The Battle of the Pips

The Faulty Radar

The ships that could not find the enemy blamed their radar.¹

Around them, American warships were opening fire. It was 26 July 1943, near the Aleutian Islands. A Japanese force was expected. Radar echoes, the “pips”, appeared to confirm its arrival.¹

Aboard USS San Francisco and USS Santa Fe, the guns remained silent. Their screens showed shells striking the water, but none of the ships those shells were supposed to hit. Their equipment seemed to be failing them at the worst possible moment.¹

But which equipment was failing? The sets that found the enemy, or the sets that could not?

Hold that question.

Imagine a child whose feed has been stopped because the volume drawn back through a gastric tube exceeded the unit’s threshold. By morning, the child has not aspirated. The team can point to the measurement. They can explain the decision. They can document the absence of the complication.

What they cannot yet establish is whether those three things are connected.

Rules of Engagement

Before the shooting began, intelligence had warned of a Japanese attempt to reinforce Kiska. The radar contacts appeared where a hostile force might plausibly be heading. Rear Admiral Robert Giffen ordered his ships into action.¹

Put yourself on that bridge. You have been warned that an enemy is coming. Instruments now appear to confirm it. How much more evidence would you require before opening fire? How would you explain waiting if your ships were attacked?

Now consider the corresponding bedside decision. A large gastric residual might indicate delayed emptying. Continuing to feed might provoke vomiting and aspiration. Withholding nutrition seems a reasonable precaution. This was the protective logic behind routine residual assessment that GASTRIC-PICU tested.²

Imagine explaining an aspiration episode after you continued feeding despite a high residual. Then imagine explaining why you stopped the feed and nothing happened. Which account feels easier to defend?

That difference should trouble us. The decision that is easier to explain may have acquired an advantage before its benefit has been established.

In a 2019 survey, 23 of 24 responding UK PICUs routinely measured gastric residual volume. Eighteen had an agreed definition of feed tolerance; every one included the residual volume.³

There is a circle here. Define intolerance partly by a high residual, and the measurement will reliably discover the condition it helps define. Stopping the feed then demonstrates that the result changed management. Whether that change helped the child remains a separate question.

The instrument has supplied a finding. Somewhere between the syringe and the feeding pump, we have turned it into an instruction.

A Quiet Night

At sea, the firing continued intermittently for half an hour. There was no convincing evidence of an enemy response. On the ships unable to detect targets, radar could still register the splashes of American shells.¹

The effects of the action were visible. Its justification remained uncertain.

Return to our child. The feed was stopped and the feared complication did not occur. Perhaps the precaution worked. Perhaps the child would have remained well if feeding had continued. Both explanations fit the morning handover.

A succession of similar nights would not settle the matter. Each could reinforce the same belief without supplying the comparison needed to test it.

This is the difficult territory of prevention: a complication avoided and a complication that was never going to happen can look identical afterwards. If we count every quiet night as evidence that the precaution worked, what experience could ever persuade us to stop?

We need to observe what happens when comparable patients receive different care. GASTRIC-PICU created that comparison.

The trial randomised 4,700 invasively ventilated children younger than 16 years across 24 PICUs in the UK and Switzerland. Children were receiving or starting gastric feeding, with extubation not planned within 48 hours. One group underwent routine residual assessment at least every six hours; the other had no routine measurement to guide feeding.²

Staff continued to assess vomiting, abdominal distension, pain or tenderness, and the child’s overall condition. Tube-position checks continued. Residual assessment remained available for clinical concerns or procedures. There were eyes on the patient even when there was no scheduled number on the chart.^2,4

Would that have felt like enough information to you?

Daylight

After the naval engagement, ships and aircraft searched for what had been hit. They found nothing. The battleships had expended 518 fourteen-inch shells against targets that nobody could subsequently locate. Unusual atmospheric conditions producing misleading radar echoes were suggested as an explanation.⁵

Daylight supplied a reason to reconsider the night’s interpretation.

