Anesthesiology Question Quality Review

Executive Summary

The 200-question candidate sample for Anesthesiology was reviewed across eight shards of 25 questions each. The sample draws from a pool of 3,585 questions and excludes the 16 benchmark and recent PYQ items that define the quality bar.

The headline finding is a severe Bloom's level imbalance: 56 of the 200 candidate questions sit at Bloom's 1 and 99 at Bloom's 2, meaning roughly 77% of the sample tests recall or basic comprehension. The benchmark and PYQ gold standard is predominantly Bloom's 3–4 clinical application and analysis. This gap is not a marginal calibration issue — it represents a structural mismatch between what the pool contains and what INI-CET and NEET-PG actually test.

Beyond the Bloom's distribution, five operationally distinct problem types were identified across the shards. These are not evenly distributed: the recall overload and the image-integrity failure are broad patterns appearing in nearly every shard; factual errors and answer-key defects are a narrower but high-risk cluster; topic fragmentation and near-duplication are concentrated in a few high-density pharmacology areas; and a small but meaningful set of questions are simply out of scope for the subject.

Approximately 60–70 questions in the sample are strong enough to keep without change. Roughly 60–70 are fixable with targeted edits. The remaining 60–80 should be disabled, either because they are factually unsafe, structurally broken, trivially easy, or redundant with better-covered concepts.

What Good Looks Like

The benchmark and PYQ items establish a clear quality bar. The best items in the gold set share four properties that the candidate sample frequently lacks.

Clinical embedding. The question places the candidate inside a patient encounter. The vecuronium reversal question (2bdbe2e6) gives a 50-year-old male, a specific surgery, and a specific agent — the pharmacology is tested through a decision, not a definition. The ARDS ventilator question (b47460b0) gives vital signs, physical findings, and a hemodynamic picture before asking for a cause. The ventilator management PYQ (77d06928) provides a full ABG, weight, PEEP, tidal volume, and SpO₂ before asking for the next step. Even the simpler items (ecb677e0 on oxygen delivery devices, c490b26b on LMA generations) are framed as applied knowledge questions rather than vocabulary tests.

Distractors that require knowledge to reject. In the rocuronium reversal PYQ (aabaa85d), edrophonium, glycopyrrolate, and neostigmine are all plausible to a candidate who has not studied sugammadex's specific mechanism. In the massive transfusion PYQ (a6caa61e), hypokalemia, hypothermia, and hypocalcemia are all genuine complications — only hypernatremia is not, and distinguishing it requires understanding the sodium content of stored blood. Distractors in the gold set are wrong for a reason, not merely unfamiliar.

Unambiguous correct answers. The gold items do not have two defensible answers. The CPR question (49b037d4) marks ventilation at 22–25/min as incorrect because the guideline figure is 10/min — the error is specific and the correct answer is not a matter of interpretation.

Appropriate Bloom's level for the concept tested. The potency-ranking PYQ (3585f80b) is correctly tagged Bloom's 4 because it requires ordering four agents by MAC, not just naming one. The FiO₂ device question (ecb677e0) is correctly Bloom's 1 because it is a classification fact — but it is used sparingly and the concept is genuinely tested in exams. The gold set does not use Bloom's 1 as a default; it uses it deliberately for high-frequency exam facts.

The candidate sample should be evaluated against this standard throughout.

Main Issue Categories

1. Recall-Only Questions With No Clinical Anchor

Why this pattern is bad

INI-CET and NEET-PG do not ask candidates to recite pharmacological facts in isolation. They present a patient, a clinical situation, or a monitoring finding and ask what to do or why something happened. A question that asks "which drug is a dissociative anesthetic?" or "what percentage of thiopentone is used for induction?" tests the same knowledge that a first-year MBBS student would have after reading a single paragraph. It does not discriminate between a candidate who understands anesthesia and one who has memorized a list. At scale, a pool dominated by these items trains candidates to memorize rather than reason, which is the opposite of the stated pedagogical goal.

How it shows up

This is the broadest pattern in the sample, appearing in every shard. It is not confined to one topic area — it spans inhalational agents, intravenous agents, local anesthetics, neuromuscular blockers, equipment, and monitoring. The Bloom's 1 count of 56 and Bloom's 2 count of 99 in the candidate set confirm this is a pool-wide phenomenon, not a cluster in one corner of the subject.

Characteristic features: no patient age, no clinical setting, no vital signs, no decision required. The stem is a direct question ("Which drug…?", "What is the…?", "Which of the following is…?") with one factually correct answer and three distractors that no informed candidate would choose. Many carry both "easy" and "blooms-1" quality flags simultaneously.

