What it is
A practice consultation, from start to finish
Since 2016, Escola de Pacientes has trained students with simulated patients: someone plays the person seeking health care, and the student conducts the whole consultation. What is trained there is not getting the diagnosis right — it is the conduct: asking what needs to be asked, in an order that makes sense, without missing the warning sign that changes management.
Since 2023 that patient can also be digital: an artificial intelligence that follows a clinical script written by the team, based on Primary Care and SUS guidelines. It answers as the patient would — including staying silent about what wasn't asked. It is practice at any time, without depending on another person to play the part.
How it works
Three steps, and the consultation begins
There is nothing to install, no sign-up, and nothing to print. You open the address, enter the code the professor gave you, and you are already seeing patients.
Choose the case and the AI
Choose from the catalog, reading the topic of each case, or ask for a random patient, without knowing what they complain about. Then choose which artificial intelligence you want to talk to.
Conduza a consulta
Ask as you would in the clinic. The patient does not hand over the diagnosis: they answer what you investigate and stay silent about what you don't ask.
Get the feedback
When you finish, the system compares what you did with what was expected in that case and shows, item by item, what you covered and what was left out.
Each case is a different person, with their own history and way of answering.
Choose your favorite AI
You are not stuck with a single artificial intelligence. You choose which one will give the patient a voice before starting and can switch for the next consultation: same case, different AI.
Switch whenever you like
See patients with one today, another tomorrow, and compare how each plays the patient's part. If one gets slow or goes down, you switch and carry on — the activity doesn't depend on a single company.
The requirement is always the same
The system does the grading, using what the clinical team defined for that case. Switching AI changes who you are talking to, not what is required of you.
Everything on one screen
The conversation, the checklist, and the result come one after another, in the same place. There is no need to open a separate form or save anything for later.
What you get at the end
The conversation is the medium; what remains is what comes after it. When you finish, each item expected for that case is marked in one of three ways — you did it, you half did it, or you did not do it — with the excerpt of your own consultation that justifies the marking.
Not every item is worth the same
What is more serious weighs more in the result: introducing yourself to the patient does not count the same as a decision that could put them at risk. And what harms the patient takes points away — it is not enough to remember everything; you must also not do what causes harm.
Seven areas, always the same
Every consultation is read through the same seven areas, in the same order — so you can compare today's consultation with last month's. When the case does not require one of them, it appears as not assessed, never as zero: zero would read as your mistake, when in fact the subject was not in play in that case.
This feedback is for you to reflect on your own consultation. It is not a grade, does not go on a report card, and does not replace the conversation with the professor — which remains part of the activity, as it always has in the School's simulations.
Two ways to start
Both lead to the same screen and the same feedback at the end. What changes is how much you know before saying good morning.
Choose from the catalog
You read the topic of each case and choose. This is the path for those who want to study a specific subject — hypertension this week, prenatal care next.
Draw a random patient
You receive someone without knowing what they complain about, as happens in real care. It is useful for practicing reasoning from scratch — which is why even the page address does not give away the topic.
How cases are organized
Each patient is placed in a group of ICPC-2, the classification used in Primary Care visits in Brazil. The group is chosen by the reason that led the person to seek care, not by the diagnosis. Someone who arrives with back pain falls under Musculoskeletal, even if the workup ends up elsewhere — if the label were the diagnosis, the answer would be given before the first question.
There is a second reason for this: SUS visits are also recorded in ICPC-2. By organizing cases the same way, what students practice can be compared with what the network actually sees.
Groups that don't have any patient yet appear in the catalog anyway, on purpose. It is this gap that shows the team which case needs to be written next.
How to access
Ask the course professor for the class code. It is given out in class, is valid for the semester, and is all you need: with the code, the artificial intelligences are unlocked and you can start seeing patients. The code says which class the consultation belongs to — never who did it. It can belong to a class, a group, or a pair, and it is how the professor follows performance.
What is recorded from your consultation
SimulaPacientes also serves the group's research. Since this involves data from people who study, it is worth saying plainly what happens.
Identification is not required
There is no field for name, email, student ID, or internet address. This is not a promise that nobody will look: it is that there is nowhere to store that data. Without an account, the consultation is not linked to any profile.
The conversation is not stored
Only the result of the check is recorded — which actions appeared and which did not. The transcript of the conversation is not stored in the SimulaPacientes database.
The reading is of the class, not of you
What the team sees is where the group stumbles most, by topic and by area. It is used to rewrite cases and plan classes — not to assess any individual.
Data collection is in a test phase, run by the team itself. Using this data in research with students depends on the project being approved by a research ethics committee, which is running in parallel with development.
