Forensic Medicine Question Quality Review
Executive Summary
This review covers 200 validated non-gold candidate questions sampled from the Forensic Medicine pool of 5,504 items, analyzed across eight shards of 25 questions each. The findings are consistent and mutually reinforcing across shards: the candidate set has a small core of genuinely usable questions surrounded by a large volume of items that are either factually unsafe, structurally broken, legally obsolete, or so cognitively thin that they add no discriminative value to a PG-level assessment.
The most important headline finding is the Blooms distribution collapse: 93 of 200 candidate questions sit at Blooms Level 1 and a further 84 at Level 2, leaving only 33 questions across Levels 3–5. The benchmark and recent PYQ sets demonstrate clearly that INICET and NEET-PG expect clinical vignette reasoning, scenario-based application, and multi-step inference — not vocabulary recall. The candidate pool is structurally misaligned with that standard.
Beyond the Blooms problem, five operationally distinct issue categories emerge from the evidence: a large cluster of questions that are factually incorrect or carry contestable answer keys; a systemic legal currency failure where IPC section numbers have not been updated to the Bharatiya Nyaya Sanhita (BNS); a recurring pattern of broken image-dependent questions that are undeliverable; a topic misclassification problem that corrupts analytics and misleads candidates; and a duplicate/near-duplicate cluster that inflates apparent coverage without adding discriminative breadth.
Of the 200 questions reviewed, a conservative estimate based on shard findings suggests approximately 40–50 questions are suitable for retention as-is, 60–80 require targeted fixes of varying complexity, and 70–90 should be disabled — either because they are factually unsafe, structurally broken, legally obsolete, or so trivially easy that strong gold-standard coverage already exists for the same concept at a higher quality level.
What Good Looks Like
The benchmark and recent PYQ sets establish a clear quality bar that the candidate pool largely fails to meet. The defining characteristics of high-quality Forensic Medicine questions in this subject are:
Clinical or investigative scenario framing, even for recall-level facts. The benchmark question on aluminium phosphide poisoning (f2f248ba) does not ask "what is the antidote for AlP?" — it places the candidate in a clinical act (gastric lavage) and asks them to interpret a chemical reaction. The drowning PYQ (5aee0274) presents a body recovered from the sea and asks the candidate to reason through the physiological consequences of seawater aspiration. Even the simplest benchmark items (e.g., 0ece3701 on grievous hurt under BNS) are anchored to a legal framework that requires the candidate to apply knowledge, not just retrieve it.
Distractors that are plausible to a well-prepared candidate. In 9adfee26 (female skull identification), all five listed features are anatomically real; the candidate must distinguish which combination is specifically female. In 5da26102 (weapon-to-injury matching), all four weapons are forensically relevant and all four injury types are real — the question tests precise association, not elimination of absurd options.
Legislative currency. The benchmark explicitly tests the Bharatiya Nyaya Sanhita (0ece3701) and the MTP Amendment Act 2021 (92f67d8f). Good questions in this subject reflect the current legal framework, not the superseded IPC.
Appropriate use of EXCEPT and multi-select formats for higher-order testing. PYQs like 7f222d6e (ante-mortem drowning EXCEPT) and e234d6ce (postmortem findings NOT seen in drowning) use the EXCEPT format to force candidates to hold multiple true facts simultaneously and identify the outlier — a genuine reasoning demand. The candidate pool's EXCEPT questions frequently have one implausible distractor that makes the correct answer obvious by elimination.
Scenario complexity that matches the Blooms tag. The autopsy order question (1e25ac6a, Blooms 4) requires the candidate to reason about why the head is examined first in a fetal death case — the answer is not retrievable from a single memorised fact but requires understanding of the docimasia pulmonum sequence. The candidate pool's Blooms 3–4 tags are frequently applied to questions that require only Blooms 1 recall.
Main Issue Categories
1. Cognitively Thin Recall Items That Fall Below PG Assessment Standards
Why this pattern is bad
The defining quality problem in this subject is not a handful of bad questions — it is a structural overrepresentation of bare-fact recall items that test vocabulary and number memorisation rather than clinical reasoning. At PG entrance level, a question that any candidate who has opened a textbook once can answer correctly in under five seconds provides no discriminative information. It cannot distinguish a prepared candidate from an underprepared one, and it wastes test real estate that should be occupied by applied reasoning items.