Example question IDs

57a8df2d — "Typical percentage of thiopentone used for induction?" Single numerical fact, no context, flagged easy+Bloom's 1. The answer (2.5%) is on the first page of any pharmacology chapter.
e66115ea — "Which is a dissociative anesthetic agent?" Ketamine as answer. Bloom's 1, flagged easy. Any candidate who has opened a textbook knows this.
7417c224 — "Which IV anesthetic possesses local anesthetic properties?" Ketamine again. Bloom's 1, flagged easy. Near-duplicate concept to e66115ea and 1a77e9c9 in the same shard.
642f85fa — "Post-operative shivering — which drug?" Pethidine. Bloom's 1, flagged easy, no clinical framing.
2c873225 — "Most common side effect of spinal anaesthesia?" Hypotension. Bloom's 1, flagged easy. Answered correctly by rote.
52b938d0 — "O₂ concentration in mouth-to-mouth?" 16%. Bloom's 1, no clinical application, pure trivia.
90ebdc07 — "GCS components except bladder function." Bloom's 1, flagged easy. Zero discriminative value for PG-level assessment.
4e22e222 — "Type E circuit used for paediatric anaesthesia." Bloom's 1, flagged easy, single-fact recall.
29c292e0 — "Central venous catheter does not measure cardiac output." Trivially obvious, Bloom's 1, flagged easy.
ee54e685 — "Maximum O₂ concentration from Venturi mask = 60%." Bloom's 1, flagged easy. The PYQ benchmark (ecb677e0) already covers this concept at a higher level.
2305f0ac — "What is NOT in Ringer's lactate?" Pure composition recall, Bloom's 1, flagged easy.
dab129b2 — "Pin index code for nitrous oxide (3,5)." Rote memorization of equipment codes, Bloom's 1, flagged easy, no tags.
4bce0367 — "Minimum O₂ concentration in anaesthesia (33%)." Bloom's 1, flagged easy, no tags.
566dcc93 — "Characteristic EEG pattern in surgical tolerance stage: Delta." Bloom's 1, flagged easy, isolated trivia.
8094cfa2 — "Gas stored in liquid form is: Nitrous oxide." Bloom's 1, flagged easy. CO₂ is also stored as liquid, making the answer debatable without a qualifier.

Recommended disposition

Questions in this category that test concepts already covered by a gold-standard PYQ or benchmark item should be disabled — there is no value in keeping a weaker version of a concept that is already well-tested. Questions that test a genuinely high-yield concept (e.g., pethidine for shivering, atracurium in renal failure) should be fixed by embedding the fact in a brief clinical scenario that requires the candidate to apply rather than recall. The threshold for fixing versus disabling should be: if the concept appears in recent PYQs and the current question is the only coverage, fix it; if the concept is already covered by a better item, disable.

2. Factual Errors and Unsafe Answer Keys

Why this pattern is bad

A question with a wrong correct answer is worse than no question at all. It actively miseducates candidates, erodes trust in the platform, and — if the item reaches a scored test — penalizes candidates who know the correct answer. In a subject like anesthesiology where drug dosing, toxicity management, and monitoring interpretation have direct patient-safety implications, factual errors are not minor editorial issues.

How it shows up

This is a narrower cluster than the recall overload, but it appears in every other shard and involves a range of error types: wrong correct answer, unit errors, historically incorrect attribution, and options that are internally inconsistent. Several errors involve high-stakes clinical content (local anesthetic toxicity treatment, ventilator pressure interpretation, bupivacaine toxicity management) where a wrong answer could reinforce dangerous clinical misconceptions.

Example question IDs

da0f6ddd — Bupivacaine toxicity treatment: the correct answer is listed as "5% dextrose." The established evidence-based treatment is 20% lipid emulsion (Intralipid). This is a patient-safety-relevant factual error in a high-stakes clinical scenario.
d935b2a1 — "Who demonstrated the anesthetic effect of ether?" marks "Moon" as correct. The historically accepted answer is William Morton (1846 public demonstration at Massachusetts General Hospital). No standard anesthesia textbook attributes this to "Moon." This is a factual error in a history question — doubly problematic because the question has no clinical value to begin with.
0c76505f — "Increased peak AND plateau pressures" marks "obstruction of tracheal tube" as correct. This is mechanistically wrong: simultaneous rise in both peak and plateau pressure indicates decreased lung/chest wall compliance (e.g., pneumothorax, pulmonary edema, endobronchial intubation). Tracheal tube obstruction raises peak pressure but not plateau pressure. This is a core ventilator mechanics concept tested in PYQs.
a4b1507e — "Most commonly used concentration of Xylocaine HCl for local anesthesia" marks 1:80,000 as correct. This is the concentration of epinephrine in a lidocaine-adrenaline combination, not the lidocaine concentration itself. The question conflates two different measurements.
c9fca928 — Marks ammonium chloride as NOT useful for metabolic alkalosis. This is factually incorrect; ammonium chloride is a recognized acidifying agent used in metabolic alkalosis treatment.
eaf22865 — "Most common complication of Bier block" lists hypotension as correct. The recognized primary complication is systemic local anesthetic toxicity (LAST) or tourniquet-related pain, not hypotension.
243b9972 — Ketamine contraindicated in epilepsy: ketamine is used cautiously in epilepsy and is not a standard contraindication; thiopentone and propofol are the anticonvulsant agents. The answer is factually contestable.
a406ea09 — PEEP in ARDS: all options use "mm Hg" as the unit. PEEP is universally measured in cm H₂O. This is a unit error that would confuse candidates learning the concept.
9e12095c — "No epileptogenic potential → Desflurane." Isoflurane is more commonly cited as the safest volatile agent in epilepsy in standard Indian PG references; the claim about desflurane is not universally supported.
b5c94c5a — "Fastest acting inhalational agent = sevoflurane." Desflurane has a lower blood:gas partition coefficient (0.42 vs. 0.65 for sevoflurane) and is faster in many references. The answer is contestable and may be wrong depending on the reference used.
e6debc1e — "All of the above" as correct answer for local anesthetic properties, but option C ("low protein binding → more active") contradicts standard teaching that high protein binding correlates with longer duration of action. The question contains an internally inconsistent correct answer.

