Seeing the patient before the numbers
Assessment is the first step of everything else in nursing, and the step examinations test in practical form because it cannot be faked by recitation. What is being examined is whether a nurse actually looks at the patient, works in a safe order, notices what matters, and acts.
Health assessment is often taught as a set of measurements, and that framing produces nurses who record numbers accurately and miss patients entirely. The purpose of assessment is to build a picture of this person: what is normal for them, what has changed, how fast it is changing, and what that change suggests. A set of observations is part of that picture and not the whole of it. A patient whose numbers are all within range and who says they feel completely different from yesterday is telling you something the machine cannot.
The UK's OSCE devotes four of its ten stations to assessment, planning, implementation and evaluation around a single scenario, because assessment cannot be tested properly by asking what a normal respiratory rate is. What is being examined is whether the candidate actually looks at the patient, does things in a safe order, notices what matters, and acts on it. A candidate who recites perfect theory while never uncovering the patient fails, correctly.
Assessment has two halves and both are evidence. Subjective data is what the patient tells you — their symptoms, their history, what they are worried about. Objective data is what you observe and measure. Neither outranks the other. A patient's report of crushing central chest pain is subjective and is more important than a normal set of observations taken in the same minute, because the observations will follow later if nobody acts on what was said.
Assessment in an unwell patient follows ABCDE — airway, breathing, circulation, disability, exposure — because that is the order in which problems kill. A blocked airway kills in minutes, inadequate breathing in minutes to tens of minutes, circulatory failure over longer, and so on. This order is not a preference; it is a ranking by urgency, and the discipline of following it means a nurse never gets absorbed in an interesting finding while something faster is killing the patient.
How an assessment begins determines most of what it will yield. An open question — what has been happening, what brought you in, what is worrying you — gives the patient room to say the thing they came to say. A closed question produces a yes or a no and shuts the door behind it. Research consistently finds that clinicians interrupt patients within seconds of their opening sentence, and that the sentence they interrupted usually contained the diagnosis.
Every symptom can be explored through the same frame: where it is, what it feels like, how severe it is, when it started, what makes it better or worse, whether anything else came with it, and whether it has happened before. Working through this systematically produces far more than asking whether the pain is bad. The question most often skipped is what the patient thinks is causing it, which frequently produces the most useful sentence in the whole conversation.
What has happened before shapes what is happening now: previous operations, chronic conditions, previous admissions. Medicines matter for three reasons — what they treat tells you what conditions exist, what they do explains findings, and what has been missed explains deterioration. Allergies must be asked directly and specifically, including what the reaction actually was, because a great many recorded allergies are side effects and a great many real allergies are recorded as nothing at all.
Who is at home, what the housing is like, whether there are stairs, whether anyone helps, what the person does for work, what they drink and smoke — these are not background colour. They determine whether a discharge will hold, whether a wound will be dressed, whether medicines will be taken, and whether the patient will return next week. A clinically perfect plan that does not survive contact with the patient's actual life is not a plan.
Patients routinely leave the most important thing until last, mention it while you are leaving, or attach it to something trivial. They may minimise out of politeness, or omit out of shame. Noticing hesitation, asking one more time in a different way, and allowing silence long enough for the patient to fill it are learnable techniques. A nurse who finishes a history exactly on time every time is probably finishing early.
Respiratory rate is the most sensitive early indicator of deterioration and the most frequently fabricated observation in healthcare, and those two facts together do enormous harm. It requires standing with the patient and counting for a full period, ideally without them knowing, because people breathe differently when watched. A rate transcribed from the previous set because the patient looks fine has removed the earliest warning available, in exactly the patient in whom it would have appeared first.
Saturation is easy to measure and easy to misread. It is unreliable in cold or poorly perfused fingers, in shock, with nail varnish, with movement, and in carbon monoxide poisoning where it may read reassuringly high while the patient is dying. It must always be recorded with what the patient is breathing, because 94 per cent on air and 94 per cent on high-flow oxygen are completely different clinical situations and an unqualified number hides that.