The Blooms distribution makes this concrete: 93 of 200 candidate questions are at Level 1 and 84 at Level 2. Only 9 reach Level 3 and 22 reach Levels 4–5. The benchmark and PYQ sets, by contrast, are dominated by Levels 3–5 with scenario framing even on their Level 1 items. The candidate pool is not just skewed — it is structurally misaligned with the exam it is meant to serve.
How it shows up
The pattern appears as a broad, pervasive problem across all eight shards. It is not a cluster in one topic area — it affects Forensic Toxicology, Asphyxial Deaths, Identification, Medical Jurisprudence, and Forensic Pathology equally. The typical presentation is a one-sentence stem asking for a single fact, with three implausible distractors and one obvious correct answer:
- Definition questions: "Thanatology deals with:" (
931432f6,e2e9e613), "Testamentary capacity is the ability to:" (22cbacdf), "Ophitoxaemia refers to poisoning due to:" (d33d1753), "What is taxidermy?" (37fc4bfb) - Number/year recall: "Legal age of consent is:" (
511f39c0,86e24429), "PNDT Act was passed in which year?" (1b79e30f), "For how many years must a person be missing to be declared dead?" (e4cc3c42), "In which year was the Mental Health Act passed?" (d9175a57) - Single-antidote or single-feature recall: "What is the antidote for sodium nitrite poisoning?" (
af43bacb), "What is the fatal dose of arsenic?" (9468016f), "Which is the last organ to putrefy?" (1b446650), "Rule of Nines is related to which of the following?" (235c691d) - Synonym/slang recall: "What is white vitriol?" (
ebc51779), "Phencyclidine street names — which is NOT?" (917c7daa), "Locard's principle of exchange of trace evidence" (cb35686e)
Recommended disposition
Items in this cluster where strong gold-standard or PYQ coverage already exists for the same concept: disable. Items where the underlying concept is genuinely high-yield but the question is merely poorly constructed: fix by converting to a clinical vignette or scenario-based stem. The fix path is only warranted when the concept itself is not already covered by a better item in the benchmark or PYQ set.
Specific disable calls from this category: 931432f6, e2e9e613, 22cbacdf, d33d1753, 37fc4bfb, 511f39c0, 86e24429, e4cc3c42, d9175a57, af43bacb, 235c691d, cb35686e, 917c7daa, 88c3a05c, 1340bb73, d2ba527c, 00c034e0, 7d062f04.
2. Factually Incorrect or Contestable Answer Keys
Why this pattern is bad
A question with a wrong answer key is worse than no question at all. It actively teaches an incorrect association, and in a subject like Forensic Medicine — where candidates are preparing for both examinations and eventual medicolegal practice — a wrong answer about brain death criteria, antidotes, or postmortem findings has real downstream consequences. This category is operationally distinct from the recall-quality problem above: these questions may be well-constructed in format but are unsafe to deploy because the keyed answer is factually wrong or seriously contested.
How it shows up
This pattern appears as a narrower but high-severity cluster, concentrated in Forensic Toxicology and Forensic Pathology. The errors are of several subtypes:
Outright factual errors in the answer key:
0ebdb4c9: Antidote for ethylene glycol poisoning marked as "Fluconazole." Fluconazole is an antifungal with no role in ethylene glycol poisoning. The correct answer (fomepizole) is present as a distractor. The answer key is simply inverted.f4e1000f: "Most specific test for OP poisoning" marked as plasma cholinesterase. Standard teaching designates RBC (erythrocyte) cholinesterase as more specific; plasma cholinesterase is more sensitive but reflects liver synthesis, not tissue inhibition.0fdbebd6: Brain death criteria listed as "ECG, EEG, decreased body temperature, and pupillary dilatation." ECG is not a brain death criterion; decreased body temperature is a consequence, not a criterion; the accepted criteria involve apnoea testing, absent brainstem reflexes, and isoelectric EEG. The marked answer is factually incorrect.ec1cefc1: Floppy baby syndrome attributed to lithium. The canonical association is with benzodiazepines (or botulinum toxin in a different context); lithium causes neonatal toxicity but is not the standard teaching for floppy baby syndrome.61190799: Caucasian skull described as "round." Round/brachycephalic morphology is characteristic of Mongoloid skulls; Caucasian skulls are typically dolichocephalic. Potential answer error requiring expert correction.