Recommended disposition

All questions with a demonstrably wrong correct answer should be disabled immediately pending expert review. Questions where the answer is contestable depending on reference source (9e12095c, b5c94c5a, 243b9972) should be flagged for expert review before any further use and treated as disabled in the interim. Unit errors (a406ea09) and conflation errors (a4b1507e) are fixable if the underlying concept is sound, but require careful rewriting rather than minor edits.

3. Image-Dependent Questions With Missing or Unverifiable Images

Why this pattern is bad

A question that references "the device shown" or "the ECG findings given" without an embedded image is not a question — it is an unanswerable prompt. In the best case, a candidate can guess the answer from context clues in the options; in the worst case, the question is completely non-functional. This is a structural defect, not a content quality issue, but it is operationally urgent because these items cannot be used in any delivery format until the image is confirmed to be attached and rendering correctly.

The gold-standard PYQ set includes image-based questions (f45d6ce7 on capnography tracing) that work correctly because the image is embedded. The candidate sample contains multiple items that attempt the same format but fail at the delivery layer.

How it shows up

This pattern appears in at least four shards and involves at least six distinct questions. The failure mode is consistent: the stem references a visual stimulus using phrases like "the device shown," "the given ECG findings," "the mode of ventilation shown here," or "the image," but no image is present in the question text. In some cases the question is a PYQ (6a601ef7 is tagged NEET-PG-2024) that was presumably functional in its original exam context but has lost its image in the content pipeline.

Example question IDs

3832e3a3 — "Vital signs assessment technique shown in the image." No image present. The question is completely non-functional as text.
274cb04e — "The device shown can deliver FiO₂ up to ___." No image. The answer (100%) is unverifiable without knowing which device is depicted.
4c7e1aa2 — "What is the mode of ventilation shown here?" No ventilator waveform image. The answer (SIMV) cannot be derived from the text alone.
ba6f79e7 — "All of the following are indicated in management of a patient with the given ECG findings, except?" No ECG image. The question is unanswerable without knowing the arrhythmia.
6a601ef7 — "What is the pressure required to open the device shown in the image?" Tagged NEET-PG-2024. No image present. This is a PYQ that has lost its image.
b1d845dc — "Which maneuver helps to improve laryngoscopic view?" The correct option reads "Improve laryngoscopic view" — a tautological restatement of the question stem. This suggests the image (likely showing a specific maneuver) was the intended anchor for the correct answer, and without it the question is both non-functional and self-referential.

Recommended disposition

All image-dependent questions should be disabled until the image is confirmed to be correctly attached and rendering in the delivery environment. Questions where the image can be identified and reattached (e.g., 274cb04e likely shows a non-rebreathing mask; 6a601ef7 likely shows an APL valve) should be fixed by attaching the correct image and verifying rendering. Questions where the image cannot be identified or where the text-only version is non-functional even with context (b1d845dc) should be disabled and rewritten from scratch as text-based questions. A systematic audit of all image-tagged questions in the full 3,585-question pool is warranted given the frequency of this failure in the sample.

4. Topic Fragmentation and Near-Duplication Within High-Density Pharmacology Areas

Why this pattern is bad

When the same pharmacological fact is tested by three or four separate questions in slightly different framings, the pool wastes coverage capacity on a single concept while leaving other high-yield areas thin. For candidates, encountering the same concept repeatedly in different surface forms creates an illusion of breadth while actually narrowing the knowledge tested. For test assembly, having five questions on ketamine's properties and one question on obstetric anesthesia complications means that any test drawing from this pool will be skewed toward the over-represented concept.

This is distinct from the recall overload problem: some of the duplicated questions are individually reasonable items. The problem is the density, not the quality of any single item.

How it shows up

Three pharmacology areas show clear fragmentation in the sample: ketamine properties, thiopentone/induction agents, and malignant hyperthermia. In each case, multiple questions test the same underlying fact from slightly different angles without adding meaningfully different test points.

