Why a hospital EHR is not an ICU clinical information system

Every few months the same conversation happens in a hospital boardroom. The ICU asks for a clinical information system; someone points out that the hospital already paid for an EHR; the request is parked. Eighteen months later the ICU is still charting on paper next to a workstation, and the EHR’s ICU module — bought, implemented, switched on — is used to write discharge summaries. Nobody did anything wrong. The two systems are built around different units of time, and no amount of configuration converts one into the other.

Order-centric versus minute-centric

An EHR’s data model is the encounter: a patient arrives, orders are placed, results return, documents are written, a bill is generated. It is a transactional record, and at that job mature EHRs are genuinely good. But its native rhythm is the order cycle — hours to days — and its chart review is built to answer “what was ordered and what resulted.”

Intensive care runs on a different clock. A ventilated patient generates more than 1,400 chartable data points a day; a 20-bed unit, upwards of 28,000 rows. The clinical questions are not “what was ordered” but “what is the trajectory over the last four hours” and “what happened in the three minutes around that alarm.” A clinical information system is built minute-first: the flowsheet is the centre of the world, time runs across the screen, and every other artefact — drugs, events, scores, notes — hangs off the timeline.

Device density is the dividing line

The practical test is what happens at the bedside. An ICU bed carries eight to twelve data-producing devices. A typical EHR integration samples a handful of vitals at fifteen-minute or hourly intervals through an interface, because its storage and screens were never designed for more. A CIS ingests the full stream — vitals at protocol cadence, ventilator and pump parameters, waveforms when needed — and keeps it reviewable. The difference is not cosmetic. Hourly sampling is what makes deterioration invisible until it is obvious; the drifting trend that early warning depends on lives between the samples.

Scores that recompute, not calculators that wait

Severity scores illustrate the gap precisely. In an EHR, SOFA or NEWS2 is typically a calculator: a clinician opens a form, keys in values that are already in the system, and files a score that is stale by the time it is saved. In a CIS, the scores are computed from the stream — SOFA, NEWS2, APACHE II and their peers recomputing continuously as values arrive, trending on the chart like any other parameter, and feeding the unit-level view that ranks beds by who is getting worse. zMed computes twelve such scores this way. A score that must be asked for measures workload; a score that recomputes by itself measures patients.

Alongside, not instead of

None of this argues for replacing the hospital’s EHR or HIS. The registration, billing, pharmacy, laboratory and ward documentation backbone is exactly where an order-centric system belongs. The architecture that works is adjacency: the CIS owns the high-frequency bedside record in the ICU, OR and ED; admissions, transfers, orders and results flow between the systems over standard interfaces — HL7 and FHIR exist for precisely this — and the discharge summary lands back in the hospital record. Each system does the job its data model was built for.

The procurement test

When a vendor says their EHR “covers the ICU,” three questions settle it. At what cadence does device data land in the chart, for how many devices per bed? Do severity scores recompute themselves as values arrive, or wait for a clinician to open a calculator? And can a consultant see the whole unit ranked by deterioration, live? If the honest answers are “hourly, a few, manually, no,” then the ICU does not have an information system. It has a place to file the summary after the fact.

zMed was built minute-first: flowsheets fed directly by the devices, twelve clinical scores recomputing continuously, a live unit-wide view, and standard interfaces to sit cleanly beside the HIS you already run. See zMed for the ICU.

References

1. Manor-Shulman O, Beyene J, Frndova H, Parshuram CS. “Quantifying the volume of documented clinical information in critical illness.” Journal of Critical Care. 2008;23(2):245-250.
2. De Georgia MA, Kaffashi F, Jacono FJ, Loparo KA. “Information technology in critical care: review of monitoring and data acquisition systems for patient care and research.” The Scientific World Journal. 2015;2015:727694.
3. Halpern NA. “Innovative designs for the smart ICU. Part 2: The ICU.” Chest. 2014;145(3):646-658.