Contestable or source-dependent answers that create exam disputes:
3343899b: "NOT typically seen in cyanide poisoning" marks "characteristic bitter almond smell" as the answer. Bitter almond smell is a classic positive finding in cyanide poisoning, not an absent one. The question logic appears inverted.f4d83c83: "Most specific feature of hanging" marked as ligature marks. Ligature marks are common to all forms of ligature asphyxia; they are not specific to hanging. A more defensible answer would reference the direction/angle of the ligature mark, Simon's sign, or fracture of C2 in judicial hanging.ad52542e: "TRUE about cadaveric lividity" marks "occurs only in dependent parts" as correct. The word "only" makes this technically false — contact flattening produces pallor in non-dependent areas, and lividity distribution is modified by position changes.3c217a93: Salt water vs. fresh water drowning — "faster death" as the correct answer is contested; fresh water causes faster cardiac death via haemolysis and ventricular fibrillation, while salt water causes slower death via pulmonary oedema. The answer is reference-dependent and the stem provides no qualifier.9c919c81: Exhumation requires "a written order from the coroner." In India, exhumation is ordered by an Executive Magistrate; the coroner system is largely defunct except in Mumbai. The answer reflects a non-Indian or outdated framework.49c1ccc8: Abrus precatorius EXCEPT question has two potentially wrong options — the active principle is stated as "N-methyl tryptophan" (it is abrin) and "symptoms resemble cobra snake bite" is the marked false option, but the active principle option is also factually incorrect.1b79e30f: PNDT Act year — the original Act was passed in 1994; the PC-PNDT amendment was in 2003. Marking 2003 as correct without specifying which version is misleading.
Recommended disposition
Where the error is clear and the correct answer is unambiguous: fix (correct the answer key and revise the distractor set). Where the answer is genuinely contested across standard Indian references (Reddy, Pillay, Nandy) and the stem cannot be made unambiguous without a major rewrite: disable. Items with two simultaneously wrong options (49c1ccc8) should be disabled pending full expert reconstruction.
Priority fixes: 0ebdb4c9, f4e1000f, 0fdbebd6. Priority disables: ec1cefc1, 3343899b, ad52542e, 49c1ccc8.
3. Legal Framework Obsolescence — IPC Questions Not Updated to BNS
Why this pattern is bad
The Bharatiya Nyaya Sanhita (BNS) replaced the Indian Penal Code (IPC) with effect from July 2024. The benchmark set explicitly tests BNS knowledge (0ece3701 on grievous hurt under BNS; 92f67d8f on MTP Amendment Act 2021). Questions that cite IPC section numbers without any BNS cross-reference are not merely outdated — they actively teach candidates the wrong legal framework for current practice and current examinations. This is a systemic problem, not an isolated one: it appears in every shard reviewed and affects Medical Jurisprudence, Sexual Offences, Injuries, and Asphyxial Deaths topics.
How it shows up
The pattern is broad and consistent across the sampled set. Questions cite IPC sections as if the code is still operative:
57336f7a: IPC 304 for culpable homicideb1b3d13c: IPC 376-C for sexual offences6aa97f00: IPC 315 for infanticide9e35e91f: IPC Section 376 for rape punishment7582bac5: IPC section for false evidence58fc2c21: IPC Section 86 for voluntary intoxicationa1351f6c: IPC 304A for criminal negligencebffe6fc2: IPC 320 for grievous hurtf0e2cbac: Section 304A IPC for medical negligence1cad4788: IPC Sections 312–316 for causing miscarriage57fe12f4: Section 498A IPC for dowry cruelty30253c4f: IPC Sections 312 and 31576ad7773: IPC 304 for culpable homicidedbdb2905and1f9f9669: IPC sections 320 and 325
None of these questions acknowledge the BNS equivalents (e.g., BNS Section 106 for criminal negligence, BNS Section 114 for grievous hurt, BNS Sections 64–70 for sexual offences, BNS Sections 88–92 for causing miscarriage).