Ketamine fragmentation (observed in shard 001): Questions 7417c224 (ketamine has LA properties), e66115ea (ketamine is a dissociative agent), 1a77e9c9 (ketamine NOT contraindicated in shock), ac49277c (ketamine for hypotensive patient), and 6366665c (ketamine for hemorrhagic shock) all converge on the same two or three facts about ketamine. The first two are pure recall; the last two are reasonable clinical vignettes. The pool does not need five questions to cover ketamine's key properties.

Thiopentone/induction agent fragmentation (observed in shards 001, 002, 004): Questions 57a8df2d (thiopentone percentage), 748c298e (which barbiturate for induction), 44fe3a7d (best induction agent), and 2008b9c2 (why thiopentone for induction) all test the same narrow concept. Questions 748c298e and 44fe3a7d are near-duplicates of each other.

Malignant hyperthermia fragmentation (observed in shard 002): Questions 3cccd6b8 (MH mechanism via ryanodine receptor), 17395fee (most reliable sign of MH), and e33302b0 (MH triggering agents) appear within a single 25-question shard. While MH is a high-yield topic, three questions in 25 represents a density that, extrapolated to the full pool, suggests significant over-coverage.

CPR parameter duplication (visible in the PYQ set itself): PYQs 49b037d4 and 60f77ea1 both test incorrect CPR parameters (ventilation rate and compression depth respectively). These are flagged as duplicates in the gold set, suggesting the duplication problem extends into the PYQ layer.

Example question IDs

748c298e and 44fe3a7d — Near-identical questions: "which barbiturate for induction" vs. "best induction agent," both with thiopentone as the answer. One should be disabled.
7417c224, e66115ea, 1a77e9c9, ac49277c — Four questions covering ketamine's properties with significant conceptual overlap. At most two (the best clinical vignettes) should be retained.
3cccd6b8, 17395fee, e33302b0 — Three MH questions in one shard; the full pool likely contains many more. The mechanism question (3cccd6b8) is the strongest; the others should be consolidated into a single clinical recognition vignette.
ce2d301e and 7cf9f041 — Both test histamine release from neuromuscular blockers (d-tubocurarine). 7cf9f041 is a well-constructed clinical vignette; ce2d301e is bare recall. The latter is redundant.

Recommended disposition

For each fragmented cluster, identify the one or two strongest questions (preferably clinical vignettes at Bloom's 3+) and keep those. Disable the weaker, more recall-oriented duplicates. Do not attempt to fix near-duplicate recall questions into vignettes when a good vignette already exists for the same concept — that creates a third version of the same test point.

5. Clinically Ambiguous or Structurally Flawed Question Construction

Why this pattern is bad

This category covers questions where the concept being tested is legitimate and the topic is appropriate, but the question itself is broken in a way that makes it either unanswerable, multiply answerable, or misleading. Unlike factual errors (Category 2), these questions are not necessarily wrong — they are poorly constructed. The practical effect is the same: candidates who know the material may be penalized, and candidates who do not may guess correctly for the wrong reasons.

How it shows up

Several distinct construction failures appear across the shards: ambiguous correct answers where two options are defensible, "None of the above" and "All of the above" as correct answers, negative framing ("NOT/EXCEPT") used without adding discriminative value, option sets that are incomplete (missing the textbook-standard answer), and stems that are grammatically incomplete or contextually underspecified.

Example question IDs

9227d65c — "Preferred position for spinal anesthesia for C-section" marks lateral decubitus as the single correct answer. Both lateral decubitus and sitting position are widely used and accepted in obstetric spinal anesthesia. Marking one as definitively correct without a guideline qualifier is clinically debatable.
c969389c — OSA patient for knee replacement: all four options (minimizing opioids, ensuring ventilation, regional anesthesia, cautious sedation) are individually correct management strategies. There is no single best answer without additional clinical context specifying the decision point.
b007788d — NMBs safe in renal failure marks "vecuronium and mivacurium" as correct, but atracurium (Hofmann elimination, the textbook "safest" choice in renal failure) is absent from the options entirely. The option set is incomplete in a way that makes the question misleading.
d0a69dc6 — "Triad of GA" marks "analgesia, amnesia, and relaxation" as correct, but the classic Woodbridge triad includes unconsciousness/narcosis. The distractor "narcosis, analgesia, and relaxation" is also defensible. Two options are correct depending on which authority is cited.
2008b9c2 — Thiopentone for induction: options B and C are near-duplicates ("Rapidly redistributed" vs. "Rapid redistribution"). The intended correct answer ("smooth induction") is clinically imprecise; the pharmacokinetically accurate reason is rapid redistribution. The question has both a duplicate distractor and a contestable correct answer.
e862107e — "Gas cylinder with single pin index?" The stem is a grammatically incomplete fragment. No clinical framing. Needs to be rewritten as a complete sentence before it can be evaluated on content.
b1d845dc — Correct option reads "Improve laryngoscopic view" — a verbatim restatement of the question stem. This is a tautological giveaway that provides no information about which maneuver is correct.
89b9792a — "Cannot be prevented by which of the following?" is awkward double-negative framing. Diltiazem as a distractor for laryngospasm prevention is unexplained and unusual.
549ef1f8 — Concentration of adrenaline with local anesthetics: 1:200,000 is marked correct, but 1:100,000 is equally standard and more commonly cited in Indian surgical and dental texts. Without specifying the clinical context, both are defensible.
c987b6b8 — "Sevoflurane is more potent than isoflurane" is marked incorrect (correct answer). Two options in the question convey the same fact (sevoflurane has higher MAC than isoflurane, and sevoflurane is less potent than isoflurane), creating a structural flaw where two distractors say the same thing.

