The Deteriorating Ward Patient — Written Clinical Reasoning

SAQ 1 — Integrated Management of the Deteriorating Patient with Septic Shock and Multi-Organ Failure (20 marks, 30 minutes)

Prompt: Outline your integrated management of Mrs O'Sullivan's acute deterioration, addressing: (a) the interpretation of her NEWS2 trajectory and the immediate escalation response; (b) the ABCDE structured assessment and the resuscitation in the first 30 minutes, including the fluid strategy in the context of her aortic stenosis; (c) the management of her hyperkalaemia and her new atrial fibrillation; (d) the drug review and the medications you would withhold and why; (e) the escalation and the ceiling-of-care decision, including the communication with the family; and (f) the common exam trap in the management of this patient. [1]

Model Answer

(a) NEWS2 trajectory and immediate escalation (3 marks): [1]

The trajectory is the central signal. A NEWS2 that has risen from 3 to 9 over six hours is a patient in accelerating physiological decompensation — the trend is more informative than any single value. At a NEWS2 of 9, she has crossed the emergency threshold (7 or more), which mandates an emergency critical care assessment within minutes, not hours [2]. At a NEWS2 of 5 to 6 she should have had urgent ward-registrar review and hourly monitoring; the failure to arrest the rise at that stage indicates the ward-level response was inadequate. My immediate action is to activate a MET call (or the equivalent local emergency response) to bring a critical care team to the bedside now, while I begin the ABCDE assessment and the resuscitation in parallel. I do not wait for the team to arrive before starting; the resuscitation and the call happen together.

(b) ABCDE and the first 30 minutes (5 marks): [1]

The ABCDE assessment is performed in order, treating life-threats as they are found. [1]

• Airway: patent — she is talking in short sentences. No intervention needed beyond positioning.
• Breathing: she is tachypnoeic (respiratory rate 30) with SpO2 of 88 per cent on room air. I apply high-flow oxygen via a reservoir mask at 15 L/min, titrating down to a target saturation of 92 to 96 per cent. She does not have documented chronic hypercapnic respiratory failure, so the standard target applies; I would send an arterial blood gas within 30 minutes to confirm. The chest X-ray shows progressive consolidation — the pneumonia is the source.
• Circulation: she is tachycardic (128 in new AF), hypotensive (84/52), with prolonged capillary refill (4 seconds), mottled peripheries, and oliguria (15 mL/hour). This is septic shock. I insert two large-bore cannulae, send bloods (full blood count, U&E, liver function, coagulation, troponin, lactate already 4.6), and blood cultures before antibiotics. I give a cautious 250 mL bolus of balanced crystalloid over 15 minutes (not the standard 500 mL, because she has moderate aortic stenosis and is at risk of pulmonary oedema), reassess, and repeat only if there is evidence of fluid responsiveness and no sign of overload (rising JVP, basal crackles, falling SpO2). The Surviving Sepsis Campaign 2021 guideline recommends 30 mL/kg crystalloid in the first three hours for sepsis with hypoperfusion, but in a patient with aortic stenosis I give this in smaller aliquots with reassessment between each, and I have a low threshold to start an early vasopressor (noradrenaline) if the blood pressure does not respond to the first bolus [4].
• Disability: GCS 14 with new confusion. Bedside glucose is essential — I check it now. Pupils normal. The confusion is part of the septic encephalopathy and will improve with treatment of the sepsis.
• Exposure: temperature 38.7, mottled peripheries. I examine the abdomen, the skin, and the lines for an alternative source; I confirm the chest findings.