Pulse is counted for rate but also felt for rhythm and strength, and an irregular pulse is a finding a machine may report as a number without flagging. Blood pressure requires the right cuff size — too small reads falsely high, too large falsely low — an arm supported at heart level, and a patient who has been sitting quietly. A blood pressure taken through clothing on an unsupported arm in a patient who has just walked to the chair is a number, not a measurement.
Temperature must be interpreted with the patient rather than alone: an older person with a serious infection may have a normal or low temperature, and a low temperature is itself a recognised sepsis sign. Consciousness is described with a scale rather than a word, because drowsy means different things to different people while a recorded score does not, and a change in that score between two nurses is only meaningful if both measured the same way.
A single set of observations is a photograph; a chart is a film. A heart rate of 104 in a patient who has run 60 all week is a significant finding, while the same 104 in a patient who lives at 100 is not. This is the argument for charting properly and for reading the whole chart rather than the last line, and it is the single most common failure in early warning scoring — the score is calculated correctly from numbers nobody has looked across.
A great deal is available before any instrument is used. Is the patient sitting up or slumped, alert or withdrawn, comfortable or guarding? Is their colour normal, pale, flushed, grey or mottled? Are they sweating? Can they complete a sentence without stopping for breath? Experienced nurses do this in the first seconds of entering a room, and students should be taught it as a deliberate step rather than assuming it will come with time.
Skin carries information about circulation, oxygenation, hydration, nutrition and pressure damage. Colour, temperature, moisture, turgor and integrity are all assessed by touching the patient, which is why an assessment conducted entirely from the end of the bed misses most of it. Pressure areas must be looked at rather than asked about, and heels in particular are the site most often missed because they sit under a blanket at the far end of the bed.
The order of abdominal assessment is deliberately different from everywhere else: look, listen, then feel. Listening comes before palpation because pressing on the abdomen changes the bowel sounds you are about to listen for. Look for distension, scars, visible peristalsis and asymmetry; listen for present, absent or altered sounds; feel gently for tenderness, guarding and masses, starting furthest from the pain so the patient does not tense before you reach it.
Calves are assessed for swelling, warmth, redness and tenderness because venous thromboembolism is common in hospital and treatable when found. Oedema is assessed for how far up the leg it extends and whether it pits. Peripheral pulses, temperature and colour are assessed for arterial supply. This is a two-minute examination that is skipped constantly and that finds conditions which kill patients who were otherwise recovering.
Older patients frequently present atypically: infection without fever, myocardial infarction without chest pain, and almost anything as confusion or a fall. Baseline matters more than population norms, because an older person's normal may sit outside the standard ranges and a value inside those ranges may represent a large change for them. The single most useful question is often to a relative: is this how they usually are?
Normal ranges for heart rate and respiratory rate vary substantially with age, so a rate that would be alarming in an adult may be normal in an infant and vice versa. Children compensate impressively and then decompensate suddenly, which means a child who looks tired and quiet after a period of distress may be deteriorating rather than settling. Parental concern is a clinical sign with real predictive value and dismissing it is a documented route to harm.
Patients who are unconscious, intubated, severely cognitively impaired, or who do not share your language cannot give a history, and their assessment depends more heavily on observation, on the record, and on people who know them. Working through an interpreter changes timing and phrasing but not the standard of assessment owed, and using a family member as interpreter for anything sensitive is a well-recognised source of both error and harm.
Assessment is affected by who the clinician finds credible, and the documented pattern is that pain and symptoms are under-assessed in older patients, in people with cognitive impairment, in people with a history of substance use, and across racial and language lines. This is not an accusation against individuals; it is a finding about systems that individuals sit inside. The practical defence is procedural: assess the same way every time, and notice when you are about to do less for someone.