Recommended disposition

Questions with ambiguous correct answers due to missing clinical context (9227d65c, 549ef1f8, d0a69dc6) should be fixed by adding a specific qualifier or guideline reference that makes one answer unambiguously correct. Questions with incomplete option sets (b007788d) should be fixed by revising the options to include the standard textbook answer. Questions with tautological correct options (b1d845dc) or grammatically broken stems (e862107e) should be fixed if the underlying concept is high-yield, or disabled if the concept is already covered by a better item. Questions where all options are correct (c969389c) should be fixed by restructuring the stem to specify a single decision point, or disabled if the concept cannot be cleanly isolated.

6. Out-of-Scope and Low-Yield Trivia Content

Why this pattern is bad

A small but consistent cluster of questions tests content that is either outside the scope of the Indian PG anesthesiology curriculum or so peripherally relevant that it has no realistic chance of appearing in INI-CET or NEET-PG. These questions consume pool capacity, inflate the apparent breadth of coverage, and — if they reach candidates — waste study time on content that will not be tested.

This is distinct from the recall overload problem (Category 1): those questions test real anesthesiology content at too low a level. This category tests content that should not be in the anesthesiology pool at all, or that is so historically obsolete or clinically irrelevant that it fails the basic relevance test for PG entrance preparation.

How it shows up

Three sub-types appear in the sample: history-of-medicine trivia (who first used ether, who demonstrated ether's effects), dental/TMJ content misclassified as anesthesiology, and equipment-code memorization with no clinical application.

Example question IDs

d935b2a1 — "Who demonstrated the anesthetic effect of ether?" marks "Moon" as correct (also a factual error — see Category 2). History-of-medicine trivia with zero clinical relevance to PG entrance preparation.
9f0b95bf — "Who was the first dentist to use ether?" Morton is the correct answer. History trivia, no clinical application, not tested in INI-CET or NEET-PG.
621d2586 — "Myofascial pain dysfunction — masticatory pain." Content is dental/TMJ, not core anesthesiology. No tags, no template membership. Misclassified into the pain management topic.
c5d01b8c — "Incisive nerve block anatomy." Dental regional anesthesia at Bloom's 1. Not relevant to the Indian PG anesthesiology curriculum tested at INI-CET/NEET-PG level.
dab129b2 — "Pin index code for nitrous oxide (3,5)." Rote memorization of equipment codes with no clinical application. Not tested in recent PYQs.
c0a3d0a0 — "Which inhalational anesthetic possesses analgesic properties?" with Xenon as the answer. Xenon is not in routine Indian clinical practice or standard Indian PG pharmacopeia. Low yield; the concept (analgesic inhalational agent) is better tested with nitrous oxide in a clinical context.
ec76ed53 — "Bronchoscope is best sterilized with?" Bloom's 1, belongs to infection control/microbiology rather than anesthesiology. No anesthetic reasoning required.
13f8f83d — "EEG beta waves = awake state." Bloom's 1 pure recall with no anesthesia-specific application. Only relevant if reframed around BIS/depth-of-anesthesia monitoring.

Recommended disposition

All history-of-medicine trivia questions should be disabled — they are not tested in INI-CET or NEET-PG and have no clinical value. Dental/TMJ content (621d2586, c5d01b8c) should be disabled and, if the content is valid, reassigned to the appropriate subject. Equipment-code memorization questions (dab129b2) should be disabled unless they can be reframed as a clinical safety scenario (e.g., why pin-index systems prevent gas mix-ups). Questions that test a real concept but in an irrelevant context (c0a3d0a0 on Xenon, 13f8f83d on EEG) should be disabled and replaced with questions on the clinically relevant equivalent (nitrous oxide analgesia, BIS monitoring).

7. Correct Concept, Insufficient Clinical Depth — Upgradeable Items

Why this pattern is bad

This is the most actionable category for the content team. These questions test real, high-yield anesthesiology concepts that appear in PYQs and benchmark items, but they do so at Bloom's 1–2 when the concept is capable of supporting a Bloom's 3–4 clinical vignette. They are not wrong, not duplicative, and not out of scope — they are simply underperforming relative to their potential. Disabling them would remove coverage of important concepts; fixing them would convert them into genuinely discriminating items.

How it shows up

The pattern is most visible in topics where the gold-standard PYQs demonstrate the correct format: neuromuscular blockade and reversal, regional anesthesia complications, preoperative evaluation, and inhalational agent pharmacology. In each of these areas, the candidate sample contains bare-recall versions of concepts that the PYQ set tests through clinical scenarios.

