Failure to rescue is a communication problem wearing a clinical costume

Almost every failure-to-rescue case looks the same in the chart. A patient deteriorates over eight to fourteen hours. The vital signs document the deterioration accurately. Somebody notices. And then a gap opens between noticing and acting that nobody in the organization can quite account for afterward.

The instinct is to treat this as a clinical knowledge problem — the nurse did not recognize the significance of the trend, so we educate. Education gets deployed, competencies get signed, and eighteen months later the same event recurs on a different unit.

It recurs because the failure was almost never about recognition. In the cases we review, the nurse usually recognized the deterioration and said so. What failed was the pathway between saying so and someone with authority acting on it.

Where the gap actually opens

Three places, consistently.

The first is the page that goes unanswered. A nurse pages a resident at 2:40 a.m. There is no response. She pages again at 3:15. In the chart this appears as two entries and nothing else — no record of what she did in the intervening thirty-five minutes, because there was no defined next step. The escalation policy says to contact the provider. It does not say what to do when the provider does not answer.

The second is the chain of command that exists on paper and nowhere else. Ask a night-shift nurse on a medical floor who she calls if the resident will not come and the attending is not returning pages. In organizations where the answer is confident and specific, failure-to-rescue rates are lower. In organizations where the answer is a pause followed by 'I'd probably call the supervisor,' they are not.

The third is the rapid response call that gets negotiated. The nurse considers calling. The charge nurse, who has been told the unit's rapid response utilization is high, suggests waiting for the next set of vitals. This conversation is not documented anywhere and it is one of the most consequential events in the entire case.

What the record shows

When we review these cases in litigation, the thing that makes them indefensible is rarely the clinical judgment. It is the record's silence during the gap. Two pages, forty minutes apart, and nothing between them. A note at 4:00 a.m. describing a patient who was clearly deteriorating at 1:00 a.m. A rapid response call placed twenty minutes after the arrest was already inevitable.

That silence is read, correctly, as inaction. And the defense that the nurse was working the problem the whole time is unprovable, because the working was never written down.

What actually changes the outcome

Give the pathway a time limit and a named alternative. Not 'notify the provider' — 'notify the provider; if there is no response within ten minutes, page again and notify the charge nurse; if there is no response within twenty minutes, contact the rapid response team directly.' The specificity is the intervention. A nurse at 3 a.m. does not need to be persuaded to escalate. She needs to know exactly what the next number is and to have permission, in advance, to dial it.

Make rapid response calls unnegotiable. If a nurse thinks the patient needs a rapid response, the call is placed. Not discussed, not deferred to the next vitals. Any metric that treats high rapid response utilization as a problem is actively producing failure-to-rescue events, and should be retired.

Measure the interval, not the event. Most organizations track rapid response calls and arrests. Very few track the time between first documented abnormal vital sign and first escalation. That interval is the actual quality measure, and it is usually available in the data you already have.

Round on the pathway at night. A pathway tested during a daytime mock code is not tested. Walk it at 3 a.m. on a weekend with the people who would actually run it, and find out what they think the answer is.