Assessment is the first step of a cycle: assess, diagnose or identify the problem, plan, implement, evaluate. The step most often dropped is the last. A plan put in place and never evaluated is an assumption, and the question that closes the loop — did that work? — is what turns a task into care. The UK OSCE examines this cycle directly across four linked stations for exactly this reason.
Assessment findings are worth only as much as they transfer to the next person. That means describing what was observed rather than interpreting it into a word: 'unable to complete a sentence without pausing for breath' transfers; 'breathless' does not. It means recording what is normal for this patient, so the next nurse can recognise a change. And it means recording the absence of a finding you specifically looked for, because that tells the next person it was checked.
An assessment that identifies deterioration and does not reach anyone has failed. The structure taught almost everywhere is SBAR — situation, background, assessment, recommendation — and the recommendation is the part most often omitted. Saying plainly what you think should happen, even when unsure, gives the listener something to act on. 'I am worried about this patient and I think they need to be seen now' is a complete escalation; a list of numbers is not.
Deterioration is a trend, so a single assessment can only ever be a starting point. After any intervention, and at intervals determined by how unwell the patient is rather than by the drug round, the assessment is repeated. The patient who was assessed thoroughly on admission and not properly reassessed for eight hours is the patient in most incident reports, and the failure is almost never a lack of knowledge about how to assess.
Rate alone is not the whole of respiratory assessment. Effort matters at least as much, and effort is assessed by looking: is the patient using accessory muscles in the neck and shoulders, are the spaces between the ribs and above the clavicles drawing in, are the nostrils flaring, is the abdomen moving paradoxically against the chest? A patient with a normal rate who is working visibly hard to achieve it is closer to failure than a patient with a faster rate who is breathing easily, because effort cannot be sustained indefinitely. Exhaustion in a patient who has been working hard for hours presents as a falling rate, which is the most dangerous reassurance in the whole of assessment.
How a breathless patient positions themselves is diagnostic information they are giving you without being asked. A patient sitting bolt upright, leaning forward on their arms, is doing so because it is the only position in which they can breathe, and laying them flat for convenience removes it. Speech is the simplest functional test available: a patient who can speak in full sentences is moving enough air, one who speaks in short phrases is struggling, and one who can manage only single words is in serious trouble. None of this requires equipment and all of it is available in the first ten seconds.
Auscultation compares one side against the other at matched points, front and back, because asymmetry is far more informative than any single sound. Absent or reduced sounds on one side in a breathless patient is a finding that must be escalated immediately. Crackles at the bases suggest fluid, wheeze suggests narrowed airways, and a silent chest in a patient who was wheezing loudly an hour ago is not improvement — it means too little air is moving to make a sound. That reversal catches students out and is precisely why it is examined.
A cough is assessed for whether it is productive, whether the patient can clear what they produce, and what the sputum looks like. A patient who cannot clear their own secretions will accumulate them, and accumulation leads to infection and to blocked airways. After abdominal or chest surgery, pain prevents effective coughing, which is why supporting the wound during a cough is a real intervention rather than a courtesy — an uncontrolled cough against a fresh wound hurts enough that patients stop coughing altogether, and stop clearing their chest with it.
Circulation is assessed at the periphery because that is where the body sacrifices supply first. Hands and feet that are cool, pale or mottled when the core is warm indicate that the body is shunting blood centrally, which it does to protect the brain and heart when circulating volume or cardiac output is inadequate. Capillary refill is tested by pressing for five seconds and counting how long colour takes to return; prolonged refill in a warm environment is a genuine finding, while the same result in a cold room on a cold hand tells you about the room.
Urine output is one of the most useful circulation measurements available and is frequently treated as a fluid balance chore rather than a clinical observation. The kidneys receive a large share of cardiac output, and they are among the first organs to have supply reduced when circulation fails. Falling urine output in a patient whose fluid intake has not changed is a circulation finding before it is a renal one, and a patient who has not passed urine for several hours needs someone to establish why rather than a note saying so.