Example question IDs

17395fee — "Most reliable sign of MH → Hypercarbia." Factually correct and PYQ-adjacent, but the stem is bare recall. The concept supports a strong Bloom's 3 vignette: a patient under volatile anesthesia develops rising ETCO₂ and muscle rigidity — what is the most reliable early sign? Fix by adding clinical context.
6354a3e4 — "Dose of non-depolarizing relaxants in neonates is decreased." A bare fact that could become a Bloom's 3 question by adding a neonatal surgical scenario and asking the candidate to select the appropriate dose adjustment rationale (immature NMJ, reduced protein binding, larger volume of distribution).
e3a0b155 — Ambulatory anesthesia drug combination question is conceptually sound (mivacurium, fentanyl, sevoflurane for daycare surgery) but reads as a list-matching exercise. Adding a daycare surgery scenario would raise it to Bloom's 3.
b8f94939 — "Lacrimation during anesthesia indicating Stage III." Factually correct but stripped of clinical context. A brief scenario (anesthesiologist notices lacrimation during induction) would make this a genuine clinical-reasoning question.
4ac73775 — "Tracheotomy dead space reduction: 30–50%." Correct but bare recall. A patient with high work of breathing post-tracheotomy would provide the clinical hook needed to justify inclusion.
946fc94c — "Fast induction and recovery is seen in?" Bare-stem recall. Fixable by converting to a clinical scenario: a patient requiring brief procedural sedation where rapid offset is needed, asking which agent is preferred.
ccd1824a — "Which anesthetic causes muscle rigidity?" with alfentanil as answer. The concept (opioid-induced chest wall rigidity) is high-yield and clinically important, but the question is too superficial. A clinical scenario involving intraoperative rigidity management would make this a strong Bloom's 3 item.
0fb5014f — "Asthmatic patient for colon surgery: preferred anesthetic agent." This is already a reasonable Bloom's 3 question but could be strengthened by specifying the intraoperative trigger (e.g., bronchospasm risk during intubation) to make the reasoning more explicit.

Recommended disposition

These questions should be fixed by adding a brief clinical stem (2–3 sentences: patient demographics, clinical setting, relevant history or findings) that requires the candidate to apply the fact rather than recall it. The correct answer and distractors typically do not need to change — the upgrade is in the stem. Priority should be given to concepts that appear in recent PYQs (MH, NMB dosing, airway management) since these are most likely to appear in future exams.

Prioritization

The seven issue categories are not equally urgent. The following prioritization reflects both the risk level of each category and the operational effort required to address it.

Immediate action required (before any further use in scored tests)

Category 2 (Factual Errors and Unsafe Answer Keys) and Category 3 (Image-Dependent Questions With Missing Images) require immediate action. Questions with wrong correct answers (da0f6ddd, 0c76505f, c9fca928, eaf22865) are actively harmful if used in scored assessments. Image-broken questions (3832e3a3, 274cb04e, 4c7e1aa2, ba6f79e7, 6a601ef7, b1d845dc) are non-functional. Both categories should be disabled from all live test templates immediately, with fixes or replacements prioritized in the next content sprint.

High priority (address in current sprint)

Category 4 (Topic Fragmentation and Near-Duplication) and Category 6 (Out-of-Scope and Low-Yield Trivia) are high priority because they affect test validity and pool efficiency. Disabling near-duplicates and out-of-scope items is low-effort (no rewriting required) and immediately improves the pool's discriminative quality. The ketamine cluster, thiopentone cluster, and MH cluster should be audited across the full 3,585-question pool, not just the 200-question sample, since the fragmentation observed in the sample almost certainly reflects a larger pattern.

Medium priority (address in next sprint)

Category 5 (Clinically Ambiguous or Structurally Flawed Construction) requires careful expert review for each question before a fix decision can be made. Some of these questions (9227d65c, d0a69dc6) require a subject-matter expert to determine which answer is correct given a specific reference standard. This work should be batched and sent for expert review rather than handled by the content operations team alone.

Ongoing / systematic

Category 1 (Recall-Only Questions) and Category 7 (Correct Concept, Insufficient Clinical Depth) represent the largest volume of work and the most significant long-term quality gap. These cannot be addressed in a single sprint. The recommended approach is to establish a conversion workflow: for each high-yield concept identified in Category 7, produce one clinical vignette version and disable the bare-recall version. Category 1 items that do not correspond to high-yield concepts should be disabled without replacement.

Category 7 items should be treated as the primary upgrade target for the subject — they represent the fastest path to improving the Bloom's distribution from its current 77% Bloom's 1–2 toward the benchmark standard of predominantly Bloom's 3–4.

Example Keep / Fix / Disable Calls

The following calls are drawn from across the eight shards and are intended to illustrate the application of the issue categories above to specific items.

