EDIC Part I common mistakes
EDIC Part I · Understand the traps and how to avoid them.
Calculating a derived cardiovascular or renal variable (cardiac index, oxygen delivery, filtration fraction, corrected calcium) from the numbers given but forgetting to convert units first - e.g. using a cardiac output in L/min directly as mL/min, or a creatinine in µmol/L where mg/dL is implied - and then selecting the answer option that matches the unconverted arithmetic.
Under time pressure candidates reach for the formula they memorised and skip the dimensional check; EDIC items deliberately supply mixed units (µmol/L, mg/dL, mmHg, cmH2O, mEq/L) so the distractor set contains the value produced by the unit error, which looks plausible and confirms the mistake.
What to do instead
Write the unit next to every number before substituting, and do a one-line dimensional sanity check on the result (CI should be roughly 2.5–4 L/min/m², DO2 roughly 500–600 mL/min/m², filtration fraction roughly 20%). If your answer sits an order of magnitude from the physiological range, you have a unit error, not a clinical insight.
Managing a vasopressor or inotrope question by naming the drug class alone ('start noradrenaline', 'add dobutamine') without committing to the target mean arterial pressure, the titration endpoint, and the fluid-status context in which the drug is being started.
Candidates learn drugs as isolated facts and answer the stem literally rather than as a haemodynamic scenario; EDIC questions are usually testing the decision rule (what MAP, what lactate clearance, what evidence of fluid responsiveness, when to add a second agent) rather than the pharmacology alone, so a bare drug name reads as incomplete.
What to do instead
For every vasoactive question, force yourself to state three things in order: the physiological target, the monitoring that confirms you reached it, and the trigger for escalating or de-escalating. Practise by converting each drug fact you revise into a 'start / titrate to / stop when' sentence.
Applying a fixed threshold (creatinine rise, PaO2/FiO2 ratio, platelet count, lactate) to classify severity without first checking whether the stem's values meet the definitional preconditions - baseline timing, rate of change, PEEP or FiO2 at the time of sampling, or whether the value is a delta or an absolute.
Definitions such as KDIGO AKI staging or the Berlin ARDS criteria are memorised as numbers but their qualifying clauses (48-hour versus 7-day windows, baseline creatinine source, minimum PEEP) are what the item is actually testing; candidates recognise the number and stop reading the qualifier.
What to do instead
Recite each staging definition as a conditional sentence with its time window and preconditions attached, not as a bare table. On the exam, underline the time reference and the ventilatory setting in the stem before choosing a stage or category.
Answering a data-science or statistics item by pattern-matching the named test ('use a t-test', 'it's a p-value problem') without identifying the data structure - paired versus independent, continuous versus categorical, the number of comparisons, or whether the question asks about confidence interval interpretation rather than significance.
Clinical training gives sparse exposure to study design, so candidates fall back on test names they have seen; EDIC data-science items are written so that the correct answer follows from the design (e.g. repeated measures, non-inferiority framing, multiplicity), and the named-test reflex lands on the distractor built for exactly that error.
What to do instead
Before answering, write one line: what is the outcome type, what is being compared, are observations paired, and how many tests. Then ask whether the question is about estimation (CI, effect size) or hypothesis testing. If the stem mentions many endpoints, suspect multiplicity.
Treating an ethics or end-of-life item as a request for the candidate's personal preference or for the 'kindest' option, rather than applying the operative framework - decision-making capacity, best-interests versus substituted judgement, necessity and proportionality, or the jurisdiction-neutral principle the stem invokes.
Ethics questions feel like opinion items, and candidates answer with what they would do clinically; the exam instead tests whether you can identify who holds decision-making authority, what standard applies when capacity is absent, and whether an intervention is proportionate.
What to do instead
For every ethics stem, answer three questions in writing first: does the patient have capacity, if not who is the legal substitute and under what standard, and is the proposed act necessary and proportionate. Choose the option that follows from that chain, not from the emotional pull of the scenario.
Recommending a feeding or rehabilitation intervention at the wrong point on the trajectory - full caloric goal on day one in a shocked patient, or early aggressive mobilisation in an unstable patient - instead of staging the intervention to haemodynamic and metabolic stability.
Guideline headlines ('early enteral nutrition', 'early mobilisation') are remembered without their contraindications and staging rules, so candidates apply the headline to a stem that was constructed to test the exception.
What to do instead
Attach to each feeding and mobilisation recommendation the explicit 'not until' condition: feeding until shock is controlled and electrolytes corrected, mobilisation until respiratory and cardiovascular stability criteria are met. When a stem describes instability, look first for the option that stages or delays.
Missing the interaction in a pharmacology item - a drug's effect altered by renal or hepatic impairment, by an interacting co-prescribed agent, or by altered protein binding - and answering as if the drug were being given to a healthy volunteer at standard dose.
Candidates revise mechanisms and side effects but under-revise dose adjustment and interaction tables; EDIC stems routinely embed the modifying factor in the past medical history or concomitant medication list, which is easy to skim past.
What to do instead
Before answering any pharmacology item, scan the stem specifically for organ function, age, weight, and every other drug listed, and ask whether each changes clearance, protein binding, or the pharmacodynamic target. Build a personal list of the interactions you keep missing and revise that list, not the whole formulary.
Reading a sepsis or infection stem as a request for the correct antibiotic name when it is testing source control, timing, or the diagnostic pathway - and choosing an antimicrobial option over the option that addresses the undrained collection, the wrong culture timing, or the need to de-escalate.
Antimicrobial choice is the most concrete and memorable part of infection management, so it dominates recall; questions about source control and stewardship are less rehearsed and get answered with the most specific-sounding drug.
What to do instead
Classify the infection question first: is it about diagnosis, timing, source control, empirical choice, or de-escalation? If the stem describes a collection, a device, or a non-resolving course, source control is the answer until proven otherwise. Justify any antibiotic answer by naming the organism coverage gap it closes.
Answering neuro-intensive-care items by reaching for the latest specific intervention (osmotherapy dose, targeted temperature protocol, specific sedation agent) without first securing the physiological priorities the stem is testing - airway, PaCO2, MAP and cerebral perfusion pressure, sodium and glucose.
Protocol details are vivid and easy to recall, whereas the ordering of priorities is abstract; items are often built so that a specific therapy is only correct after the physiological derangement has been addressed, and candidates jump to step three.
What to do instead
For every neuro stem, run a fixed checklist before looking at the options: airway and ventilation, PaCO2, oxygenation, MAP/CPP, sodium, temperature, glucose, seizure control. Eliminate any option that presupposes a step you have not yet secured.
Treating a health-services-research or outcome item as a clinical management question - answering 'what should be done for this patient' when the stem asks about a population-level measure such as sensitivity, number needed to treat, attributable risk, or the effect of a service configuration on outcome.
The clinical reflex is strong and the stem may open with a patient vignette, so candidates answer the vignette rather than the question actually posed; outcome-research items deliberately use clinical framing to test whether you can shift level from individual to population.
What to do instead
Read the final sentence of the stem twice and label the level of the question - individual or population - before reading the options. If the words 'screening', 'service', 'cohort', 'NNT', or 'proportion' appear, compute the population measure explicitly rather than reasoning clinically.