These assess the other direction: not too little circulating volume, but too much, or a heart that cannot move what it has. A raised jugular venous pressure, peripheral oedema that pits, breathlessness worse on lying flat, and rapid weight gain together suggest fluid overload. The assessment that distinguishes overload from hypovolaemia matters enormously because the treatments are opposite, and giving fluid to a patient in pulmonary oedema causes direct and immediate harm.
A monitor reports a number; a finger reports rhythm, volume and character. An irregular pulse may indicate atrial fibrillation, which changes risk and management substantially and which an automated reading may simply average into a plausible figure. A weak, thready pulse suggests low volume; a bounding pulse suggests the opposite. This is the clearest example in assessment of information available only to someone who touches the patient, and it is lost entirely in a ward where observations are collected by machine alone.
Consciousness is assessed on a scale rather than described in a word, because scales transfer between people and adjectives do not. The widely used approach grades eye opening, verbal response and motor response, each with defined levels, producing a score that can be compared across shifts and across staff. The critical teaching point is that a falling score is an emergency regardless of the absolute number, and a two-point fall in an hour matters more than a stable low score in a patient who has been that way for a week.
Pupils are assessed for size, equality and reaction to light, and the finding that must never be filed is a newly unequal or newly unreactive pupil, which can indicate rising pressure inside the skull and is a neurosurgical emergency. Comparing against the patient's own earlier record matters because some people have naturally unequal pupils and some medications and eye conditions affect the response. As everywhere else in assessment, change outranks absolute value.
Power is assessed by asking the patient to move against resistance and comparing sides. New weakness on one side, new difficulty speaking, or a new facial droop is a stroke until proven otherwise and is time-critical, because treatment options close within hours. The single most useful thing a nurse can establish is when the patient was last known to be well, since that time determines what treatment is possible, and it is a question only someone at the bedside can ask.
Any patient with reduced or altered consciousness has their blood glucose checked, without exception and early. Hypoglycaemia is rapidly reversible and rapidly fatal, it can mimic almost any neurological presentation including stroke and intoxication, and it is missed most often in patients whose confusion is attributed to something more interesting. It is the cheapest test on the ward and the one with the worst consequences for skipping.
Every other observation in this manual has an instrument. Pain has only the patient and the nurse, which makes it the observation most vulnerable to assumption. The principle is that pain is what the patient says it is, occurring where they say it does — and the documented reality is that some patients are believed less than others, with under-assessment concentrated in older patients, those with cognitive impairment, those with a history of substance use, and across language and racial lines.
A numerical scale is useless in a patient who cannot use it, and their pain is no less real. Behavioural assessment substitutes: facial expression, guarding a body part, restlessness or agitation, changes in breathing, vocalisation on movement, resistance to being turned, and change from that person's own baseline. The last is the most reliable and is available only to a nurse who knows the patient, which is a clinical argument for continuity rather than a sentimental one.
Some pain is an assessment finding demanding escalation rather than analgesia. Severe pain out of proportion to the injury, particularly in a limb, with pain on passive stretch, may be compartment syndrome and is a surgical emergency in which analgesia masks the only warning. Sudden severe abdominal pain in a post-operative patient, new chest pain, or a sudden severe headache described as the worst ever are all findings, not symptoms to be medicated while waiting to see.
Assessment does not end when analgesia is given. Reassessment at an interval appropriate to the route establishes whether the pain has actually improved and whether the treatment has caused harm. For opioids the observation that matters most is sedation level, since increasing sedation precedes respiratory depression, and a patient becoming harder to rouse is giving the warning while still looking peaceful.
Hydration is assessed from several findings together because no single one is reliable. Dry mucous membranes, reduced skin turgor, sunken eyes, thirst, concentrated urine, reduced urine output and a postural drop in blood pressure each point the same way, and any of them alone can mislead. Skin turgor in particular is unreliable in older patients, whose skin loses elasticity with age regardless of hydration, which is exactly the group in whom dehydration is most common and most consequential. The most dependable measurements are daily weight taken the same way each day, and an honestly completed fluid chart — and the word doing the work in that sentence is honestly.
