Long-Term Follow-up of Preventive Interventions in CHR-P Populations

confidence: high · 6 sources

⚠
Working research — not a recommendation.

This page records what the evidence says a policy does, not what anyone should do about it. It has been synthesized from the cited sources but has not been independently reviewed.

What the evidence says

The sources on this page describe the natural course of the clinical high-risk state for psychosis (CHR-P) well. They say much less about whether preventive interventions change that course over the long term. Individuals meeting CHR-P criteria have a 20% risk of developing psychosis at 2 years and a 22% risk at 3 years. Attenuated psychotic symptoms last on average 2 years. Higher baseline attenuated positive symptoms (OR=2.56) and negative symptoms (OR=2.68) are associated with transition, and good functioning is associated with lower risk (OR=0.59). These are descriptive and associational findings. They set the baseline against which any intervention is measured.

For the intervention question, the weight of the meta-analytic evidence is null on comparative effects. The most recent network meta-analysis found no robust evidence favouring any indicated intervention over another or over needs-based interventions. A Cochrane pairwise meta-analysis reached the same conclusion. A review of 42 meta-analyses found no robust evidence that any indicated intervention beats another, including needs-based care and control conditions, for preventing psychosis or improving any other outcome. That review notes the uncertainty is high. A third source describes the advantages of specific pharmacological or psychosocial interventions over high-quality monitoring and supportive care as modest and at times inconsistent.

The null result extends across outcomes. Aggregate-data meta-analyses summarised in one review found no measured effect of preventive treatments on these outcomes:

Some of these syntheses are old. The quality-of-life meta-analysis dates from 2014, and the depression meta-analysis (5 trials, 714 participants) from 2013. Trials may also be underpowered to detect small effects.

Direct evidence on follow-up is thin. One 2018 source says CBT "may lead to" a reduction in transition rates or a delay of onset, citing prevention trials with 12-month and longer follow-ups. It gives no effect sizes. The one primary trial here with follow-up is a small feasibility RCT (n=40) of the SleepWell sleep intervention. It found large decreases in insomnia severity (d = 2.67 at 3 months, d = 1.91 at 9 months). It also reported preliminary, unquantified decreases in paranoia, anxiety and depression. For psychosis transition, 0 participants in the SleepWell group and 1 in usual care transitioned. The authors state this single event supports no causal inference.

No source on this page reports long-term outcomes beyond transition, such as functional recovery or remission, for any preventive intervention.

Where the evidence splits

The CBT claim and the meta-analytic null findings point in different directions on whether a specific psychosocial intervention changes transition. The 2018 source says CBT may reduce transition rates or delay onset. The network meta-analyses, the Cochrane review and the 42-meta-analysis review find no robust evidence favouring any indicated intervention over another or over needs-based care. The CBT claim is phrased as a possibility and gives no comparator or effect size. Its comparator may differ from the "versus each other" comparisons in the null findings. The sources on this page do not state which difference explains the split.

The SleepWell trial found decreases in depression, while an older pairwise meta-analysis found no effect of preventive treatments on depression. The SleepWell result comes from one 40-person feasibility study of one intervention, and the authors call it preliminary. The meta-analysis pooled 5 trials. The sources do not state whether the intervention type, the population or the sample size accounts for the difference.

Thresholds and dose-response

No threshold established in the current sources.

Open empirical questions

Findings

CHR-P status carries a 20% to 22% risk of psychosis within 2 to 3 years

confidence — direction: highconfidence — magnitude: highconsensus: settled
mechanism
not established; the claims are descriptive cohort rates
effect
20% transition at 2 years; 22% at 3 years (denominator: individuals meeting CHR-P criteria)
conditions / moderators
whether rates generalise across CHR-P services, assessment instruments and settings, and whether longer follow-up shows different cumulative rates, is not established

Attenuated psychotic symptoms last about 2 years, and baseline symptoms and functioning are associated with transition

confidence — direction: highconfidence — magnitude: mediumconsensus: settled
mechanism
not established; associational
effect
attenuated symptoms last 2 years on average; transition is associated with baseline positive symptoms (OR=2.56) and negative symptoms (OR=2.68); good functioning is associated with lower risk (OR=0.59)
conditions / moderators
whether the associations hold across assessment tools and service settings is not established

No indicated preventive intervention shows a measured difference in preventing psychosis onset versus other interventions or control

confidence — direction: highconfidence — magnitude: mediumconsensus: contested, with the weight leaning toward no demonstrated difference (see the CBT finding below)
mechanism
not established
effect
no measured effect difference between intervention types, needs-based interventions or control conditions; one source describes advantages of specific interventions over monitoring and supportive care as modest and at times inconsistent
conditions / moderators
trials may be underpowered to detect small effects; whether longer follow-up reveals delayed or differential effects is not established; the evidence is described as high in uncertainty

CBT may decrease transition rates and delay onset in CHR-P

confidence — direction: mediumconfidence — magnitude: lowconsensus: contested; the comparative meta-analyses above find no robust evidence favouring any intervention
mechanism
not established in the claim
effect
decreases transition rates and delays onset, with amelioration of symptoms; magnitude not specified
conditions / moderators
the source uses the language of possibility ("may lead to"); effect sizes are not quantified; the comparator is not stated

Preventive treatments show no measured effect on symptom, functioning, distress, quality-of-life and acceptability outcomes

confidence — direction: highconfidence — magnitude: high for symptoms, distress, functioning, depression and acceptability; medium for quality of lifeconsensus: settled for the outcomes covered, with the weight leaning toward no measured effect; dated for some outcomes
mechanism
not established
effect
no measured effect on attenuated positive symptoms, attenuated negative symptoms, depression, symptom-related distress, social functioning, general functioning, quality of life or acceptability (dropout)
conditions / moderators
syntheses are pairwise or network meta-analyses of RCTs published between 2013 and 2018; the quality-of-life meta-analysis is from 2014 and newer evidence may exist; the depression meta-analysis covers 5 trials with 714 participants; the general functioning meta-analysis covers 9 trials; the acceptability network meta-analysis covers 14 trials with 1,848 participants

The SleepWell sleep intervention decreases insomnia severity in ultra-high-risk youth

confidence — direction: highconfidence — magnitude: mediumconsensus: thin
mechanism
not established in the claim
effect
decreases insomnia severity; d = 2.67 at 3 months and d = 1.91 at 9 months, versus usual care
conditions / moderators
feasibility RCT, n=40; a larger powered trial would raise confidence in the magnitude

SleepWell shows preliminary decreases in paranoia, anxiety and depression

confidence — direction: mediumconfidence — magnitude: lowconsensus: thin
mechanism
not established
effect
decreases in paranoia, anxiety and depression; magnitude not given
conditions / moderators
described as preliminary; no effect sizes; a larger powered study is needed

SleepWell shows no measured effect on psychosis transition

confidence — direction: lowconfidence — magnitude: lowconsensus: thin
mechanism
not established
effect
no measured effect; 0 transitions in the SleepWell group versus 1 in usual care
conditions / moderators
the trial was not powered for transition; the difference is a single event; the source says no causal inference about preventing psychosis onset can be drawn