FMGE common mistakes
FMGE · Understand the traps and how to avoid them.
Answering integrated clinical vignettes from pattern recognition on the last two lines of the stem, without naming the single best-supported diagnosis before looking at the options.
FMGE items are deliberately written so that the final clause (a lab value, a duration, a drug history) changes the answer; candidates trained on one-line recall questions skim to the keyword and choose the 'classic' association. Time pressure in a single-sitting paper reinforces this shortcut.
What to do instead
Adopt a fixed 20-second read protocol: underline the age/sex, the duration, the one abnormal value, and the exact question asked ('most likely diagnosis' vs 'next step' vs 'most appropriate investigation'). State the diagnosis in your head before reading options, then eliminate options that do not fit the stated duration or the stated question type. Practise 50 vignettes where the last line flips the answer from the obvious association to the correct one.
Choosing the 'next best step' as the definitive treatment instead of the immediate stabilising step, or vice versa, in emergency and management items.
Candidates memorise disease–drug pairs but not the sequence of management. They also conflate 'best initial test' with 'most specific test' and 'definitive treatment' with 'first thing to do'.
What to do instead
Build one-page management ladders for high-yield emergencies (upper GI bleed, DKA, status epilepticus, acute severe asthma, shock, obstetric haemorrhage, raised ICP, acute coronary syndrome) as numbered sequences: stabilise → confirm → treat → long-term. For each step, write the one phrase that identifies it. Then answer 30 items and label each option as 'step 1', 'step 2', 'step 3' or 'wrong' before selecting.
Guessing on calculation items (dilutions, drug doses, paediatric fluid and drug dosing, acid–base, epidemiological measures) rather than working them out on paper.
The exam allows rough work but candidates try to do arithmetic mentally under time pressure, or they have never practised the specific formula families used in community medicine (odds ratio, attributable risk, sensitivity/specificity, sample size) and paediatrics (maintenance fluids, defibrillation energy, drug mg/kg).
What to do instead
Keep a written formula card: prevalence vs incidence, sensitivity/specificity/PPV/NPV, relative risk vs odds ratio, attributable and population attributable risk, number needed to treat, standard error and confidence interval, 4-2-1 maintenance fluid rule, Holliday-Segar, and the common paediatric resuscitation doses. Do 10 timed calculations per day with the working shown, and always sanity-check the units and whether the answer is plausible.
Treating pharmacology as a list of drug names without attaching mechanism, class-specific adverse effect, and contraindication to each drug.
Undergraduate revision often stops at 'drug of choice' tables. FMGE items frequently test the mechanism behind an adverse effect (why does an ACE inhibitor cause cough, why does metoclopramide cause dystonia) or a contraindication in a named comorbidity.
What to do instead
For each of the high-yield classes (antihypertensives, antidiabetics, antimicrobials, anticoagulants, anti-TB drugs, antipsychotics, antidepressants, antiepileptics, chemotherapeutic agents), write three columns: mechanism in one line, signature adverse effect with its mechanism, absolute contraindication. Add the antidote or reversal agent where one exists. Review the card weekly and self-test by covering the drug name and reading only the mechanism.
Learning pathology and microbiology as isolated facts instead of linking morphology and organism to the clinical syndrome, the investigation, and the treatment.
Pre-clinical preparation rewards rote memorisation of lists, so candidates can name a tumour marker or a culture medium but cannot move from a clinical picture to the organism or from a histology description to the diagnosis. Integrated items then feel unfamiliar.
What to do instead
Convert each high-yield topic into a four-column chain: risk factor → pathogenesis/morphology → clinical syndrome → investigation and treatment. Do this for the major carcinomas, glomerulonephritides, anaemias, and for organisms such as Mycobacterium tuberculosis, Salmonella, Staphylococcus aureus, hepatitis viruses, and malaria. Test yourself by starting from any single column and reconstructing the other three.
Reading forensic medicine and toxicology only as a list of sections and poisons, without the clinical toxidrome and the medicolegal reasoning that items actually ask for.
This subject is often deferred to the end of preparation and skimmed. Candidates memorise poison names but not the toxidrome, the specific antidote, the sampling timing, or the difference between cause, manner and mode of death, so they lose the reasoning items.
What to do instead
Build toxidrome tables (cholinergic, anticholinergic, opioid, sympathomimetic, sedative-hypnotic) with clinical signs, likely agents, antidote and supportive measures. Separately learn the definitions that items hinge on: cause/manner/mode of death, lucid interval, cadaveric changes with timelines, age of wounds, and the legal thresholds (consent, negligence, insanity, dying declaration). Answer 20 short reasoning items and write the justification in one sentence.
Skipping anatomy and physiology diagrams and imaging descriptions because they cannot be 'read quickly', leaving whole question blocks unattempted or guessed.
Candidates avoid topics that require spatial or physiological reasoning when time is short, and they never practise translating a described section, nerve course, or pressure–volume relationship into an answer without a picture.
What to do instead
Practise text-only anatomical reasoning: describe a structure by its relations, its blood supply, its nerve supply and the deficit produced by its injury, then check against an atlas. For physiology, work through the pressure–volume loop, the nephron segment functions, the cardiac cycle timings and the respiratory mechanics as written sequences. Ten such text-only reconstructions per week removes the dependency on images.
Answering obstetrics and gynaecology items with a single diagnosis and ignoring the gestational age, parity, and the stated investigation finding that determine the management.
Management in obstetrics is stage-dependent and time-critical; candidates apply the same answer across trimesters or across the latent and active phases of labour, and they confuse the interpretation of a cardiotocograph with the action required.
What to do instead
For each major topic (hypertensive disorders, antepartum haemorrhage, postpartum haemorrhage, preterm labour, diabetes in pregnancy, multiple pregnancy, abnormal labour, gynaecological malignancies) draw a trimester- or stage-wise decision tree with the defining finding at each node. Learn CTG interpretation as a fixed sequence (baseline rate, variability, accelerations, decelerations, then the action). Rehearse 15 items where the only change is the gestational age.
Leaving community medicine as an afterthought and then guessing at epidemiology, biostatistics, and national health programme details.
It feels non-clinical and is postponed, yet it is a large, high-yield, and highly predictable component. Candidates also confuse definitions (incidence vs prevalence, screening vs diagnostic test, epidemic vs endemic) and programme-specific thresholds they never checked.
What to do instead
Allocate fixed weekly time to community medicine from the start. Master the calculation families with written practice, learn the study-design hierarchy and their biases as a table, and maintain a one-page summary of national programme eligibility criteria, immunisation schedule, and notifiable diseases. Use the most recent programme guidelines available and note the source and year beside each fact so you know what is current.
Spending the final weeks re-reading notes instead of doing timed full-length mock papers with structured review of every wrong and every guessed answer.
Re-reading produces a feeling of familiarity that is mistaken for recall, and it never exposes the reasoning errors or the time allocation problems that actually cost marks. Candidates also review only the wrong answers and ignore lucky guesses.
What to do instead
In the last six to eight weeks, sit at least two full-length timed papers per week under exam conditions. Review every question: mark each as confident-correct, guessed-correct, or wrong, and for the latter two write the specific reason (missed last line, wrong management step, unknown fact) and the corrected fact on a single running error log. Re-test the error log weekly and track whether the same failure mode recurs; adjust time per section based on where you overran.