Nutritional assessment begins with screening on admission, combining body mass index, recent unintentional weight loss and the effect of acute illness on intake. It continues with something simpler and more often neglected: noticing what actually gets eaten. Three trays returned untouched is a clinical finding, and it is visible only to somebody who looks at the tray rather than the chart. Malnutrition in hospital slows wound healing, increases infection, costs muscle and lengthens stay, and it is very rarely caused by an absence of food.
Assessing swallow matters because a patient who aspirates develops pneumonia, and aspiration pneumonia kills. The warning signs are coughing during or after swallowing, a wet or gurgling voice afterwards, food held in the cheek, drooling, very slow eating and recurrent chest infections. After a stroke, screening before any oral intake — including before oral medication — is standard practice almost everywhere, and giving a tablet with a sip of water to an unscreened patient is a common shortcut with a serious consequence.
Bladder and bowel assessment is routinely skipped because it is unglamorous and because patients are reluctant to raise it. Both are wrong reasons. Urinary retention causes pain, confusion and kidney damage and is frequently missed in patients who cannot report it. Constipation causes confusion, abdominal pain, nausea, overflow diarrhoea mistaken for infection, and in extreme cases obstruction. Asking directly, recording accurately, and noticing the absence of a bowel movement over days is basic assessment with a large effect on outcome.
The admission assessment is not merely the first in a series. It establishes what normal looks like for this person, and every later observation is interpreted against it. A patient whose usual blood pressure is 160 systolic is in trouble at 110, and nobody will recognise that unless somebody recorded the 160 on the first day. Recording a patient's baseline mobility, continence, cognition and the support they have at home does the same work for discharge that baseline observations do for deterioration.
The frequency of reassessment is set by how unwell the patient is and how fast they are changing, not by the convenience of the round. This sounds obvious and is routinely violated, because observation rounds are scheduled for the ward rather than the patient. A patient who has just deteriorated, just been given a treatment intended to change something, or just returned from theatre needs assessment sooner than the timetable suggests, and a nurse who waits for the next round because it is nearly due has chosen the schedule over the patient.
Assessment transfers only as well as it is written. Three habits make the difference: describe what was observed rather than interpreting it into an adjective; record what is normal for this patient so change is recognisable; and record the absence of a finding you specifically looked for, since that tells the next person it was checked rather than missed. A note reading 'chest clear, no calf tenderness, heels intact' is worth more than one reading 'observations stable' because somebody can act on it.
The hardest judgement in this subject is a patient who looks unwell while every measurement is normal. The instinct is to trust the numbers because they are objective, and that instinct is wrong often enough to matter. Compensation holds numbers normal until it fails; some patients cannot mount the expected response at all; and a nurse's impression that something has changed is built from information no chart holds. The correct action is to escalate the concern as a concern, saying plainly that the observations are normal and that you are worried anyway. That sentence has saved lives and is one of the most professionally difficult things a junior nurse ever says.
Candidates prepare for practical assessment stations by memorising sequences, and then lose marks on things that have nothing to do with recall. Examiners mark whether you introduced yourself and confirmed who the patient is with two identifiers. Whether you gained consent before touching anyone. Whether you washed your hands at the moments that required it, including after touching the patient's surroundings. Whether you exposed only what you needed and covered the patient again. Whether you spoke to the patient as a person throughout rather than narrating past them to the examiner. Whether you acted on what you found rather than simply reporting it. Almost every one of these is a courtesy that is also a safety behaviour, which is why they are marked rather than assumed.
Being watched changes performance, and the defence is structure. A candidate following a fixed sequence has something to fall back on when nerves remove the ability to improvise, which is precisely what nerves do. Saying the sequence aloud as you work serves three purposes at once: it keeps you in order, it tells the examiner what you are doing and why, and it keeps the patient informed about what is about to happen to them. A candidate who works in silence may be assessing perfectly and will not be credited for findings the examiner cannot see them make.