In parallel, I deliver the Sepsis Six: oxygen, blood cultures, broad-spectrum IV antibiotics (escalating from her current regimen to a broader-spectrum agent such as piperacillin-tazobactam given the progression on first-line therapy, per local guideline), 30 mL/kg crystalloid in aliquots as above, lactate already measured, and hourly urine output via the catheter. [1]

(c) Hyperkalaemia and new atrial fibrillation (4 marks): [1]

The hyperkalaemia (potassium 5.8) with likely ECG changes (the question does not give the ECG but at 5.8 I would expect peaking of the T waves and possibly widening of the QRS) is a medical emergency. I give calcium gluconate 10 mL of 10 per cent intravenously over 5 to 10 minutes to stabilise the myocardial membrane, followed by insulin-dextrose (10 units of Actrapid in 50 mL of 50 per cent dextrose over 15 minutes) and a salbutamol nebuliser to shift potassium into cells. I recheck the potassium and the ECG within 30 minutes. If the hyperkalaemia is refractory or she is in established renal failure with a rising creatinine, she will need renal replacement therapy — another indication for ICU admission. [1]

The new atrial fibrillation is almost certainly secondary to the sepsis, the fever, the electrolyte disturbance (hyperkalaemia and likely hypomagnesaemia), and her underlying ischaemic heart disease and aortic stenosis. My priority is to treat the sepsis and correct the electrolytes; the AF will usually revert as the acute illness resolves. I would NOT give a beta-blocker or a calcium-channel blocker in her hypotensive state. If the AF is itself causing haemodynamic compromise that does not respond to the resuscitation, the options are amiodarone 300 mg intravenously over one hour (a negative inotrope but less hypotension than a calcium-channel blocker) or, if she is peri-arrest, synchronised DC cardioversion. The teaching point: in the deteriorating patient, treat the cause of the arrhythmia first; specific antiarrhythmic therapy is reserved for the arrhythmia that persists or that is itself the primary haemodynamic problem. [1]

(d) Drug review and withholding (3 marks): [1]

The drug chart is the highest-yield document in the unexplained deterioration, and in this patient several drugs must be withheld: [1]

• Metformin — withhold immediately. She has acute kidney injury (creatinine risen from 140 to 210) and septic shock; metformin accumulates in renal failure and in shock states and carries a risk of lactic acidosis. Her glucose is managed with a sliding-scale insulin infusion.
• Empagliflozin — withhold immediately. SGLT2 inhibitors carry a risk of euglycaemic diabetic ketoacidosis, are nephrotoxic in volume depletion, and have no place in the acutely deteriorating, hypovolaemic, septic patient.
• Any ACE inhibitor, ARB, or diuretic — withhold for the acute illness. These worsen hypotension and AKI.
• NSAIDs — withhold if she is on any; nephrotoxic in AKI.
• Any nephrotoxic antibiotic (e.g. an aminoglycoside) — review and substitute if possible. [1]

I review every drug on the chart, hold the harmful ones, and continue only what is essential (her antiplatelet and statin for the ischaemic heart disease, an appropriate antibiotic, the insulin sliding scale). [1]

(e) Escalation and the ceiling of care (3 marks): [1]

I call ICU outreach now — she has met the criteria (refractory hypotension, multi-organ failure, likely need for vasopressor and possibly renal replacement therapy). The decision about ICU admission weighs three factors: the reversibility of the acute problem (the pneumonia is potentially reversible with antibiotics and support), her baseline function (she is described as independent, living with her husband — favouring escalation), and her expressed wishes. I ask the family, gently and early, whether she has an advance care directive and what she would want. A previously independent 78-year-old with reversible sepsis is, in most institutions, a candidate for a time-limited trial of ICU care, with a defined review point at 48 to 72 hours and an honest conversation with the family throughout. I do not frame this as "all or nothing" — the trial of ICU is a defined period of maximal support after which we reassess, and if the treatment is not working, we refocus on comfort. [1]

(f) The common exam trap (2 marks): [1]

The trap in this patient is the fluid strategy in aortic stenosis. The standard sepsis fluid resuscitation is 30 mL/kg crystalloid in the first three hours — but in a patient with moderate aortic stenosis and ischaemic heart disease, an uncritical large-volume bolus risks precipitating acute pulmonary oedema and worsening the outcome. The correct approach is smaller aliquots (250 mL) with reassessment of fluid responsiveness and signs of overload between each, a lower threshold for early vasopressor (noradrenaline) support, and — if available — dynamic assessment (passive leg raise, focused echocardiography) to guide the fluid. The candidate who gives 30 mL/kg as a single bolus to this patient has applied the guideline without weighing the comorbidity. The second trap is treating the AF before treating the sepsis — the AF will not be controlled until the underlying cause is addressed. [1]

