Lecture 5 - Early Intervention in Psychosis RCT evidence

Question 1

Should we be intervening those in ARMS

Answer 1

For:
-Evidence shows a high transition rate from ARMS to psychosis, so it is worth ‘catching it early’

Against:

• Waste of time and NHS resources
• Issue of incorrectly diagnosing and stigma involved in labelling of the disorder
• Only 20/30% transition, so you might be inducing MH problems in them by worrying them that theyll become psychotic.

Question 2

What did Marshall et al (2005) find

Answer 2

Meta-Analysis of outcomes for FEP patients, uncovered that people who have longer period of untreated psychosis, tend to be more unwell after treatment and have poorer outcomes in several domains (quality of life, social functioning, depression/anxiety)

Question 3

What is the rationale for early intervention

Answer 3

Rather than preventing psychosis, it is about reducing the length of untreated psychosis.

Question 4

Findings from Yung et al (1996) and FEP patients

Answer 4

Retrospective accounts of 21 FEP patients said they had life threatening and disabling suicidal symptoms during the prodrome

-but can we believe retrospective accounts now?

Question 5

Findings from Melle et al (2006)

Answer 5

Rates of severe suicidality, significantly higher in communities without early detection programmes

Question 6

What is the added benefit of early intervention programmes for ARMS individuals who do not transition

Answer 6

Subthreshold and prodromal symptoms are still very difficult for the individuals, It is not just about preventing transition, but treating the individual and their symptoms - just because they do not transition to full-blown psychosis, does not mean they do not have poor outcomes still.

Question 7

Positives to early intervention

Answer 7

• May improve engagement with services, particularly later on if the individual gets worse
• May reduce the trauma and stigma of hospitalisation and use of the MH act
• May reduce psycho-social deterioration
• There is the possibility of preventing the transition to psychosis in young people

Question 8

When were the first anti-psychotics introduced

Answer 8

Around the 1950s onwards. But there were a lot of side-effects: weight gain, nausea, headaches

Question 9

What was the first intervention study into psychosis

Answer 9

the PACE trial - McGorry et al (2002)

Question 10

What was the methodology for the PACE trial

Answer 10

• Treatment included a combination of low dose antipsychotic (risperidone) and CBT
• Werent particularly interested in any specific treatment, just if ANYTHING led to a chance in transitions rate
• N=59 at one site, intervention for 6m, followup 12m
• had 2 groups (specific prevention intervention and needs based intervention i.e. treatment at usual)

Question 11

Findings of the PACE trial

Answer 11

• There was a sig diff at the end of treatment (6 months) between control and treatment group
• However, there was no sig difference at the end of follow up (12m)
• Transition rates as 12m were: 19% in treatment, 36% in control and all happened within the first 6 months for controls, with an additional 3 transitions in the treatment arm after 6 months.

Question 12

What did the PACE trial find when considering the treatment types separately

Answer 12

• Adherence for CBT was high, but variable for anti-psychotics
• If you split up the treatment group into adherent and non, those who did not adhere, were more likely to transition
• This was the case for those not adhering to the anti-psychotics.

Adherence to risperidone provides a more lasting effect on transition than CBT but generally people do not like taking them. CBT does not provide as much of a lasting effect, but people prefer doing it so adherence is higher

Question 13

Who conducted the PRIME trial, and what was it observing

Answer 13

McGlashan et al (2006) - tried to look at the effects of antipsychotics (Olanzapine) alone, on transition to psychosis

Question 14

Methodology for McGlashan et al

Answer 14

• 60 patients across 4 sites (this separates effect of intervention from the principle invsstigator)
• Double blind
• Olanzapine and placebo
• 1 year treatment, 1 year follow up

Question 15

Findings for McGlashan et al

Answer 15

-No significant difference on transition, but there still was an advantage of using the treatment:
Treatment group = 16% transition
Control = 38% transition

Question 16

Problems with PRIME study

Answer 16

• they only used data from the treatment year due to 12/60 patients completing follow up
• hugely underpowered findings due to drop outs

Question 17

Methodology of Morrison et al (2006) EDIE-1 trial.

Answer 17

• First trial of cognitive therapy alone
• N=58 at one site (37 CT, 23 control) randomised
• 6m intervention, 1 and 3 year follow up

Question 18

Findings of Morrison et al (2006)

Answer 18

• At 12m, significant effect of intervention (6% CT vs 22% control)
• However, by 3 years, this effect had disappeared - once you stop engaging in the treatment, the effects and protection stop too - which supports the PACE trial findings of long term effects of CT

Question 19

What did the replication of EDIE-1 find? (Morrison et al, 2012)

Answer 19

It was a larger, 5 site trial, in order to see if insignificant effects were due to power.
But again, found no significant effect of CT after 3 years.