KEEP — No changes needed

Question ID	Reason
7cf9f041	Clinical vignette: histamine release from d-tubocurarine causing bronchospasm and hypotension, resolved by antihistamine. Bloom's 3, plausible distractors, applied pharmacology. Matches benchmark standard.
c86c7a8b	Pediatric oculocardiac reflex during strabismus surgery. Appropriate clinical context, correct management (inform surgeon + atropine), good distractors. Bloom's 3.
12b4128b	Lorazepam in liver dysfunction: tests phase II (conjugation-only) metabolism vs. phase I-dependent drugs. Clinically relevant, good distractors, Bloom's 2 applied correctly.
eda71a36	Preoperative cardiac deferral: Mobitz type II requiring pacemaker evaluation before elective surgery. Distractors represent common but lower-risk conditions. Bloom's 3.
352c92f3	Elderly patient, fracture NOF, halothane GA, sudden hemodynamic collapse + capnogram fall → pulmonary thromboembolism. Bloom's 4, integrates monitoring interpretation with clinical diagnosis.
c5b62a34	Post-CABG scenario with quantitative hemodynamic data (cardiac index, wedge pressure, HR, BP). Requires application of cardiac output physiology to select pacing. Bloom's 3.
7fcf2d48	Obstetric anesthesia vignette: hyperventilating parturient, physiological consequence on uteroplacental circulation. Bloom's 3, clinically relevant.
7671bd8c	Pulse oximeter readings and hemoglobin variants (MetHb, CarboxyHb, HbS, SulfHb). Bloom's 4, requires understanding of oximetry physics.
6b92d6c0	Tracheotomy patient with post-procedure apnea; hypoxic drive as the mechanism. Bloom's 4, well-constructed distractors.
b238fc6a	Drug-eluting stent patient on clopidogrel pre-operatively. Bloom's 3, realistic perioperative pharmacology scenario.
4390184c	Tension pneumothorax after failed subclavian catheterisation. Bloom's 3, plausible distractors, FMGE-2025 PYQ tag.
2e559302	Post-extubation chest wall rigidity with conscious patient after fentanyl-pancuronium. Requires distinguishing incomplete reversal from respiratory depression. Bloom's 3.
6366665c	Ketamine for hemorrhagic shock emergency laparotomy. Clear clinical vignette, hemodynamic context, Bloom's 3.
485952f6	Ruptured globe with recent meal, choice of anesthetic technique. Good clinical reasoning demand (full stomach + open globe tension). Bloom's 3.
23937bb2	Fetal ion trapping with epidural bupivacaine in acidosis. Bloom's 4, obstetric anesthesia, tests pH-dependent ionization.
0df961cc	ASA classification with obese + angina at rest + COPD. Bloom's 3, NEET-PG tagged, angina at rest is the key discriminator.
f147772c	ETCO₂ as early indicator of air embolism. Bloom's 3, monitoring application, clinically important.
ad845855	Atracurium in renal failure: correct answer, good distractor set. Bloom's 2 with real clinical utility.
54e685de	RSI, Sellick's maneuver "always required" as false stem. PYQ-tagged, tests nuanced understanding that cricoid pressure is now controversial.
9254dba1	Best method to confirm double-lumen tube placement (fibreoptic bronchoscopy). Clinically relevant, medium difficulty, tempting distractors.

FIX — Targeted edits required

Question ID	Required fix
9227d65c	Add a guideline qualifier or rewrite to test a less contested obstetric spinal concept (e.g., left lateral tilt post-spinal).
0c76505f	Correct the answer key: simultaneous peak + plateau pressure rise = decreased compliance, not tube obstruction.
a4b1507e	Rewrite stem to explicitly ask about epinephrine additive concentration, or correct the answer to reflect lidocaine percentage (1–2%).
2008b9c2	Deduplicate options B and C; reassign correct answer to rapid redistribution; rewrite stem to ask pharmacokinetic basis.
6639d6cc	Rewrite the correct option to specify the concentration/uptake mechanism of the second gas effect rather than "facilitation of inhalation."
a406ea09	Correct all option units from mm Hg to cm H₂O; consider revising the PEEP range to reflect current lung-protective guidelines.
b4e374cb	Replace "None of the above" distractor with a specific mechanistic distractor (e.g., "blocking K⁺ channels to hyperpolarise the membrane").
549ef1f8	Specify the clinical context (dental infiltration vs. field block) to make one adrenaline concentration unambiguously correct.
17395fee	Add clinical context: patient under volatile anesthesia develops rising ETCO₂ and muscle rigidity — what is the most reliable early sign?
da10b136	Reclassify from Anesthetic Equipment to General Anesthesia/Pharmacology; correct Bloom's tag from 1 to 2.
e862107e	Rewrite as a complete sentence; add brief clinical context (e.g., obstetric analgesia setting) to raise Bloom's level.
274cb04e	Attach the correct image (non-rebreathing mask) or rewrite as a text-based question naming the device explicitly.
4c7e1aa2	Embed the ventilator waveform image or rewrite as a text-based scenario describing SIMV waveform characteristics.
ba6f79e7	Specify the arrhythmia in the stem text (e.g., "a patient in ventricular fibrillation") to make the question functional without the ECG image.
6a601ef7	Attach the correct image (APL/pop-off valve) or rewrite as a text-based question identifying the device by name and asking about its opening pressure.
243b9972	Change the correct answer to methohexital or reframe around a drug with a clearer contraindication profile; add a clinical qualifier.
eaf22865	Change correct answer to systemic local anesthetic toxicity (LAST) or tourniquet-related pain as the recognized primary complication of Bier block.
c969389c	Add a specific intraoperative decision point (e.g., "for postoperative pain management") that makes regional anesthesia the clearly superior choice, or restructure distractors to include clearly incorrect options.
ad52729e	Add a qualifier ("in conventional adult ARDS management") to the stem to avoid ambiguity about high-frequency ventilation.
0fb5014f	Add brief clinical context specifying the intraoperative trigger (bronchospasm risk during intubation) to make the reasoning explicit.

