The best clinical decision support is the one you don't have to look for

Physicians have access to more medical information than ever before. Clinical guidelines, research papers, treatment algorithms and AI-powered search tools can answer almost any clinical question.

But there's a problem: having the right information available doesn't necessarily mean a physician can access it when it matters.

And in a busy oncology clinic, that distinction can make all the difference.

Doctors don't need another place to look

We recently spoke with a senior community oncologist who spends much of his time developing technology to support clinical decision-making. He liked Open Medicine's Living Algorithms, particularly the ability to visualize different treatment pathways and see options he might not otherwise have considered.

But he raised an important concern. "I don't have time to read an algorithm on my phone at the point of care, when I'm talking to a patient. I want information to be surfaced exactly and instantly as it's delivered."

His point was simple. Even the best clinical resource has limited value if a physician doesn't have time to consult it.

Community oncologists may see dozens of patients a day, often with entirely different cancers. They're reviewing scans, interpreting molecular tests, managing toxicities, discussing treatment options and helping patients navigate some of the most difficult moments of their lives.

Asking them to open another website, search for an algorithm and interpret it during a consultation adds yet another task to an already demanding workflow.

The problem isn't always access to information. It's getting that information to the physician without forcing them to go searching for it.

What if the right information appeared automatically?

Imagine an oncologist seeing a patient with metastatic lung cancer. The patient's molecular testing reveals an EGFR mutation. Instead of requiring the physician to search for the relevant treatment algorithm, the clinical system recognizes the mutation and automatically surfaces the appropriate treatment considerations.

Or imagine a patient starting a new regimen. The system automatically highlights recommended supportive medications, monitoring requirements or additional tests that might otherwise be overlooked.

The physician doesn't have to remember to ask the question. The relevant information just appears as part of the existing workflow.

This is particularly interesting as AI becomes capable of interpreting clinical conversations, extracting structured information from medical records and connecting that information to clinical decision-support tools.

A treatment algorithm no longer has to be something a physician opens and reads. It can become a source of structured knowledge that other clinical systems use to deliver guidance at precisely the right moment.

The physicians who need it most may be the least likely to use it

Our conversation revealed another important insight. The oncologist explained that he reads extensively and actively seeks out new clinical information. He would use a resource like Open Medicine because he's already accustomed to researching treatment decisions.

But he was worried about physicians who are less comfortable with technology, less familiar with recent advances or simply too busy to seek out additional information. These physicians may benefit enormously from clinical decision support, yet they're also the least likely to open another website.

This creates an uncomfortable paradox. The physicians who most need additional clinical guidance may be the least likely to go looking for it.

Making medical knowledge freely available is an important first step. Making sure that knowledge reaches physicians when they need it is an entirely different challenge.

From clinical algorithms to clinical intelligence

Open Medicine organizes complex treatment decisions into practical, expert-authored pathways. They capture not only what treatments are recommended, but also the reasoning, exceptions and clinical considerations that experienced physicians use when making those decisions. Today, physicians can explore these pathways directly.

But structured clinical knowledge has another advantage: it can potentially be integrated into the systems physicians already use.

A Living Algorithm could help inform an AI assistant, an electronic health record or another clinical application. Instead of asking physicians to navigate an entire decision tree, a system could identify the relevant branch based on the patient's clinical characteristics and surface the appropriate considerations.

The physician remains responsible for interpreting the information and making the treatment decision. The technology helps ensure that relevant knowledge is available at the right time.

Of course, doing this well requires accurate patient information, clinically validated rules, appropriate safeguards and integration into real-world workflows. An irrelevant alert can be just as frustrating as another unnecessary click.

The goal isn't to interrupt physicians more often. It's to make the information they receive more useful.

The future of clinical decision support is fewer clicks

Healthcare technology has a bad habit of creating new destinations. Another portal, another dashboard, another application, another login. Each promises to make physicians more efficient. Yet collectively, they often create more work.

The next generation of clinical decision support should move in the opposite direction.

Instead of asking physicians to adapt their workflow to the technology, the technology should adapt to the workflow to the physician.

That means understanding the clinical context, recognizing when a decision is being made and providing relevant, expert-authored guidance without unnecessary interruptions.

It also means recognizing that physicians need to spend their time with patients, not navigating software.

As the oncologist told us: "I don't want to add more clicks at the point of care. My patient doesn't need someone to tell them what the standard of care is. They need someone to hold their hand and tell them that they're dying."

That observation captures something important about the future of medical AI. The objective shouldn't be to make physicians spend more time interacting with technology.

It should be to help physicians spend more time with patients.