SAQ 2 — NEWS2 Interpretation and the Evidence for Rapid Response Systems (10 marks)

Prompt: A junior doctor asks you to explain: (a) the rationale for the NEWS2 score and how the response thresholds are set; (b) why the MERIT study found no significant reduction in the composite primary outcome despite an increase in MET calls; and (c) the synthesis of the evidence on rapid response systems that you would give to a hospital executive considering implementation. [1]

Model Answer

(a) Rationale for NEWS2 and the response thresholds (4 marks): [1]

NEWS2 is a track-and-trigger system that converts the subjective clinical judgement "this patient looks unwell" into an objective, reproducible, aggregate score, thereby reducing the inter-observer variability and the hesitation that delay the response to deterioration. It is built from seven physiological parameters — respiratory rate, SpO2 (with a second scale for hypercapnic respiratory failure), supplemental oxygen, temperature, systolic blood pressure, heart rate, and conscious level (AVPU/ACVPU) — each scored 0 to 3, with the sum giving a NEWS2 from 0 to 20. The thresholds are graded: a score of 0 is low-risk (routine 12-hourly monitoring); 1 to 4 is low (registered nurse review, 4- to 6-hourly monitoring); 5 to 6 is the key medium threshold (urgent ward-registrar review within 30 minutes, hourly monitoring); and 7 or more is a clinical emergency (emergency critical care assessment within minutes, continuous monitoring). A single score of 3 in any parameter — a "red score" — mandates urgent review regardless of the aggregate, because a single markedly abnormal value may be diluted to an apparently acceptable aggregate [2]. The score is a tool, not a substitute for clinical judgement — but used consistently, it closes the gap between deterioration and response.

(b) Why MERIT was negative despite increased calls (3 marks): [1]

The MERIT study — the only cluster-randomised controlled trial of a MET system, across 23 Australian hospitals — found that introducing the MET system increased emergency team calls from 3.1 to 8.7 per 1000 admissions but did not significantly reduce the composite primary outcome of cardiac arrest, unexpected death, or unplanned ICU admission (5.31 versus 5.86 per 1000 admissions, p=0.640) [3]. Several factors explain the negative result: the trial was underpowered for the composite outcome; there was contamination (the control hospitals increased their own MET-like calls during the study, diluting the between-group difference); the implementation period was short, and the cultural change required for a MET system to work takes longer to embed; and a composite outcome that includes unplanned ICU admission may rise as a consequence of appropriate escalation (more calls means more ICU admissions, which is not a failure). The negative primary outcome does not invalidate the system; it sets the context for the meta-analytic evidence.

(c) Synthesis for a hospital executive (3 marks): [1]

The synthesis I would give the executive is that rapid response systems have a consistent, meta-analytically supported effect in reducing non-ICU cardiopulmonary arrests by approximately a third (Chan's 2010 meta-analysis), with a less certain but plausible effect on overall hospital mortality. The mechanism of benefit is broader than the arrest count: the system changes the safety culture of the hospital, surfaces the deteriorating patient earlier, supports ward staff, and facilitates appropriate (not just more) escalation. The MERIT trial is the only randomised evidence and its primary outcome was negative, but the trial's limitations and the consistent before-and-after and meta-analytic evidence justify implementation. The implementation must attend to both limbs of the system: the afferent (recognition) limb — the vital sign measurement, the NEWS2, the staff empowered to call — and the efferent (response) limb — the team that arrives, the competencies it brings, the escalation pathway. A system with a strong response team but weak recognition will fail; a system with strong recognition but no one to respond will fail. The executive's investment must be in both, and in the culture that connects them. [1]