

Mounting cognitive fatigue

According to a 2023 study published in the National Library of Medicine, EHRs score lower on usability than other technologies when factoring in efficiency, effectiveness and user satisfaction. As the study’s author notes, “the volume and organization of data along with alerts and complex interfaces require a substantial cognitive load and result in cognitive fatigue.” Increasing documentation requirements have burdened physicians with clerical tasks that force them to spend excessive amounts of time on their computers.

Alert fatigue is another major contributor to burnout. When clinicians are getting 15 or more alerts every hour of every workday, and some or most are not essential to their decision-making, they eventually will start ignoring some. This makes it more likely these clinicians will miss information that does indicate a potentially serious health situation. And, as you can imagine, excessive alerts slow down workflow efficiency.

Most providers will also tell you that EHRs have not lived up to their promise of enabling data-sharing and coordination. Historically insufficient interoperability standards have made it difficult to share data between systems. That’s why hospitals still use fax machines.

To improve how EHRs perform, provider organizations must overcome technology- and process-related challenges to create an efficient clinical workflow that fully enables the sharing of meaningful data to complete the patient picture. That’s where AI comes in.

Automating intelligently

Today, advanced AI technologies are being deployed to simplify clinical documentation and improve interoperability and population health management. For example, we’ve been talking about ambient voice for years. It’s finally here, and it is being used to help clinicians break free from the EHR. Instead of having their heads buried in the computer, they can actually focus on the patient because ambient listening is capturing all the details of the conversation and creating structured documentation to easily review. This allows physicians to be fully in the moment with their patients.

AI-based analytics are also having a major impact across hospitals and health plans. Predictive analytics are now being used for bed management and block time management, functions that are essential to operational efficiency and staffing optimization. Another use case for AI-powered predictive analytics is identifying at-risk patients by interpreting subtle trends that a physician could miss because there may be multiple different variables that are only slightly abnormal. When viewed together, however, these variables may indicate an increased risk for a particular illness or condition. This is a powerful example of how AI can discern these connections in the data and truly assist clinicians.

AI also enhances interoperability by presenting community data to providers in a meaningful way. A clinician seeking community data typically must wade through unstructured data, duplicate data and huge amounts of other data irrelevant to their needs for a particular patient. AI can instead scan, normalize and aggregate the data in a digestible, meaningful manner so clinicians can make decisions without unnecessary hurdles.

Similarly, AI can quickly compile accurate, well-organized chart summaries and discharge summaries, both of which consume vast amounts of clinician time. Within these summaries, AI can provide clinicians with recommendations based on documentation, order sets and medication history, while providing links to the original sources for verification.

As a physician myself, I am optimistic that we are finally turning the corner on provider burnout by adding automation to a wide variety of time-consuming tasks including documentation, prior authorizations, referrals, patient reminders and billing. As AI helps modernize EHR workflows, clinicians can practice in a way that benefits both themselves and those under their care.

Truly engaging patients

On the patient side, AI tools can be integrated into patient engagement platforms that enable a far more user-friendly and intuitive experience than traditional patient portals. AI algorithms inside patient engagement platforms can bring a proactive approach to appointment reminders, medication nudges, education and patient communication across multiple channels.

By analyzing social determinants of health (SDOH) data, AI can also help tailor education for individuals based on factors such as access to transportation, access to healthy food and community resources to provide assistance. These types of tools can help empower patients to be active participants in their own care, which research ties to better outcomes. This kind of engagement will be critical as our health care system moves toward value-based care.

As we expand AI—particularly in patient-facing tools—it is crucial to prioritize safe and responsible usage. Greater efficiency and personalization are much needed in health care, but these benefits must be secondary to patient safety. AI should enhance, not replace, the human touch and clinical expertise that are foundational to quality care. With proper oversight, AI can support providers while driving better health care experiences for everyone involved.

Collaborating with the right partner

When assessing AI applications, provider organization decision-makers should look for a technology partner that prioritizes clinical usability and workflow simplicity as they develop and integrate new AI capabilities into their system. Providers should ask:

• Whether the vendor engages with customers throughout the development process to ensure new features and releases meet client needs today as well as their strategic imperatives for the future
• What guardrails the vendor has in place to maintain ethical use of bias-free AI and support incremental, sustainable client adoption
• If the vendor ensures full interoperability and easy access to well-organized clinical data
• How the vendor is incorporating tools in clinical documentation, patient engagement and analytics
• What the vendor does to protect patient privacy and prevent breaches
• What the vendor’s technology roadmap looks like—and how future-forward they are

Finally, provider organizations should look for a vendor that is financially stable and likely to be around in the long-term to support their own long-term success.

Leveling up the EHR

The limitations and inefficiencies of traditional EHRs undermine care quality, cost money and contribute to clinician burnout. New AI-based technologies such as ambient listening, intelligent automation and enhanced interoperability tools are improving clinical usability while increasing patient engagement. The journey to get to this point has had bumps in the road, but the sun is now rising on a bright new dawn for clinicians.

Laura Kohlhagen is a physician executive.