If you ask our assistant whether a headache is normal, it will not answer. It will tell you that the question is a medical one and point you to where it can be answered. It does that by design, not for lack of capability.
Assessing a symptom is a clinical act. Doing it well demands context an app does not have; doing it badly reassures someone who should not be reassured. The cost of referring you unnecessarily is one tap; the cost of reassuring you wrongly is not one we would be the ones to pay.
That is why the warning-signs screen has not a single box to fill in: we publish the official list, cited, identical for everyone, and every action it offers escalates — "it is probably nothing" does not exist in our vocabulary.
That boundary is written into the code, into the database and into our automated tests. If you ever see it broken, report it: we treat it as the most serious defect we can have.
This blog is about paperwork and how the health system works. It does not assess symptoms and it does not replace the appointment with your doctor or your midwife. If it is an emergency, call 911.
Back to the blog