Recommended disposition
This is a fix category, not a disable category — the underlying concepts are high-yield and examinable. The remediation path is clear: update section numbers to BNS equivalents, or where both IPC and BNS are pedagogically relevant (e.g., for candidates who may encounter older case law), explicitly frame the question as "under the Bharatiya Nyaya Sanhita" with the new section number. Questions that test only the section number as a bare fact (e.g., 7582bac5, d2ba527c) should be disabled rather than fixed, since section-number memorisation is low-yield even under BNS. Questions that test the legal concept with the section number as context (e.g., a1351f6c on criminal negligence, bffe6fc2 on grievous hurt) should be updated and retained.
4. Broken Image-Dependent Questions
Why this pattern is bad
Several questions in the candidate set are structurally undeliverable because they reference images, X-rays, or figures that are absent from the question text. These are not questions with minor formatting issues — they are completely unanswerable without the missing visual. A candidate presented with "What type of injury is shown in the image?" with no image cannot engage with the question at all. Serving these items in any assessment context would generate candidate complaints, score disputes, and loss of platform credibility.
How it shows up
This pattern appears as a narrower cluster but is confirmed across multiple shards, suggesting it is not isolated to a single authoring batch:
37fda070: "What is the estimated age from the given X-ray?" — no X-ray present. The question is entirely dependent on the radiograph showing epiphyseal fusion landmarks.94d02d50: "What type of injury is shown in the image?" — no image present. The correct answer (brush burn) cannot be inferred from the stem alone.e30584cb: "What does the following image show?" with fracture type options — no image embedded.bb2281f1: Stem reads "The following injury is:" with no image or clinical description — the stem is entirely empty of content.063b1d14: "All are true about the image shown in the figure except:" — no figure present.9de839cd: "What is incorrect about the image shown below?" — no image present.
A related but distinct sub-problem is the truncated stem: 267437b0 ("Which of the following statements regarding Eunuchs is true?") presents options referencing "both statements" but no statements are visible in the question text — the stem is truncated at the data level. 92879955 has the same defect. These are data entry failures rather than image-linkage failures, but the operational consequence is identical: the question is unanswerable.
Recommended disposition
Two remediation paths exist. If the correct image or missing stem content can be retrieved and attached: fix (attach image or restore stem). If the image cannot be recovered or the stem cannot be reconstructed: disable. Given that these questions have been in the pool without images, the probability that the assets are readily available is uncertain. The content team should attempt image recovery first; any item where the image cannot be confirmed within one review cycle should be disabled immediately to prevent accidental deployment.
Specific items: 37fda070, 94d02d50, e30584cb, bb2281f1, 063b1d14, 9de839cd, 267437b0, 92879955.
5. Topic Misclassification Contaminating Forensic Medicine Categories
Why this pattern is bad
Topic misclassification has two distinct harms. First, it corrupts topic-level analytics: if questions about skull morphology are filed under Asphyxial Deaths, performance data on that topic becomes meaningless. Second, it misleads candidates who are studying by topic — a candidate drilling Asphyxial Deaths should not encounter a question about Caucasian skull shape. A third harm is specific to Forensic Medicine: some misclassified items are not from a different FMT sub-topic but from an entirely different subject (oral surgery, clinical medicine, general law), which means they are testing knowledge that belongs to a different discipline entirely.
How it shows up
Two subtypes are observed in the reviewed set:
Within-FMT misclassification (wrong sub-topic):
58fc2c21(IPC Section 86, voluntary intoxication and criminal responsibility) filed under Asphyxial Deaths — belongs to Medical Jurisprudence4de5cdb2(suspended animation in drowning) filed under Identification — belongs to Asphyxial Deaths0940f202(Burking) filed under Injuries and Their Significance — belongs to Asphyxial Deaths61190799(Caucasian skull shape) filed under Asphyxial Deaths — belongs to Identification
Cross-subject contamination (non-FMT content):
c0ba8b97(painful lockjaw EXCEPT tetany) filed under Injuries and Their Significance — this is an oral surgery/general medicine question testing trismus aetiology with no medicolegal relevanceae31d478(water test diagnoses pneumothorax) — in forensic context the "water test" (hydrostatic/docimasia pulmonum) is used to determine live birth, not pneumothorax; this appears to be a clinical medicine question misclassified into FMTa357aa82(patient refuses blood transfusion) tagged USMLE Step 1/2 only with no Indian exam tag — generic Western bioethics content with no Indian medicolegal framing
Recommended disposition
Within-FMT misclassifications: fix by correcting the topic tag. The question content may be usable once correctly classified. Cross-subject contamination: disable from the Forensic Medicine pool. Items like c0ba8b97 and ae31d478 may have a home in another subject (Oral Surgery, General Medicine) but should not appear in FMT analytics or FMT-tagged assessments.
