Duration of untreated psychosis: early intervention rarely shortens it, though it helps those who reach it

The services help those who reach them

Across 10 randomised trials with 2,176 people treated for 9 to 24 months, early intervention services did better than usual care on all 13 outcomes at the end of treatment (Correll 2018). Fewer people stopped treatment (RR 0.70, 95% CI 0.61–0.80), fewer were hospitalised (RR 0.74, 0.61–0.90), and more were in school or work (RR 1.13, 1.03–1.24).

A 2024 review found better quality of life at follow-up in early intervention services (g = 0.600), but no difference in remission or in positive symptoms (Salazar de Pablo 2024).

These are benefits of the services for people who reached them. They do not show that the services shorten the time it took those people to get there.

On average, the time to treatment barely changes

A 2024 review of 14 cohorts from 12 services, with 2,938 people, found that early detection or early intervention strategies, counted together, shortened the duration of untreated psychosis slightly: g = 0.168 (95% CI 0.055–0.283). Its authors call the impact of early detection on this delay limited (Salazar de Pablo 2024). The studies were controlled but mostly not randomised, and they defined the start of psychosis differently.

An earlier review of 16 controlled studies found g = −0.12 (95% CI −0.25 to 0.01) (in this review a negative g means a shorter delay): the interval includes zero, so no reduction could be shown (Oliver 2018). The exception was services for people at clinical high risk of psychosis, g = −0.386. Differences between studies were mostly explained by different definitions of the delay. This review shares authors with the 2024 one.

An independent review of 11 studies covering 8 initiatives found that educating general practitioners and early intervention services on their own did not shorten the delay; evidence on campaigns aimed at several groups at once was mixed (Lloyd-Evans 2011). That review dates from 2011.

The exception: a continuous public campaign in Norway

In the TIPS programme in Norway, one area ran a continuous public campaign together with early detection teams. In 1997–2000, among 281 people across the programme and comparison areas together, the median time to treatment in the programme area was 5 weeks, against 16 weeks in the comparison areas (Melle 2004).

After the campaign stopped, the median in the same area went back to 15 weeks, among 108 and 75 people in the two periods compared. The early detection teams kept working in both periods (Joa 2008). In this programme, the shorter delay lasted while the campaign ran; the teams alone did not keep it.

The limit: this is one programme, in one country, from 1997, described in two reports by the same research group, and the areas were not assigned at random.

Other programmes: partial results or no comparison

The table below lists the programmes. In Birmingham (ED:IT), with two areas and 151 people in the programme period alone, the median in the intervention area fell from 71 to 39 days, by 45%. But the control area itself shifted: its tables show 31.5 days before and 79.5 after, though the abstract reports no change there (Connor 2016). So the result is partial.

In New Haven, USA (STEP, the Mindmap campaign), a non-randomised trial compared the campaign site with a concurrent comparison site in Boston (PREP), over one year before and four years of the campaign (2014–2019), with 171 and 75 people. The median delay went from 311.5 to 149 days at the campaign site and from 324.5 to 430 days at the comparison site, but the difference between the sites was not statistically significant, and the groups before the campaign were small (24 and 12 people) (Srihari 2022). The authors write that TIPS is the only other initiative that shortened the delay in a community with a concurrent comparison group. This is also a partial result from one programme, not a second confirmation of the Norwegian one.

In Hong Kong (EASY, an early intervention service), the delay for adults fell from 180 to 93 days, with no significant change for young people, but this compares 2000 with 2009–2011 without a comparison group (Chan 2018). In Montreal (PEPP), educating all first points of contact did not change the delay over 3 years before and 3 years after, again without a comparison group (Malla 2014). In London (Camden and Islington), time to contact with the service went from 133.5 to 116.5 days, which was not significant; the catchment area widened during the programme (Lloyd-Evans 2015).

A related briefing: how the wait is spread

A related briefing shows that among people who reached care for psychosis, the mean time to treatment is several times the median, because some people wait a very long time: https://noosinstitute.org/research/psychosis-untreated-duration-mean-vs-median. For depression and anxiety, no controlled study we found has even measured whether anything shortens the time to first help: https://noosinstitute.org/research/depression-anxiety-time-to-help-untested

How to quote this

Finding: early intervention services for psychosis help people who reach them, but on average early detection or early intervention barely shortens the time to treatment; only one programme with a comparison group clearly did: Norway's TIPS, with a continuous public campaign together with early detection teams (median 5 weeks against 16), and the delay returned to 15 weeks after the campaign stopped.

Limit: most studies were not randomised and dated the delay differently, and the Norwegian result comes from one programme in one country from 1997, reported by one research group.