This fall, imaging leaders have repeatedly asked us a critical question: Are we better off waiting for our EHR to address diagnostic follow-up?
It’s a fair question. EHR vendors are pouring money into AI, and nobody wants to buy something their EHR will give them next year. For hospitals on Oracle Health (Cerner), the question is tangled up with bigger ones about where the platform is headed. Oracle Health lost a net 56 hospitals in 2025, its third straight year of losses, according to KLAS data reported by Becker’s. For community hospitals on MEDITECH, CPSI, or similar platforms, the question is usually simpler: how long would we be waiting?
Meanwhile, the patients with findings in this week’s reports can’t wait. In a July 2026 focus group, we asked ten radiology and imaging leaders how they manage follow-up. Only one could reliably say what share of recommended follow-ups actually happen.
What EHR-based tools do well
Give credit where it’s due. Tools built into the EHR, or bolted onto it, have gotten good at the first step: finding the recommendation in the report. Natural language processing can read a free-text radiology report, spot “recommend follow-up CT in 6 months,” and drop it on a worklist. That used to take a person hours. At East Alabama Medical Center, automating that step alone cut report review from five hours a week to 15 minutes.
If identification were the whole problem, waiting for your EHR would be reasonable.
Where patients still fall through
It isn’t the whole problem. The leaders in our focus group described the same pattern again and again: the technology handles the easy step and leaves the hard part to people.
“They might go through fifty patients to get two that didn’t have something scheduled already.”
-Focus Group Participant
The hard part is the middle of the workflow, where patients are actually lost:
- Sifting through the noise to find the signal. Identifying that a report has a recommendation isn’t the same as identifying when a navigator needs to step in.
- Reaching the right provider. The ordering physician may be outside your system, on a different EHR, or no longer the patient’s doctor.
- Getting the order placed. A notification in an inbox is not an order.
- Getting the exam scheduled and completed. Patients now read their results in the portal before anyone calls. Older patients often don’t use the portal at all.
- Confirming closure. Without it, no one can say what share of follow-ups happened. Without that number, there’s no business case.
Two more gaps make the wait harder to justify. First, most hospitals have mature programs for breast and lung, and almost nothing for those without a program, including hepatobiliary, renal, adrenal, or vascular findings. Second, the work that does happen depends on navigators that leaders told us they can’t hire or can’t get approved. One participant summed up the volume problem: “We can’t turn it on for everything.”
An EHR roadmap may eventually improve identification. It’s much less likely to identify follow-ups that need escalation, staff your navigator bench, reach providers outside your walls, or measure closure across every finding type.
What to look for in anything you add
Whether you wait, build, or buy, these seven questions separate tools that find findings from tools that finish them.
- Does it sort through recommendations to give you actionable information? Some recommendations will close on thier own, some need a little nudge, and some require a human touch.
- Does it measure closure, not just identification? Ask for your closure rate, meaning how many recommended follow-ups were actually completed, not how many were flagged.
- Does it cover every finding type? Breast and lung are table stakes. Look for one workflow that also handles all findings across your organization and recommendation types: incidental, explicit, and implicit.
- Does it reach providers where they work? Notifications should land in the ordering provider’s existing workflow, including providers outside your health system.
- Does it reach patients beyond the portal? Design outreach for everyone – not just the tech-savvy.
- Does it escalate only when automation fails? Your navigators should manage exceptions, not work lists start to finish. That’s how a small team covers every finding type without new FTEs.
- Will it survive an EHR change? If you might switch platforms in the next few years, your follow-up program and its history shouldn’t have to start over.
A good first step needs none of these decisions. Run a retrospective analysis of 12 to 24 months of radiology reports to find your real closure rate. That number turns “should we invest?” into “how far behind are we?”
The bottom line
Waiting for your EHR is a decision too. Every month you wait, that month’s findings go untracked. You don’t need a new EHR to close the loop. You need a way to measure closure, a clear owner, and technology that finishes the work instead of just finding it.
Want to go deeper?
- Read Closing the Loop, our report on what ten imaging leaders told us about follow-up, with a four-stage framework.
- See how East Alabama Medical Center increased completed lung nodule follow-ups by 74% without adding staff.
- Heading to RSNA (Nov 29 to Dec 2)? Stop by booth 5528. No pitch required.