6. Duplicate and Near-Duplicate Question Pairs
Why this pattern is bad
Duplicate questions waste pool capacity and, when both versions are deployed in the same assessment or study plan, create an artificial sense of coverage while actually testing the same narrow fact twice. In a pool of 5,504 questions, duplication is particularly costly because it inflates apparent breadth while the actual conceptual coverage is narrower than it appears. Near-duplicates are especially problematic when one version is factually stronger than the other — the weaker version may be served preferentially by randomisation, teaching a less precise version of the concept.
How it shows up
The pattern appears as a narrower cluster but is confirmed across multiple shards, suggesting it is likely more prevalent across the full 5,504-question pool than the 200-question sample can reveal:
c5d2d0bdand66ea17be: Both ask about soot in the respiratory passage as the indicator of antemortem burns in a charred body. Nearly identical stems and correct answers. One should be disabled.a084a89dandc777efff: Both ask why oral evidence is more important than written testimony, with the same correct answer and nearly identical distractors. One is PYQ-tagged; the other is not. The non-PYQ version (c777efff) should be disabled.0078525cand082561b6: Both ask which form of phosphorus causes "phossy jaw" with the same correct answer (yellow/white phosphorus) and overlapping distractors. One should be disabled.96b48603ande735fd1f: Both present a patient with cholinergic features pointing to organophosphorus poisoning. The richer version (96b48603, with full cholinergic picture including convulsions, bradycardia, miosis, and secretions) should be retained; the stripped-down version (e735fd1f) should be disabled.1da07ecdand3760c285: Two opioid symptom questions in the same shard, both Blooms-1/2, no scenario. The EXCEPT-format version (3760c285) is marginally stronger;1da07ecdshould be disabled.
A related problem is the aluminium phosphide / silver nitrate question, which appears in both the benchmark set (f2f248ba) and as a near-duplicate in the recent PYQ set (fd4b2d37). Any candidate-pool version of this question is redundant given gold-standard coverage already exists.
Recommended disposition
In each confirmed duplicate pair: keep the stronger version (typically the one with richer clinical context, PYQ tagging, or more discriminating distractors) and disable the weaker version. Do not attempt to fix the weaker version into a distinct question unless the concept genuinely warrants two different angles of testing.
7. Distractor Construction Failures That Undermine Question Validity
Why this pattern is bad
A question with a correct answer key but broken distractors is not a valid assessment item. If one distractor is anatomically absurd, if the correct answer is also a distractor in a different option, or if "all of the above" is the keyed answer, the question fails to discriminate between candidates who know the material and candidates who are simply eliminating implausible options. This is operationally distinct from the factual accuracy problem (Category 2) because the answer key may be correct — the failure is in the option set, not the key.
How it shows up
This pattern appears as a moderate-frequency cluster, most visible in Forensic Toxicology and Forensic Pathology:
8976bd5e(Harvard criteria for brain death): The distractor "Small obturator foramen" is anatomically absurd and has no plausible link to brain death criteria. Any candidate will eliminate it immediately, reducing the question to a three-option item.27741f4b(Bevelling of skull in firearm injury): The first option is a verbatim copy of the question stem ("Bevelling of the skull is seen in the:"). This is a data entry error that makes the option set incoherent.ea00ad70(Arrest type in drowning): "All of the above" is the keyed answer. This is a known test-construction flaw — it rewards candidates who recognise the format rather than those who know the content.c115ffd3(Anal coitus terminology): "Sin of Gomorrah" is itself a near-synonym for the correct answer in some texts, making it a flawed distractor that could be argued as also correct.5ca3aab6(Nitrate hypostasis): The distractors (CO, cyanide) are too obviously wrong — both are well-known causes of cherry-red hypostasis, not nitrate-related changes. A well-prepared candidate eliminates them instantly.3f7fc207(Formication): The difficulty field reads "easy" as a string rather than an integer — a metadata error that, while not a content problem, signals authoring quality issues in this item.
