Recognising Deterioration in Remote Nursing

Deterioration recognition is taught almost entirely inside a hospital model. Observations are recorded, a score is calculated, the score crosses a threshold, and a team with more resources than you arrives. The training focus is on detection and escalation, because response is somebody else's problem.

Remove the team and the whole structure has to be rebuilt. On a vessel, an installation or a remote clinic, you are the detection system, the response team and the intensive care unit, and the only variable you genuinely control is how much warning you give yourself.

If you want to practise catching it early, run a deteriorating-patient case in the app.

Trends beat thresholds, and here is why it matters more remotely

A single set of observations locates a patient. Two sets, spaced deliberately, tell you where they are going, and direction is the more actionable of the two.

Consider two patients with the same numbers: respiratory rate 22, heart rate 108, blood pressure 118/74, temperature 38.1, saturation 95% on air, alert. One has been sitting there for eight hours. The other was entirely normal ninety minutes ago. The score is identical; one of them needs your attention now and the other needs a plan.

In a hospital, that distinction affects how urgently a team is called. On a ship, it affects something much larger: whether you begin the evacuation conversation now or in six hours. Because the logistics have a long lead time, early detection converts directly into options — a helicopter window, a diversion, a port with capability — and late detection removes them permanently.

Which reframes the purpose of the score. In a hospital it triggers a response. Remotely, it buys lead time for a decision that cannot be made quickly.

Scoring properly, or not at all

Early warning scores discriminate reasonably well when used as designed and poorly when used casually. Four disciplines make the difference.

• Count the respiratory rate for a full period, by watching. It is the single most predictive parameter in most scoring systems and the one most often estimated or copied forward. A rate recorded as "20" for three consecutive sets is almost always a rate nobody counted.
• Score every parameter every time. A partial set produces a falsely low total, and a falsely low total is worse than no score because it provides reassurance you have not earned.
• Record the oxygen the patient is on. A saturation of 94% breathing air and 94% on supplemental oxygen are entirely different findings, and any score that ignores that is misleading.
• Write the score down with the time. The trend only exists if it is recorded. Remembering that "they seemed better earlier" is not a trend.

Two limitations to hold alongside the score. It performs less well where the patient's baseline is unusual, and a total driven entirely by one deranged parameter deserves individual thought rather than the escalation the total implies. The number is a prompt to think, not a conclusion.

The findings that precede the numbers

Observations lag. Experienced nurses routinely detect deterioration before the chart shows it, and the cues they use are describable rather than mystical. It is worth naming them, because a describable cue can be documented and handed over, while a hunch cannot.

• Change in behaviour or interaction. Someone who was chatty and is now answering in short sentences. Someone who has stopped complaining about the thing they were complaining about.
• Work of breathing rather than rate. Accessory muscle use, positioning, an inability to complete a sentence. All can precede a rate change.
• Skin and perfusion. Colour, temperature to touch, mottling, sweating that does not fit the environment.
• Fluid balance and output. Reduced urine output frequently precedes haemodynamic change by hours.
• Family or colleague concern. "He is not himself" from someone who knows the person is a genuine clinical finding, and dismissing it is a recognised failure pattern.
• Your own unease. If you find yourself checking on someone more often than the plan requires, that is data. Escalate the reason rather than the feeling: "I am checking every twenty minutes rather than hourly because his breathing has changed" is a clinical statement.

Escalating when there is nobody on the other end

The word "escalate" hides four different actions, and remotely they separate out.

Escalate the frequency. The first and most available action. Shorten the observation interval and say the new interval out loud with a reason. This is the cheapest intervention you have and it is frequently the one that saves the patient, because it is what generates the trend that justifies everything else.

Escalate the assessment. Move from an observation round to a structured primary survey. A rising score is a trigger to go and look properly, not a trigger to write a number down more carefully.

Escalate to remote advice. The telemedicine or shore-side clinical call. This has a lead time of its own, and it works far better when made early with an uncertain picture than late with a crisis. Bring a structured handover, a described trend rather than a single set, and an explicit question.

Escalate the logistics. The conversation with whoever controls movement — the bridge, the control room, the operations centre. Crucially, this can and often should start before a clinical decision to evacuate has been made, because you are asking what options exist and how long each takes, not requesting one.

That last point is the biggest practical difference from hospital practice. Asking "if I needed to move this patient in four hours, what would be possible?" costs nothing, commits to nothing, and converts an unknown into a planning constraint. Waiting until the decision is certain routinely finds that the weather window closed an hour ago.

A written plan beats a good instinct

For any patient you are worried about, write down four things. It takes two minutes and it survives shift change, fatigue and your own memory.

1. The interval. Full observations at a stated frequency, and who is doing them.
2. The triggers. Specific, named findings that mean call me or act now — a further rise in respiratory rate, a fall in conscious level, a systolic below a stated figure, a urine output below a stated volume. Pre-committing to the trigger removes the hesitation that costs the time.
3. The escalation path. Who you contact, in what order, with what information, and how long each step takes to produce a response.
4. The logistics question you have already asked. What movement options exist, and what the deadline is on each.

Handing that plan over is also the point at which vague concern becomes shared concern. A colleague who receives "watch him" has been given a task. A colleague who receives the four items above has been given a decision they can make without you.

Practising this

Deterioration cannot be rehearsed in a static format, because the whole skill is temporal. What you need is a case that continues after you have finished assessing it, so you have to choose an interval, commit to a trigger, and then live with the consequence of having chosen a long one.

Two specific rehearsals repay the time. Run a case where the patient deteriorates slowly enough that a single set of observations looks acceptable at every point — that is the pattern that gets missed. And run one where you have to make the logistics call early on an uncertain picture, because that is the decision most people report finding hardest.