This is for doctors preparing for the Diploma in Tropical Medicine & Hygiene (DTM&H) who want to use an AI tutor to accelerate learning without letting it think for them. It addresses the knowledge that underpins the two best-of-five MCQ papers and, to a lesser extent, the parasitology paper. The principal limitation is one you must hold in view throughout: Open Exam Prep is a general healthcare-certification tutor that does not cover tropical medicine, so using it for DTM&H means using a general-purpose AI without DTM&H-specific curation — which raises, not lowers, your verification burden.
Current state: what Open Exam Prep actually is
Open Exam Prep (open-exam-prep.com) is a real, largely free platform, but it is not a DTM&H product. Verified 21 July 2026, it presents itself as a broad healthcare-certification site with an "Ask AI" tutor and a very large question pool, oriented to nursing, pharmacy and allied-health certifications.
| Attribute | Open Exam Prep — verified 21 July 2026 |
|---|---|
| DTM&H / tropical-medicine coverage | Not found. Tropical medicine and the DTM&H are not listed; the scope is nursing, pharmacy and allied-health exams. |
| AI tutor | Yes — an "Ask AI" feature is present. |
| Question pool (vendor-reported) | 265,364 practice questions across 2,018 exams in 298 families. |
| Price | "Free Exam Prep" is advertised; no pricing tiers are stated. |
| Adaptive learning | Not mentioned. |
Read that top row twice. Because there is no DTM&H-specific content, you cannot treat Open Exam Prep as a DTM&H bank; you can only treat its "Ask AI" as a general reasoning aid that you feed with your own DTM&H material and then rigorously check. Everything below is therefore written to be transferable to any general AI tutor, and the discipline matters more, not less, precisely because the model has no curated tropical-medicine grounding.
Exam anchor: the DTM&H syllabus and official material
The DTM&H written examination is set by the Society of Apothecaries and, per the current Guide (verified 21 July 2026), has four online invigilated components: Papers 1 and 2 (50 best-of-five MCQs each, 90 minutes each, no negative marking); Paper 3 (five short-structured preventative-medicine and international-public-health questions, 60 minutes); and Paper 4 (50 parasitology/entomology images with questions, 90 minutes). The blueprint is roughly 60% clinical infectious diseases and tropical medicine, 25% preventative medicine and public health, and 15% non-communicable disease. The syllabus in the Apothecaries Guide is the official reference; an AI tutor's output is a third-party claim to be verified against it, WHO guidance and the SmPC/eMC.
Prompt taxonomy: six jobs an AI tutor can do well
Confine the tutor to six jobs, each of which supports your reasoning rather than replacing it:
- Mechanism — explain why a rule holds (pathophysiology, pharmacology, vector biology).
- Discriminating features — separate two entities you keep confusing.
- Option elimination — critique your reasoning for excluding distractors.
- Guideline verification — state the current first-line approach and its source and date.
- Counterfactuals — change one variable and ask how management changes.
- Retrieval testing — have the tutor quiz you without revealing answers.
Twenty-five reusable prompts, grouped by the six commonest error types
Use your own words for the clinical detail; never paste proprietary question text from any bank into the tutor.
Mechanism gaps
- "Explain the mechanism by which P. falciparum causes cerebral malaria; keep it to five steps."
- "Why does severe malaria cause hypoglycaemia, and how does quinine change that risk?"
- "Explain the immunology of the schistosome egg granuloma and how it drives periportal fibrosis."
- "Why does visceral leishmaniasis cause pancytopenia and hypergammaglobulinaemia?"
- "Explain how Vibrio cholerae toxin produces secretory diarrhoea at the level of the enterocyte."
Discriminating-feature confusions 6. "Give me three features that reliably distinguish cutaneous from mucocutaneous leishmaniasis." 7. "How do I tell Trypanosoma brucei gambiense from rhodesiense clinically and epidemiologically?" 8. "Contrast the blood films of P. vivax and P. falciparum in three bullet points." 9. "Distinguish typhoid from typhus by incubation, vector and first-line treatment."
