A first episode of psychosis rarely arrives without warning. It is usually preceded by months of gradual change, and how long that goes on before anyone treats it is one of the few things that reliably predicts how the person does afterwards.
Our overview of psychosis covers what the condition involves and how it is treated; this article is about the period beforehand and why noticing it has a measurable payoff.
What the early period actually looks like
The changes are usually gradual and social before they are ever strange. What families describe afterwards is withdrawal, sleep coming apart, school or work performance dropping, and motivation falling away, over weeks to months rather than overnight.
Unusual experiences, when they appear, often arrive in a hedged form first. The person finds them odd rather than believing them outright, and many people in this phase know something has shifted and cannot put it into words. That partial insight is easy to dismiss precisely because the person is describing it uncertainly, and it is worth treating as information rather than as reassurance.
Why the delay is the part that matters
The length of time psychosis goes untreated predicts outcome, which is unusual in psychiatry and is what makes this worth acting on. Reviewing the first-episode schizophrenia literature, a meta-analysis found longer untreated periods associated with more severe symptoms, poorer functioning and a weaker response to treatment. [perkins-2005-dup]
A schematic of the direction of the association reported in Perkins et al. (2005), drawn to show the shape rather than measured values from any single cohort.
Read that as a direction rather than a measurement. The finding is an association across many studies, not a promise about one person, and shorter delay is not the only thing that differs between the people at either end of it.
What follows from it is narrower than it first appears. Reviewing early detection and intervention strategies, a more recent meta-analysis found that early intervention services improved quality of life, employment, negative symptoms, relapse rates, admission rates and functioning, while campaigns aimed at shortening the delay had only limited effect on the delay itself. [salazar-2024-dup] So the reliable lever is reaching the right service, not public awareness in general.
Is this worth an assessment?
This is about whether to seek an assessment, not about reaching a conclusion. Think about the last few months compared with how this person usually is.
0 of 6 ticked
Several changes together, over months, in a direction that is unlike the person, is the pattern worth taking to a doctor. Describe what changed and over what period rather than naming a condition: the timeline is the part hardest to reconstruct later.
One or two of these describe a great many ordinary situations, including depression, heavy cannabis use and a difficult few months. Asking directly about what you have noticed costs nothing and usually tells you more than watching does.
None of this matching is genuinely reassuring. If you are worried for a reason that is not on this list, that is still worth raising with a doctor, because the list describes a common pattern rather than the only one.
No screener on this site assesses psychosis. The hub above covers anxiety, depression, stress, sleep, burnout, self-esteem and loneliness.
What this does not establish
Almost nothing on the list above is specific to psychosis. Withdrawal, sleep disruption, falling performance and lost motivation describe depression at least as well, and they also describe ordinary adolescence, heavy cannabis use and the aftermath of something hard. Our guide to cannabis and psychosis covers one of those overlaps in detail.
That overlap is an argument for assessment rather than for waiting. The whole point of an assessment is to distinguish between explanations that look identical from outside, which is not something a family can do at home and not something this article can do either.
And the association between delay and outcome is an association. Longer delay may partly mark a slower, more insidious onset that would have gone worse regardless, and the studies cannot fully separate those.
When to seek help
Speak to a doctor if someone has changed in several of these ways over weeks to months, and say what changed and when rather than what you think it is. Ask specifically whether an assessment is warranted; in many places there are specialist early intervention services, and being referred to one is the step that the evidence most supports.
Seek help the same day if the person is frightened by what they are experiencing, is acting on beliefs that put them at risk, cannot care for themselves, or is talking about harming themselves or anyone else. Urgency here is not an overreaction, and the delay is the thing this whole article is about.
If you are having thoughts of harming yourself, treat that as urgent and contact your local emergency services or a crisis helpline.
How MyFreud can help
MyFreud is useful here for the timeline, which is the single most useful thing to bring to an assessment and the hardest to reconstruct from memory. Dated notes on sleep, mood and what changed give a clinician the shape of the last few months instead of an impression formed from the worst week.
Download MyFreud and start today: App Store or Google Play.