DISABLE — Remove from pool

Question ID	Reason
da0f6ddd	Factual error: bupivacaine toxicity treatment listed as 5% dextrose; correct treatment is 20% lipid emulsion. Patient-safety-relevant error.
d935b2a1	Factual error: "Moon" credited for demonstrating ether; correct answer is William Morton. Also history trivia with no clinical value.
c9fca928	Factual error: ammonium chloride marked as NOT useful for metabolic alkalosis; it is a recognized treatment.
92a8378d	Exact duplicate of 8d121126. One must be removed; the surviving version needs clinical reframing before use.
9f0b95bf	History-of-medicine trivia (first dentist to use ether). No clinical relevance to PG entrance preparation.
621d2586	Dental/TMJ content (myofascial pain dysfunction) misclassified as anesthesiology. Out of scope.
c5d01b8c	Dental regional anesthesia anatomy (incisive nerve block). Not relevant to INI-CET/NEET-PG anesthesiology curriculum.
57a8df2d	Pure Bloom's 1 recall (thiopentone percentage), flagged easy. No clinical reasoning required. Concept covered by better items.
e66115ea	"Which is a dissociative anesthetic?" Bloom's 1, flagged easy. Near-duplicate of 7417c224 and 1a77e9c9.
7417c224	"Which IV anesthetic has LA properties?" Bloom's 1, flagged easy. Concept better tested in a clinical vignette.
ab7b00ce	"Longest-acting local anesthetic?" Bloom's 1, flagged easy. Concept better tested via dibucaine number in pseudocholinesterase deficiency.
44fe3a7d	Near-duplicate of 748c298e; vague superlative ("best induction agent") with no clinical context.
642f85fa	Post-operative shivering → pethidine. Bloom's 1, flagged easy, no clinical context.
ce2d301e	Maximum histamine release → d-tubocurarine. Bloom's 1, flagged easy. Concept covered by the better clinical vignette 7cf9f041.
3832e3a3	Image-dependent question with no image present. Non-functional in text delivery.
b1d845dc	Correct option is a verbatim restatement of the question stem. Non-functional as a question; requires complete rewrite.
90ebdc07	GCS components except bladder function. Bloom's 1, flagged easy. Zero discriminative value for PG-level assessment.
52b938d0	O₂ concentration in mouth-to-mouth (16%). Bloom's 1, trivia-level recall, no clinical application.
2c873225	Most common side effect of spinal anaesthesia (hypotension). Bloom's 1, flagged easy, answered by rote.
4e22e222	Type E circuit for paediatric anaesthesia. Bloom's 1, flagged easy, single-fact recall.
29c292e0	CVP does not measure cardiac output. Trivially obvious, Bloom's 1, flagged easy.
ee54e685	Venturi mask max O₂ = 60%. Bloom's 1, flagged easy. Redundant with PYQ ecb677e0 which covers the same concept better.
2305f0ac	NOT in Ringer's lactate. Pure composition recall, Bloom's 1, flagged easy. No anesthetic reasoning.
dab129b2	Pin index code for N₂O. Rote equipment-code memorization, Bloom's 1, flagged easy, no tags.
4bce0367	Minimum O₂ concentration in anaesthesia. Bloom's 1, flagged easy, no tags.
566dcc93	EEG delta waves in surgical tolerance stage. Bloom's 1, flagged easy, isolated trivia.
8094cfa2	Gas stored in liquid form = N₂O. Bloom's 1, flagged easy. Answer debatable without qualifier (CO₂ also stored as liquid).
c0a3d0a0	Xenon as analgesic inhalational agent. Not in Indian clinical practice or standard PG pharmacopeia. Low yield.
ec76ed53	Bronchoscope sterilization. Belongs to infection control/microbiology, not anesthesiology. No anesthetic reasoning required.
13f8f83d	EEG beta waves = awake state. Bloom's 1, no anesthesia-specific application. Only relevant if reframed around BIS monitoring.
bb57f18b	Most common cause of death after anaesthesia (cardiac arrest). Conflates mechanism with cause; factually ambiguous and educationally inert.