Option-elimination failures 10. "Here is my reasoning for excluding four options in a fever-in-a-returning-traveller scenario — tell me where my logic is weak, without giving me the answer." 11. "Critique my rule for choosing between artesunate and quinine in severe malaria." 12. "I ruled out dengue because of the timing; challenge that assumption." 13. "Play devil's advocate against my choice of praziquantel here."
Guideline and first-line errors 14. "State the current WHO first-line treatment for uncomplicated falciparum malaria and give the source and its date." 15. "What is the current first-line for severe malaria, and how confident are you in the date of this guidance?" 16. "Summarise current rabies post-exposure prophylaxis, flagging any point where guidance has recently changed." 17. "Give the current recommended regimen for visceral leishmaniasis in East Africa and cite the guidance."
Counterfactual and edge cases 18. "Take this uncomplicated malaria case and make the patient 30 weeks pregnant — what changes?" 19. "Same case, now the patient is a two-year-old — how does management differ?" 20. "If this returning traveller were also HIV-positive with a low CD4 count, what shifts?" 21. "Change the setting from a UK hospital to a rural clinic with no PCR — how does your approach adapt?"
Retrieval and spacing 22. "Quiz me with five short unlabelled vignettes on haemorrhagic fevers; do not reveal answers until I respond to each." 23. "Ask me to name the vector and one control measure for five arthropod-borne diseases, one at a time." 24. "Give me three cloze-deletion prompts on antimalarial mechanisms and wait for my answers." 25. "Test me on the incubation periods of five imported infections; correct me only after I commit."
The anti-answer-leak rule
Before you open the tutor on any item, commit an answer and a one-line rationale in writing. If you open the explanation first, the model's confident prose overwrites your reasoning and you learn nothing about your own process. This is the single most important habit when the tutor is a general model with no DTM&H grounding: your committed rationale is the thing being tested, and the tutor is there to interrogate it, not to supply it. Our pillar on auditing an AI tutor develops this into a full grounding, leakage, hallucination and retention check.
The jurisdiction check — a reusable prompt
Because a general tutor may default to US framing, keep one prompt permanently to hand: "Does the answer you just gave reflect current UK/DTM&H-relevant guidance and WHO recommendations, or another jurisdiction? State which body and the date of the guidance you are relying on." Tropical-medicine regimens, prophylaxis schedules and notifiable-disease rules vary by jurisdiction and change over time, and a general model will not flag this unless forced to. Cross-check anything it asserts against WHO guidance and, for UK-licensed medicines, the SmPC/eMC.
Create a misconception record
For every corrected error, write four lines and no more: one sentence stating the incorrect rule you held; one sentence stating the corrected rule; one transfer question that tests the corrected rule in a new context; and one review date. This record, not the tutor transcript, is what you revisit. It converts a fleeting correction into a scheduled retrieval event, which is where the AI-graded-feedback discipline in our calibration pillar pays off.
Weekly verification sample
Once a week, take a small random sample of the tutor's outputs — say five — and check them independently against official or primary sources: WHO guidance, the Apothecaries syllabus, a peer-reviewed reference or the SmPC/eMC. Log every discrepancy. If your error rate is non-trivial, you have quantified exactly why you must not outsource reasoning to this tool, and you can down-weight it accordingly. A general model with no tropical-medicine curation should be assumed to hallucinate specifics until your own audit says otherwise.
Worked example: a seven-day plan for a travel and global-health doctor
The tutor has one job — mechanism and discrimination coaching — and iatroX has another: unseen transfer measurement. This is a pattern, not a proprietary algorithm.
- Day 1: Attempt a mixed set of DTM&H items from your own materials; commit answers before any explanation.
- Day 2: Feed only your wrong items' principles to the AI tutor using the mechanism and discrimination prompts; write misconception records.
- Day 3: Run the jurisdiction check on every guideline claim from Day 2; correct your records against WHO/SmPC.
- Day 4: Test transfer on a fresh, timed unseen block in iatroX — same principles, new wrappers.
- Day 5: Paper 4 image drill plus retrieval-testing prompts (22–25); no answer-leak.
