Training 3,000 medical students in five months: how the cascade model works
Buying training seats one at a time does not scale past a few hundred people. Certifying instructors inside the institution does. Here is how that worked across a state.
Between May and September 2025, working with the Andhra Pradesh Department of Medical Education under Project EASE, we trained close to 3,000 first-year medical students across 17 government medical colleges in peer mental health support and QPR gatekeeper skills.
The interesting part is not the total. It is the arithmetic that made the total possible.
Why buying seats stops working
Gatekeeper training bought per person has a cost per head that does not fall with volume. For a department of forty, that is fine. For a state programme aiming at hundreds of thousands of students, teachers and health workers, it is not a budget line anyone approves twice.
It also creates a dependency. Every new intake needs the supplier again, which means the institution never develops the capacity itself, and the programme stops the moment the funding does.
What a cascade does instead
Certify a small number of people inside the institution as instructors. They train their own cohorts, at their own pace, without paying per head. The supplier trains the trainers and then steps back.
The economics invert somewhere around a few hundred people, and after that the gap widens fast. A state programme is not a large purchase of training. It is a small purchase of instructor certification followed by a large amount of internal delivery.
What it demands in return
Cascade models fail in predictable ways, and it is worth naming them.
Fidelity drifts. Each layer of delivery is slightly further from the source material, and without a fixed curriculum and periodic refresher the content erodes into a general talk about being kind to one another.
Certified instructors leave. If you certify one person in a department, you have bought a single point of failure. Certifying three or four costs marginally more and survives a resignation.
Delivery goes unrecorded. Institutions that cannot say who has been trained cannot answer a regulator, and increasingly will be asked to.
And the referral end has to exist. Training a thousand people to recognise a student in crisis is actively unhelpful if there is nowhere for them to refer that student. On the Andhra Pradesh programme the referral route ran to Tele-MANAS and to institutional counselling, which is why the training could be delivered honestly.
Who to certify first
Not the most senior people available. The ones who will actually run sessions afterwards.
In our experience that means training officers, hostel wardens, student counsellors and the faculty members students already go to informally. Seniority is close to irrelevant. Willingness to stand in front of a room every term is the whole qualification.
Project EASE is a suicide prevention framework of the Government of Andhra Pradesh's Department of Medical Education, delivered in partnership with QPR Institute India, the Suicide Prevention India Foundation and the American Association of Physicians of Indian Origin. Figures cited are for the May to September 2025 medical college programme.
Training a college or a department?
Tell us the institution, roughly how many people, and any date you are working towards. We reply with formats and a quote within a working day.