Recommended disposition
Items where the distractor failure is a simple authoring error (verbatim stem copy, absurd option): fix by replacing the broken distractor with a plausible alternative. Items where "all of the above" is the keyed answer: fix by restructuring as a positive-stem question or converting to an EXCEPT format. Items where the distractor set is too obviously wrong to discriminate: fix by replacing weak distractors with plausible near-miss options drawn from the same conceptual neighbourhood.
Specific fix calls: 8976bd5e, 27741f4b, ea00ad70, c115ffd3.
Prioritization
The table below ranks the seven issue categories by the combination of severity (how badly the problem harms assessment quality) and breadth (how many questions in the sample are affected).
| Priority | Issue Category | Severity | Breadth in Sample | Primary Action |
|---|---|---|---|---|
| 1 | Factually Incorrect or Contestable Answer Keys | Critical | Moderate (~15–20 items confirmed) | Fix or disable immediately; unsafe to serve |
| 2 | Broken Image-Dependent Questions | Critical | Narrow (~8–10 items confirmed) | Fix (attach image) or disable; undeliverable |
| 3 | Legal Framework Obsolescence (IPC → BNS) | High | Broad (~15–20 items confirmed) | Systematic fix pass; outdated for current exams |
| 4 | Cognitively Thin Recall Items Below PG Threshold | High | Very broad (~80–100 items estimated) | Bulk disable where gold coverage exists; vignette upgrade otherwise |
| 5 | Topic Misclassification | Moderate | Narrow (~8–12 items confirmed) | Fix tags; disable cross-subject contamination |
| 6 | Duplicate and Near-Duplicate Pairs | Moderate | Narrow (~10–15 pairs estimated) | Disable weaker version in each pair |
| 7 | Distractor Construction Failures | Moderate | Moderate (~10–15 items confirmed) | Fix distractor sets; straightforward remediation |
Immediate action items (before next deployment cycle): All items in Categories 1 and 2 should be pulled from active serving. The factual errors in 0ebdb4c9, f4e1000f, and 0fdbebd6 are particularly dangerous because they involve clinical management decisions (antidotes, brain death criteria) where a wrong answer could reinforce harmful practice.
Short-term action items (within one review cycle): The IPC-to-BNS update pass (Category 3) should be treated as a subject-wide sweep, not a question-by-question fix, because the pattern is systemic. All questions citing IPC sections should be identified across the full 5,504-question pool and routed through a single legislative update review.
Medium-term action items: The Blooms distribution problem (Category 4) requires new question creation, not just editing of existing items. The candidate pool cannot be fixed into a higher Blooms distribution by editing alone — the subject needs targeted commissioning of Blooms 3–5 clinical vignette questions in the high-frequency topic areas (Asphyxial Deaths, Forensic Toxicology, Identification, Medical Jurisprudence).
Example Keep / Fix / Disable Calls
The following examples are drawn directly from the reviewed sample and illustrate the application of the issue categories above to specific items.
KEEP — 443b7ae5 (Disaster Victim Identification — STR profiles, building collapse scenario)
Well-constructed Blooms-4 applied question presenting a realistic DVI scenario with conflicting fingerprint and DNA evidence. Requires the candidate to reason through post-mortem body part displacement. Unique topic area, good discriminative potential, matches the clinical-reasoning style of benchmark PYQs. No structural issues.
KEEP — bc9b1ee7 (Phenol poisoning — brownish oral mucosa and greenish-brown urine)
Clean vignette format, factually sound, appropriate Blooms-2 application. Matches PYQ style. The clinical presentation anchors the recall fact in a diagnostic reasoning context.
KEEP — 13107463 (Arsenophagia clinical vignette — elderly patient on traditional medicine with cardiac arrhythmias and skin changes)
Blooms-3 applied reasoning, well-constructed distractors, clinically grounded. Matches benchmark vignette style. Keep with minor difficulty calibration check.
KEEP — e5cc6ca1 (Snake bite vignette — haemotoxic features pointing to viper)
Genuine Blooms-3 application; distractors are plausible. Matches the clinical vignette style of recent PYQs. Keep with minor distractor review.
KEEP — b36694c1 (Drunk surgeon nicking artery — criminal vs. civil negligence)
Well-constructed vignette, Blooms-5, clinically grounded, high-yield for Medical Jurisprudence. Needs minor review to confirm PYQ tag accuracy.
