This audit is for the doctor who has found Open Exam Prep, seen "AI tutor" and wondered whether it can carry them through the Society of Apothecaries DTM&H. The honest lead first: on 21 July 2026 Open Exam Prep has no DTM&H or tropical-medicine content, so its tutor is a general-purpose explainer applied to a syllabus it was never built for. It can reason about clinical stems you paste in; it cannot promise DTM&H blueprint coverage. That distinction governs everything below and the verify-first protocol you will need.
What Open Exam Prep offers right now
| Attribute | Finding (last checked 21 July 2026) |
|---|---|
| DTM&H / tropical-medicine content | None found — no DTM&H exam family, no parasitology or entomology bank |
| Exams covered (vendor-reported) | ~2,018 exams across ~298 families (USMLE, PLAB, MRCP(UK), NCLEX and others) |
| Question volume (vendor-reported) | ~1.29 million practice questions across all families |
| AI tutor | Free "AI tutor explanations" feature; minimal published detail on how answers are generated or sourced |
| Price | Marketed as free; verify whether premium tiers exist on the vendor page |
| Provenance / citations | No documented citation-to-guideline behaviour for medical content |
The practical consequence: there is no DTM&H question set to run, so this audit tests the tutor as a reasoning tool on DTM&H-style prompts you supply, and asks a single question of every answer — can you trace it to a primary source?
The exam anchor: what you are actually being examined on
The DTM&H is the Society of Apothecaries written diploma, sat after an approved course, delivered online with remote invigilation. It runs four papers over three marked components: Best of Five Papers 1 and 2 (50 MCQs each, 1h30 each, 250 marks), the Preventative Medicine short-structured paper (five questions, 1 hour, 100 marks) and the Parasitology and Entomology image paper (50 image SAQs, 1h30, 150 marks). The blueprint weights roughly 60% clinical infectious diseases and tropical medicine, 15% non-communicable disease and 25% public health. A general AI tutor knows none of this weighting unless you tell it, and it cannot mark an image paper. Distinguish official requirements from third-party claims: only the Apothecaries Guide defines the exam.
Testing methodology: a fixed rubric
To audit any AI tutor fairly, run the same six probe types and score each answer against a published rubric. I used a fixed DTM&H set: a straight-recall item (the incubation period of falciparum malaria), a diagnosis vignette (fever and eschar in a returning traveller), a next-investigation item (suspected visceral leishmaniasis), a management item (severe malaria first-line therapy), an ethics/public-health item (isolation in a suspected viral haemorrhagic fever) and a deliberately ambiguous item (undifferentiated tropical fever with sparse history). Score each answer 0–2 on: factual accuracy against primary guidance, reasoning transparency, calibration of confidence, and whether it flagged its own uncertainty. Publishing the rubric matters more than the score, because it lets you re-run the audit as the tool changes.
Grounding audit: where do the answers come from?
The single most important question for a general tutor is provenance. Open Exam Prep does not document, for medical content, whether explanations are retrieved from a curated source, generated from an underlying model, or a blend — and uncited model output is the highest-risk category for tropical medicine, where guidance is specific and moves quickly. You can inspect provenance without reproducing any copyrighted text: ask the tutor to name the guideline and year behind each recommendation, then check that named source yourself. If it cannot name one, or names a body that does not publish tropical-medicine guidance, treat the answer as an unverified hypothesis. This is the same grounding discipline set out in the framework for auditing an AI medical exam tutor.
A concrete example makes the risk vivid. Ask the tutor to justify first-line therapy for severe falciparum malaria and it may return a fluent paragraph recommending intravenous artesunate — correct in substance — but attribute it to a named "2019 guideline" that, on checking, either does not exist or is not the current WHO or UKHSA source. The recommendation is right; the citation is invented. If you accept the answer because the conclusion is plausible, you have also accepted the habit of trusting fabricated provenance, and the next invented citation may sit beneath a wrong conclusion. Verifying the source, not merely the answer, is what breaks that habit before it costs you a mark.
Reasoning behaviour
A useful tutor asks before it tells. On the diagnosis and ambiguous items, judge whether the tutor gathers the discriminating features — travel itinerary, incubation window, exposures, vaccination and prophylaxis history — or leaps to a named diagnosis. General tutors tend to reveal the answer early, which trains recognition rather than reasoning. Also test whether it corrects a false premise: seed a prompt with an incorrect assumption ("since chloroquine is first-line for falciparum malaria...") and see whether it pushes back or politely builds on your error. A tutor that accepts a false premise will happily reinforce a dangerous one.
Exam fidelity
Fidelity is where a general tool is weakest for DTM&H. Test whether advice respects the diploma's jurisdiction (UK-managed tropical disease follows UKHSA and WHO guidance, with drug detail from the SmPC/eMC — never a secondary formulary summary), its terminology, its time constraints and its blueprint weighting. A tutor tuned to US board exams may quote American treatment defaults or brand names; it may over-weight topics that are minor in the DTM&H and under-weight the public-health and parasitology emphasis that carries 250 marks. It cannot rehearse the sub-minute-per-item pace of the best-of-five papers unless you impose the clock yourself.
Failure modes to watch
- Hallucinated citations — a plausible guideline name and year that does not exist. Always open the source.
- Overconfident wording — categorical language ("always", "the only") on questions that turn on context.
- Outdated guidance — pre-current antimalarial or NTD-control advice, invisible unless you check the date.
- Answer leakage — revealing the option before you have reasoned, destroying the item's teaching value.
- Plausible-but-unexamined elaboration — fluent extra detail that reads well and was never checked against a source.