- Day 6: Weekly verification sample of five tutor outputs against primary sources; log discrepancies.
- Day 7: Retest the week's misconception records; reschedule the survivors. Leave your reserved unseen pool untouched.
Decision checklist: continue, supplement, switch or stop
- Continue using the tutor if your weekly verification error rate is low and your unseen scores are rising.
- Supplement with a properly curated infection or tropical-medicine source if the tutor's specifics are unreliable — likely, given no DTM&H grounding.
- Switch your reliance to primary sources and your own materials if verification repeatedly catches confident errors.
- Stop using the tutor for factual claims (keep it only for mechanism coaching) the moment it costs you more time to verify than it saves.
Bottom line
An AI tutor is a reasoning partner, not an oracle — and a general one with no tropical-medicine curation, like Open Exam Prep's "Ask AI", must be kept on a short leash of committed answers, jurisdiction checks and weekly verification. Used that way it sharpens discrimination and mechanism; used passively it launders confident errors into your revision. Keep the reasoning yours, and let iatroX measure whether it transfers.
FAQ
Is Open Exam Prep enough for DTM&H on its own? No. Verified 21 July 2026, Open Exam Prep does not cover tropical medicine or the DTM&H at all — its scope is nursing, pharmacy and allied-health certifications — so it cannot be your DTM&H resource. Its "Ask AI" tutor can coach general reasoning if you feed it your own material, but you still need the approved DTM&H course, syllabus-mapped question practice and image drills. Its vendor-reported figures (265,364 questions across 2,018 exams) describe other exams, not this one.
Which DTM&H component does Open Exam Prep not reproduce well? All of them as bespoke DTM&H practice, because it holds no DTM&H content. Even as a general tutor it reproduces the parasitology image paper (Paper 4) and the structured public-health paper (Paper 3) least well — image recognition and planned written answers are not what a text-based general tutor trains. Use it, if at all, for mechanism and discrimination on the MCQ knowledge, and rehearse Papers 3 and 4 with dedicated material.
How should I verify Open Exam Prep AI answers for DTM&H? Commit your own answer and rationale first, then run the jurisdiction check prompt to force the tutor to name the guideline body and date, then cross-check every specific against WHO guidance, the Apothecaries syllabus and the SmPC/eMC for UK-licensed medicines. Take a weekly random sample of five outputs and audit them against primary sources, logging discrepancies. Assume a general model with no tropical-medicine curation will occasionally hallucinate specifics until your audit shows otherwise.
When should I stop using Open Exam Prep and move to mixed mocks? Stop leaning on it for factual claims as soon as your verification sample keeps catching confident errors, and in any case move to mixed, timed mocks two to three weeks out. Late-stage revision should be dominated by full-length timed MCQ practice, timed Paper 4 image sets and structured Paper 3 answers under exam conditions — not by open-ended chat with a general tutor.
How should I combine Open Exam Prep with iatroX without duplicating practice? Give each a distinct job. Use the AI tutor only to interrogate the reasoning behind items you have already attempted and committed to, and use iatroX to measure whether the corrected rule transfers to unseen, timed items. Because you never paste bank items into the tutor and iatroX supplies fresh unseen wrappers, the two never rehearse the same question, and your calibration stays intact.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 21 July 2026. Open Exam Prep's figures (265,364 questions; 2,018 exams; free access) are vendor-reported on that date and cover non-tropical-medicine exams; verify on the product site, and note the platform does not list DTM&H or tropical medicine. Disclosure: iatroX operates a competing question bank and a Socratic Tutor; here iatroX's role is confined to unseen-MCQ transfer measurement and spaced retrieval, jobs a general third-party tutor does not perform for DTM&H. Corrections are welcome via the feedback route on iatrox.com.
References: the Society of Apothecaries — Diploma in Tropical Medicine and Hygiene and the Guide to the Diploma (2025); Open Exam Prep (vendor-reported); and, for method, the iatroX pillars on auditing an AI tutor, calibrating AI-graded feedback and "Your Q-Bank Percentage Is Not Your Exam Score", plus the comparison hub.