KEEP — 0c620bbc (Antemortem burns — soot in airways, NEET-PG tagged)
Tagged NEET-PG, Blooms-4, difficulty-3. Distractors include plausible postmortem features. Discriminatory and well-constructed.
FIX — 0ebdb4c9 (Antidote for ethylene glycol — "Fluconazole" marked correct)
Category: Factually Incorrect Answer Key. The correct answer (fomepizole) is present as a distractor. Swap the answer key to fomepizole and replace fluconazole with a plausible distractor (e.g., N-acetylcysteine or thiamine). Priority fix before next deployment.
FIX — 37fda070 (Age estimation from X-ray — no image present)
Category: Broken Image-Dependent Question. Attempt to recover and attach the relevant radiograph (likely showing medial clavicle epiphyseal fusion). If image cannot be recovered, rewrite as a text-based question describing the radiological finding explicitly.
FIX — a1351f6c (Criminal negligence under IPC 304A)
Category: Legal Framework Obsolescence. Factually correct under old IPC. Update to BNS Section 106 framing or add explicit "(as per IPC, now replaced by BNS Section 106)" qualifier. The underlying concept is high-yield and should be retained.
FIX — 27741f4b (Bevelling of skull — first option is verbatim copy of stem)
Category: Distractor Construction Failure. Replace the first option with a meaningful distractor (e.g., "Entry point of bullet" or "Tangential gunshot wound"). Content is otherwise sound.
FIX — 9c919c81 (Exhumation — "written order from coroner")
Category: Factually Incorrect Answer Key. In India, exhumation is ordered by an Executive Magistrate, not a coroner. Rewrite the correct answer option and verify against current Indian legal framework (CrPC/BNSS provisions).
FIX — 58fc2c21 (IPC Section 86 — filed under Asphyxial Deaths)
Category: Topic Misclassification + Legal Framework Obsolescence. Reclassify to Medical Jurisprudence. Also add BNS Section 26 equivalent or update framing.
FIX — 98396c61 (Cadaveric spasm "also known as rigor mortis")
Category: Factually Incorrect Stem. Cadaveric spasm and rigor mortis are distinct phenomena. Rewrite to ask specifically about one or the other; the current stem conflates them and will teach an incorrect association.
DISABLE — 931432f6 (Thanatology deals with — options include "Snakes" and "Fingerprints")
Category: Cognitively Thin Recall. Trivially easy, no discriminatory value, implausible distractors. The concept is covered at a higher quality level in the benchmark set.
DISABLE — ec1cefc1 (Floppy baby syndrome attributed to lithium)
Category: Factually Incorrect Answer Key. The canonical association is with benzodiazepines. Until expert verification confirms or corrects, this item risks teaching an incorrect clinical association. Disable immediately.
DISABLE — c0ba8b97 (Painful lockjaw EXCEPT tetany — oral surgery content)
Category: Topic Misclassification / Cross-Subject Contamination. This is an oral surgery question with no medicolegal relevance. Disable from the Forensic Medicine pool entirely.
DISABLE — 0078525c (Phossy jaw — duplicate of 082561b6)
Category: Duplicate Content. Near-identical to 082561b6. The clinical vignette version (082561b6, fire worker) is marginally stronger. Disable 0078525c; consider vignette upgrade for 082561b6.
DISABLE — 7d062f04 (INTERPOL DVI yellow form colour)
Category: Cognitively Thin Recall / Niche Trivia. Colour-code memorisation with no reasoning demand and no PYQ precedent. No clinical or medicolegal reasoning is tested. Disable unless a dedicated DVI question set is being built with appropriate context.
DISABLE — ad52542e (TRUE about cadaveric lividity — "occurs only in dependent parts")
Category: Factually Incorrect Answer Key. The word "only" makes the marked correct answer technically false. The question is poorly constructed and the correct answer is itself inaccurate. Disable and rewrite from scratch.
DISABLE — ae31d478 (Water test diagnoses pneumothorax — clinical medicine content in FMT)
Category: Topic Misclassification / Cross-Subject Contamination. In forensic context, the water test (docimasia pulmonum) determines live birth, not pneumothorax. This is either a misclassified clinical medicine question or a factual error in the forensic context. Disable from FMT pending clarification of intended context.