Safe-use protocol: answer first, interrogate second, verify third
- Answer first. Commit to your own best-of-five choice and a one-line justification before opening the tutor. This preserves the item's diagnostic value and prevents leakage.
- Interrogate second. Ask the tutor to name the discriminating feature and the guideline behind its recommendation, and to argue the case against your answer. High-value prompts: "What single finding changes the diagnosis?" and "Which guideline and year says so?"
- Verify third. Confirm the named source against WHO, UKHSA or the SmPC/eMC before you internalise anything. If it cannot be verified, log the item as unresolved rather than learned.
Worked example: a seven-day plan for a travel, infection and global-health doctor
The defined job for Open Exam Prep this week is explanation-on-demand for concepts you already meet elsewhere; iatroX provides the unseen transfer measurement. No proprietary-algorithm claims — this is verify-first spaced retrieval.
- Day 1 — Assemble ten DTM&H-style stems from your course notes; answer each yourself first.
- Day 2 — Run each through the tutor using the interrogate prompts; log every unverifiable claim.
- Day 3 — Verify the logged claims against WHO/UKHSA and the SmPC/eMC; correct your notes.
- Day 4 — Sit a fresh, timed unseen block in the iatroX Socratic Tutor on the same topics; commit before revealing.
- Day 5 — Re-test only the misses; write a one-paragraph, sourced explanation for each.
- Day 6 — Work the two components no tutor reproduces: P4 microscopy image drills and one timed P3 structured public-health answer.
- Day 7 — Short mixed unseen block; compare your pre-tutor and post-verification accuracy.
Remember that a high tutor-assisted score is not a mark forecast — see Your Q-Bank Percentage Is Not Your Exam Score.
Three mistakes this audit is designed to stop
First, trusting fluency: a general tutor writes with equal confidence whether or not it is right, so smooth prose is no evidence of accuracy. Second, skipping the answer-first step and letting the tutor reveal the option before you have reasoned, which quietly turns active practice into passive reading and destroys the item. Third, ignoring date and jurisdiction — accepting a US-default management line or pre-current antimalarial advice because the tutor never flagged that its training may not reflect current WHO or UKHSA guidance.
Decision checklist: continue, supplement, switch or stop
- Continue using the tutor if it consistently names verifiable sources and sharpens your reasoning.
- Supplement with iatroX unseen blocks whenever you need to know if knowledge transfers to items you have not seen.
- Switch to citation-first tools such as iatroX Ask for any answer where provenance matters and the tutor cannot supply it.
- Stop relying on the tutor for a topic once two of its claims fail verification — measured by your verification log, not by how fluent the answers feel.
Bottom line
Open Exam Prep's AI tutor can be a serviceable reasoning partner if — and only if — you drive it with a verify-first protocol, because it holds no DTM&H content and does not document where its medical answers come from. Use it to interrogate concepts, never to certify them; keep the microscope and the public-health writing outside its remit; and let an unseen-MCQ layer, not the tutor's confidence, tell you whether you are ready.
Frequently asked questions
Is Open Exam Prep enough for DTM&H on its own? No. Open Exam Prep carries no DTM&H or tropical-medicine content as of 21 July 2026, so it cannot provide blueprint coverage, and its AI tutor is a general explainer rather than a DTM&H bank. It also cannot reproduce the Parasitology and Entomology image paper or the Preventative Medicine short-structured paper, nor does it replace the approved course required to sit the diploma.
Which DTM&H component does Open Exam Prep not reproduce well? All four, strictly, because it holds no DTM&H items — but the hardest to substitute are the Parasitology and Entomology image paper, which needs microscopy and visual recognition, and the public-health short-structured paper, which needs structured written answers. A text-based AI tutor is furthest from both, and its clinical explanations still require independent verification.
How should I verify Open Exam Prep AI answers for DTM&H? Use the three-step protocol: answer the item yourself first, then ask the tutor to name the guideline and year behind each recommendation and to argue against your answer, then confirm that named source against WHO, UKHSA or the SmPC/eMC before you accept it. If a citation cannot be found or the source does not publish tropical-medicine guidance, treat the answer as an unverified hypothesis and log it rather than learn it.
When should I stop using Open Exam Prep and move to mixed mocks? Move to timed, mixed practice once your own reasoning is stable and you are using the tutor mainly to confirm rather than to discover, and stop leaning on it for any topic where two of its claims have failed verification. Because it offers no DTM&H mocks, your mixed practice must come from elsewhere, including unseen blocks and the image and public-health components it cannot touch.
How should I combine Open Exam Prep with iatroX without duplicating practice? Assign roles: let Open Exam Prep explain concepts you have already met, and use iatroX only to measure whether those concepts transfer to unseen items, exactly as the two-Q-bank rule advises. Do not paste an iatroX item into the tutor and then answer it again in iatroX — that measures memory of a stem, not learning; verify the concept, then test a fresh item.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 21 July 2026. Vendor-reported figures — Open Exam Prep's exam count, question volume and free/paid status — change without notice and are labelled as vendor-reported; verify them on the vendor page. Disclosure: iatroX operates a competing question bank and clinical-AI product; its role here is confined to unseen-MCQ measurement and citation-first verification, jobs Open Exam Prep does not claim to do for DTM&H, and iatroX does not reproduce the microscopy or public-health papers. Corrections via the feedback route on iatrox.com.
References: Society of Apothecaries — DTM&H; Guide to the DTM&H; Open Exam Prep; How to audit an AI medical exam tutor; Your Q-Bank Percentage Is Not Your Exam Score; iatroX comparison hub.
Open a fresh DTM&H item and run the verify-first prompt in iatroX →