The handover is the highest-risk procedure in the ICU

Count the procedures in an ICU by frequency and one dwarfs all the others. A 20-bed unit with two nursing shift changes a day performs roughly 15,000 bedside handovers a year, plus thousands more at medical shift changes and every transfer in and out. Each one moves the entire mental model of a critically ill patient from one head to another — and in most units it is performed verbally, from memory, under time pressure, with no instrument check and no record of what was actually said. Reviews of serious adverse events keep returning the same verdict: communication failure at transitions of care sits behind the majority of them. If a central line had that risk profile, it would have a checklist, a competency sign-off and an audit programme by Friday.

What actually gets dropped

Studies that record handovers and compare them against the chart find that a fifth to a third of clinically relevant items never cross the shift boundary. The losses are not random; they cluster in predictable categories. Pending results — the culture that will come back at 21:00 — vanish, because nothing is owed yet. Stop dates and review dates disappear: the antibiotic that should end tomorrow, the catheter on day six, the sedation hold due in the morning. Soft trends go missing — urine output drifting down over eight hours is invisible to a nurse who has seen only the last reading. And the context that prevents repeated harm, like the family conversation that already happened or the difficult airway noted at intubation, survives one handover and rarely two.

The pattern is worth naming: verbal handover preserves the present and loses the trajectory. The incoming nurse learns what the patient is, and not what the patient is becoming — which in intensive care is the half that matters.

The photocopy problem

Information that survives only in speech degrades like a photocopy of a photocopy. The day nurse tells the night nurse, who tells the next day nurse, and by the third retelling the lactate that was “trending up” has become “stable” because the inheritor never saw the curve. Writing it down on the handover sheet helps less than expected, because the sheet is itself a manual transcription — a summary of a summary, assembled in the last twenty minutes of a twelve-hour shift, by the most tired person in the room.

Draft from the chart, not from memory

The structural fix is to stop asking the outgoing nurse to reconstruct the shift. The chart already knows it. A record that captures device data, drugs, fluids, events and assessments continuously can draft the handover itself: significant events in sequence, the trends with their curves, fluid balance totalled, active infusions with their rates, pending tasks and due reviews carried forward as live items rather than as things to remember. The outgoing nurse reviews, corrects and adds the human layer — what the family was told, what to watch for, what the numbers do not show — and signs. The incoming nurse reads one screen instead of three notebooks and a verbal blur.

Two properties make this safe rather than merely convenient. The draft is attributable — every line traces to its source data, so the handover stops being hearsay and becomes part of the record. And the pending items are persistent: a task not closed on this shift presents itself again on the next, instead of depending on three consecutive memories.

What stays human

None of this replaces the conversation; it upgrades it. When the facts arrive pre-assembled, the five minutes at the bedside are spent on judgement — concerns, priorities, the patient you are worried about and why — instead of on dictating infusion rates that were already in the chart. Units that work this way report shorter handovers and fewer of the morning discoveries that begin with “nobody told me.”

The handover will always be the ICU’s most frequent procedure. It should be its most instrumented. zMed’s ICU charting drafts structured handover summaries directly from the live record — events, trends, balances and carried-forward tasks — so the shift change starts from the chart, not from recall. See zMed for the ICU.

References

1. The Joint Commission. “Sentinel Event Alert 58: Inadequate hand-off communication.” 2017.
2. Starmer AJ, et al. “Changes in medical errors after implementation of a handoff program.” New England Journal of Medicine. 2014;371(19):1803-1812.
3. Haig KM, Sutton S, Whittington J. “SBAR: a shared mental model for improving communication between clinicians.” Joint Commission Journal on Quality and Patient Safety. 2006;32(3):167-175.
4. Riesenberg LA, Leitzsch J, Cunningham JM. “Nursing handoffs: a systematic review of the literature.” American Journal of Nursing. 2010;110(4):24-34.