Interoperability costs now rival the clinical harm of incomplete charts
Cost has become the dominant obstacle to interoperability, and the shift is changing how health systems talk about data exchange. In a survey conducted by CliniComp with the CHIME Foundation, 47% of chief information officers named cost as their top barrier. Sandra Johnson, senior vice president of client services at CliniComp, framed the problem as a move away from a purely technical challenge toward a business and vendor-relations problem. The practical result reaches the bedside: physicians are still making decisions with fragmented records, missing context that would normally shape diagnosis, medication review, and follow-up.
Treating a patient with only part of the chart is not a software annoyance. It is a clinical handicap.
When data remain trapped in separate systems, the physician’s view of the patient becomes partial by design. That can affect everything from medication reconciliation to recognition of prior imaging, specialist notes, and downstream care plans. Johnson described a world in which the chart exists, but the pieces do not come together when clinicians need them. In acute care, where seconds matter and assumptions can harden into decisions, that gap can slow care or distort it.
Cost, vendor behavior, and the new interoperability bottleneck
Johnson’s central claim is that the industry has already solved part of the interoperability puzzle. Vendor consolidation, policy changes, and government incentives have connected more systems than existed in earlier eras of health IT. The next challenge is less about whether data can move and more about whether vendors and health systems will invest enough to make that data usable at the point of care. In other words, transport alone is no longer the finish line. Usability is.
Health systems have often approached interoperability by layering new tools on top of old infrastructure. Johnson argued that this strategy falls short because bolt-on solutions do not create a durable foundation. The deeper problem is architectural: if the core EHR environment is not built to normalize and surface data across sources, clinicians end up with multiple versions of the same patient instead of a single usable record. For physicians, that means more scrolling, more reconciliation, and more reliance on memory when the system should be doing the synthesis.
Vendor blocking still sits inside that conversation, even if the terminology has changed. Johnson characterized it as a market failure that requires participation from every part of the industry. The implication is blunt. If one vendor, one health system, or one procurement decision treats exchange as optional, the entire care chain absorbs the penalty.
Patient identity matching determines whether the data mean anything
The conversation around interoperability often turns technical too quickly, but Johnson returned to a more basic issue: matching the patient correctly across systems. Without reliable identity matching and data normalization, even a technically successful exchange can leave clinicians with records that do not line up cleanly. Duplicate entries, mismatched demographics, and inconsistent formatting can all fracture a patient’s story before the clinician ever sees it.
The result is not merely an administrative nuisance. It can lead to a record that appears complete in volume but incomplete in meaning. A physician may see encounters, lab values, or imaging from multiple settings and still not have a coherent picture. The value of interoperability lies in integration, not accumulation.
AI cannot correct for a fragmented clinical record
Artificial intelligence is now tied to the interoperability discussion because every CIO in the survey said AI running on interoperable systems mattered to the organization. That is a revealing point. Health systems understand that AI built on siloed data will only reproduce the incompleteness already present in the chart. If an algorithm can only see part of the patient, it will generate partial answers with more speed, not more truth.
AI depends on the quality, continuity, and context of the data it receives. A fragmented record can amplify blind spots rather than remove them. In practice, that means a model may miss longitudinal patterns, cross-site medication histories, or the context needed to interpret a clinical event. Johnson’s warning is simple enough to be unsettling: advanced tools do not rescue broken infrastructure. They inherit its failures.
Physicians are already feeling the workflow cost
Johnson urged physicians to tell their IT teams exactly where fragmented data slows down clinical decisions. That advice places the burden where it belongs: on the daily workflow. Clinicians know when a record is missing the outside discharge summary, when medication lists do not reconcile, or when a prior imaging study exists somewhere else in the system but cannot be found in time. Those are the moments when interoperability stops being abstract and becomes a direct determinant of efficiency and safety.
There is also a human cost in making clinicians reconstruct care manually. Each extra step steals attention from the patient in front of them. Each unresolved data gap increases the chance that a physician must choose between moving forward with incomplete information or delaying care while searching for it. Neither option is ideal, and both carry consequences.
A five-year view centered on usable data at the bedside
Johnson’s five-year vision is not a future crowded with more software. It is a future in which the clinician can open the chart and see a complete view of the patient at the point of care. In that setting, technology should support the decision rather than interrupt it. The idea is straightforward, but achieving it will require health systems to treat interoperability as core infrastructure, not an add-on purchased after the budget is already strained.
The immediate challenge is financial because cost now dominates the barrier conversation. The larger challenge is operational because data that cannot be normalized, matched, and surfaced at the bedside remain clinically incomplete even when they technically exist. Every CIO in the survey recognized the importance of interoperable AI, but the more urgent question is whether health systems will build the underlying foundation needed for that promise to work in real clinical settings.
Until then, many physicians will keep practicing with partial visibility. The chart may be digital, but the picture is still incomplete.