Question 20

What was a key issue in the EDIE-1 trials

Answer 20

The study was severely damaged by overall low transition rates.
The control group already had very few transitions, so in order to see a statistical significant effect of treatment, the treatment group would need to have even fewer transitions (which is extremely difficult to find)

Question 21

Who studied into the effects of Omega 3 Fatty Acids

Answer 21

Amminger et al (2010) - Vienna Study

Question 22

Methodology of the Vienna Omega 3 Study

Answer 22

• N=81 at one site
• 12week intervention and 40 week follow up
• Double blind with placebo (coconut oil)
• HUGE does of omega 3, equivalent of 1.5kg of fish
• Was a needs based intervention, meaning if they saw the people were becoming unwell, they gave them therapy and medication where needed but balanced on both arms of the trials

Question 23

Findings of Vienna study

Answer 23

• 5% transition in Omega3, 28% in placebo
• Very significant effect, with reduction in positive, negative and general symptoms.
• HOWEVER needs replicating with multiple sites

Question 24

What are the benefits of Omega 3 fatty acid

Answer 24

They are proven to help in brain functioning and development

Question 25

What are the issues with RCTs of EI in psychosis

Answer 25

• Many trials have been inconclusive and insignificant due to underpowered samples, drop outs, and use of only one site
• Argument surrounding if transition rates are meaningful and useful - can we really be drawing a dichotomous line in diagnosis for ‘you are/you are not’ psychotic and unwell.
• Issue of investigator inflating the patients initial scores to get onto the trial in the first place - meaning we are never going to see a sig difference in TR, since the individuals were never the unwell to begin with.
• Time consuming and expensive
• Hard to replicate: nobody wants to fund something that has already been done
• No biological marker for treatment success: solely reliant on self report and clinician report
• High placebo effect, especially for psychological therapies

Question 26

What is important to consider with the UHR group regarding their outcomes?

Answer 26

They are not just psychosis-prone, but more general psychopathology-prone and at risk for compromised social functioning (Yung et al, 2010)
Just because they do not transition to psychosis in a particular EI trial, does not mean they are healthy and functioning well - they are still suffering from anxiety or depression and other social functioning limitations (van der Gaag et al)

Question 27

Findings from the meta-analysis by van der Gaag et al

Answer 27

• There is a marginal effect of treatment in general and a marginally sig effect at longer follow ups
• But as a whole, there are a lot of underpowered trials, and even when looking at general factors like having a good social life, keeping a job, having friends, all the effects of the interventions disappear.
• So even if the EIs prevent transition, they are not actually presenting any functional changes in the individual.

Question 28

What did van der Gaag et al find out about preventive interventions

Answer 28

They are effective at reducing the risk of onset of psychosis:
After 12m - 52-54%
After 2-4yrs - 35-37%

Question 29

What does number-need to treat (NNT) mean and what did van der Gaag find for this?

Answer 29

NNT refers to how many UHR individuals would need to receive the treatment, in order to prevent one more transition to psychosis (almost like a transition ratio)

van der Gaag et al found in their meta-analysis a NNT of 9-12 which is quite good compared to other disorders

Question 30

What are the future directions for RCTs of EI in psychosis

Answer 30

van der Gaag et al:

• antipsychotic medications need more trials
• omega-3 FA need replications
• intergrated psychological interventions need replicating with better methodological considerations

Stafford et al (2012):

• use of CBT with and without family therapy
• there is good evidence for family interventions reducing relapse rates in EP and shizophrenia - the whole family is working to support the individual

Question 31

What has been considered strange with transitions rates recently

Answer 31

Transition rates seem to be getting lower in more recent studies, and they vary (reduce) depending on the individuals entry into the study

Cohort effect?

Question 32

What is lead time bias?

Answer 32

Recent studies have recruited individuals earlier in their psychotic symptoms, and need to follow them for longer to look at transition rates, so essentially they have more years to live with and get better, because we have detected the disease earlier than previous studies pre 1998 have.

E.g. we can now detect disease X from birth, meaning the person has a prognosis of 65 years to live, whereas before we detected disease X at age 50, and the individual only had 15 years to live.
One would say the survival rate for the disease has increased, but it hasn’t, we have just got better at detecting it earlier.

Question 33

What are the current guidelines for treating ARMS

Answer 33

CBT with or without family therapy but NO prescription of antipsychotic medication

Question 34

Considerations for the reduced transition rates over the years

Answer 34

• Lead time bias
• ‘Treatment as usual’ might have just improved over the years and become more effective e.g. higher medication dosages
• A dilution effect = identifying more people with lower actual risk despite still meeting UHR criteria e.g. the lower half of the prodrome
• Different types of attenuated psychotic symptoms are associated with different levels of risk for psychotic disorders (Yung et al 2009; Nelson & Yung, 2009)