Stations are frequently lost in the final thirty seconds. Finishing means stating what you found, stating what you think it means, saying what you would do next, and saying who you would escalate to and how urgently. It also means leaving the patient comfortable, covered, with their call bell in reach and their belongings where they can get them — a candidate who identifies deterioration correctly and walks away leaving a half-undressed patient unable to reach a call bell has demonstrated something real about their practice.
None of the above is examination technique invented for examinations. Two identifiers prevents wrong-patient error. Hand hygiene prevents infection. Explaining before touching is the difference between care and processing. Stating a recommendation is what makes escalation work. Leaving a call bell in reach prevents the fall that happens when a patient tries to reach something alone. The examination is constructed from the behaviours that prevent the commonest harms in real wards, which is the most useful thing a student can understand about it.
The most common failure is an assessment conducted entirely without touching or uncovering the patient. Skin temperature, capillary refill, pulse character, calf tenderness, pressure areas, abdominal tenderness and oedema are all unavailable from a distance, and every one of them finds conditions that kill patients who were otherwise recovering. The habit forms honestly — wards are busy, patients are asleep, uncovering someone feels intrusive — and it removes most of the examination. Looking at a patient's heels takes ten seconds and is skipped more often than almost anything else in nursing.
Once a plausible explanation exists, subsequent findings get bent to fit it. A post-operative patient labelled anxious has their tachycardia attributed to anxiety; a patient with a history of alcohol use has their confusion attributed to withdrawal; an older patient's breathlessness is attributed to age. Each of these is sometimes correct and each is a documented route to missing bleeding, hypoglycaemia and pulmonary embolism. The defensive question is simple and worth asking aloud: if I did not already have an explanation, what would I think this was?
An automated blood pressure on a moving arm, a saturation probe on a cold finger, a respiratory rate carried forward from the last set — each produces a number that looks identical to a real measurement. A patient who says they feel dreadful while the machine reports normal values is providing information the machine cannot. The correct response is neither to dismiss the numbers nor the patient, but to repeat the measurement properly and escalate the discrepancy as a discrepancy.
The last error is documentary and it is serious: writing observations that were not taken, ranges that were not checked, or a systems review that did not happen. It occurs under pressure, usually with the intention of catching up later, and it corrupts the record for everyone downstream — the next nurse believes the heels were checked, the audit believes the score was calculated, and nobody knows the patient was never uncovered. An honest gap in a chart is recoverable. A fabricated entry is not, and in most jurisdictions it is treated far more seriously than the omission it was hiding.
The subtlest error is structural rather than individual. A shift organised entirely around tasks — observations at these times, medications at those, dressings after lunch — produces nurses who visit patients to do things to them and never simply look at them. The patient is seen six times and assessed none. This is how a deteriorating patient can be attended all day and still be found in trouble at handover, with every task signed for and the deterioration unrecorded because nobody was looking for it. The countermeasure is deliberate: when you are at a bedside for any reason, spend ten seconds assessing before you begin the task and ten seconds after. Across a shift that is a few minutes of work, and it is the whole difference between having seen a patient and having processed them.
On a ward where several people see the same patient, each can reasonably assume another has already checked the thing nobody checked. The patient came from theatre, so recovery will have looked at the wound. The physiotherapist was here this morning, so someone will have seen the legs. The night staff handed over that everything was fine, so it was. Diffusion of responsibility is a well-described failure in every safety-critical industry and hospitals are no exception. The defence is to assess what is in front of you rather than what the record says was assessed, and to say plainly at handover which things you personally checked and which you did not — a handover that distinguishes between the two is far more useful than one that reports everything as fine. The phrase that makes this concrete at a handover is simple and is worth practising until it feels natural: I checked the chest and the calves myself, and I have not seen the heels.