The clinical comparison was more demanding. Gastric contents are real. Aspiration is a real danger. The question was whether routinely measuring the former, and adjusting feeding accordingly, improved the outcomes of children exposed to the latter.²

For the clinical co-primary outcome, GASTRIC-PICU ranked survival and days free from ventilation at 30 days, with death assigned the worst outcome. Among 4,460 children analysed, omitting routine checks met the prespecified noninferiority criterion: the adjusted odds ratio was 0.95 (95% CI 0.86–1.05), with the lower confidence limit above the boundary of 0.833. The per-protocol analysis also met that criterion.²

This established noninferiority within the chosen margin. It did not prove that every safety outcome was identical. The open-label design, greater consent withdrawal and more protocol deviations without routine checks also deserve attention, although the consistent per-protocol finding strengthens the conclusion.²

The amount of bedside activity changed substantially: 48,516 residual assessments with usual care, compared with 5,903 without routine checks. High residual volume was recorded as a reason for stopping feed in 406 of 2,131 children with available data in usual care, compared with 80 of 2,009 without routine checks.⁴

Here was a practice that generated thousands of measurements and repeatedly interrupted treatment. The trial had made it possible to ask what patients gained from that activity.

Omitting routine checks modestly improved nutritional delivery. Among 2,936 children with nutritional outcome data, the mean percentage of estimated energy requirements achieved by 72 hours was 80.3%, compared with 76.8% in usual care: an adjusted difference of 3.2 percentage points (95% CI 1.3–5.2). Nutritional data collection had stopped after a prespecified blinded reassessment established that sufficient data had accrued.²

Before celebrating the extra nutrition, apply the same scrutiny. More feed delivered does not, by itself, demonstrate better recovery. The trial did not demonstrate a survival or ICU length-of-stay advantage.²

Nor did it abolish uncertainty about harm. Vomiting leading to feed stoppage occurred in 10.2% without routine checks and 9.0% with usual care; ventilator-associated pneumonia occurred in 6.0% and 5.4%, respectively. Neither difference was statistically significant, but the confidence intervals allowed some increase in risk.²

We should resist turning a modest nutritional gain into a clinical victory simply because it fits the story we now want to tell. That would repeat the error we have spent the night examining.

Changing the Orders

The naval engagement ended when Giffen ordered his ships to cease fire.¹

In a PICU, an instruction can remain embedded in a protocol long after the meeting at which everyone agreed to reconsider it. Imagine the next nurse facing a feeding chart with a space for a residual volume, a colleague expecting that number at handover, and a recommendation to omit the check. Who has made it possible to act on the evidence?

ESPNIC recommendations had already advised against routine residual measurement in critically ill children in 2020. The GASTRIC-PICU editorial highlights the imbalance: limited evidence may establish a practice, while removing it demands a large randomised trial.^6,7

Suppose your unit had never introduced routine residual checks. Would these results persuade you to start? Would thousands of additional measurements and more interruptions for high residuals look like an improvement worth adopting?

If you would decline the proposal, what changes because the practice is already familiar?

For children resembling those enrolled, the findings provide a strong basis for adopting the tested approach: omit routine residual measurement, retain clinical assessment, and investigate concerns when they arise. The trial excluded children below 37 weeks’ corrected gestation, those with current or recent gut pathology or intestinal surgery, those receiving postpyloric feeding, and those on long-term invasive ventilation. Its findings should be applied within those boundaries.^2,4

That change needs agreement between nurses, dietitians and doctors, reflected in the feeding protocol, documentation and handover. Leaving an individual clinician to defend each omitted measurement preserves the old order while claiming to have accepted the new evidence.

For clinicians outside paediatric practice, the question travels further than the recommendation. Choose a routine measurement that alters treatment in your unit. Follow the result through to the action it triggers. Then ask what establishes the benefit of that action beyond the reassurance of having done something.

Perhaps the practice has excellent evidence behind it. Perhaps its place rests largely on a succession of uneventful nights. We should know which before the next result arrives.

The crews who could not see the enemy had questioned their instruments.¹ At the bedside, we must also be willing to question the instructions attached to ours.

If we would not introduce this practice today, what are we waiting for before we stop?