Prevention of psychosis with a cognitive behavioural intervention in help-seeking young people with an at risk mental state for developing psychosis

Schizophrenia has an unfavourable course in 80 percent of the cases. These
patients are either chronically psychotic or have a relapsing course. The
relapses cause high costs in mental health services because of multiple
rehospitalisation and continuous care. The cognitive deficits and negative
symptoms cause an unemployment rate of more than 80 percent and an unmarried
status with little offspring. Prior to the first psychotic episode 80 to 90 %
of the patients have experienced retrospectively symptoms of escalating
severity over 1 to 2 years; the so called prodromal phase.
Prospective longitudinal studies have found that Psychotic-like experiences
(PLEs) are quite prevalent in the population. These PLEs are transient in
nature and only develop into florid psychosis when persistent and combined with
other risk factor such as urbanisation, trauma or cannabis use.
Studies have examined the possibility of detecting individuals in the prodromal
stage, prior to the development of florid psychosis. Mc Gorry and colleagues
have developed operational criteria to identify a group among help-seeking
young people of an 'at risk mental state' (ARMS). Such an ARMS makes this group
at ultra high risk for developing a psychosis within a year time.
This ARMS group is divided into subgroups. The first group is the group with a
genetic endowment. When a person declines in social functioning and a first
degree relative has schizophrenia or the person itself has the diagnose of
schizotypy, then the person is at risk. This group forms about 15% of the total
ARMS group. Another small group of about equal size is the group that has had
Brief Limited Interval of Psychotic Symptoms (BLIPS). These transient psychotic
symptoms lasted less then a week. The largest subgroup of about 70 % is formed
by the group with attenuated psychotic symptoms. These groups could well be
considered as different endophenotypes and genotyping might be of prognostic
value. The ARMS group develops psychosis in 13 to 54 percent of the cases
within a year. The average transition rate is 37 percent in a year. The ARMS
group comprises many late prodromal patients. The CIDI psychosis section can
detect an ARMS group with high transition rates as well. Two PLEs on the CIDI
with a emotional disorder has a transition rate of 40 percent. Based on these
findings the question arose whether effective interventions could be developed
that would prevent or delay the transition into florid psychosis in the late
prodromal state.

Three innovative studies have been undertaken to intervene in the ARMS group to
reduce the transition rate into psychosis. The Melbourne study compared
needs-based intervention with low dose risperidone (1.3 mg/d) and CBT combined
(n=67). The results were that 35% of needs-based intervention developed a
psychotic episode versus 10% in the risperidone CBT group. The Manchester trial
compared treatment as usual with CBT (n=58). In the TAU group 26% developed a
DSM-IV psychosis diagnosis while the CBT group showed only 6 % transitions into
psychosis. The Copenhagen trial compared integrated treatment with standard
treatment (n=79). After 12 months 33% was diagnosed with psychotic disorder in
schizophrenia spectrum in the standard care group and 8% in the integrated
treatment group. These are encouraging results in the short term. The reduction
in relative risk (RRR) varies from 46% to 78% at 9 to 12 month follow-up. The
Copenhagen study also has two year follow-up data that show an RRR of 48%. The
three year follow-up from the Manchester trial shows a RRR of 34%, which was no
longer statistically significant.

As persistence of PLE is a risk factor for psychosis, the reduction of this
risk factor could prevent transition to psychosis in future. At this moment,
CBT and other psychological interventions are capable to reduce the risk for a
psychosis in the short term. Although this is only a delay for some patients, a
health gain of one year of more is a significant finding in a devastating
condition such as schizophrenia.

The specificity of the intervention is still not known. Melbourne combined CBT
with anti-psychotic medication, so the relative contribution is unclear. The
Copenhagen study provides a whole package of interventions which still has to
be sorted out. A replication of the Manchester study seems the best option. The
intervention has the largest effects sizes reported so far and is directed at
thinking styles that form the basis of the development of psychosis. A
manualised protocol has been developed.

The innovative part of this study is to examine the effectiveness of a specific
intervention targeted at the prevention of florid psychosis in both a
population based sample and a clinical sample. This will take place in
help-seeking young people from 18 until 35 years of age.

Research has shown that people with ARMS have a current axis 1 disorder in 76
percent of the cases. We will not miss many of the patients with ARMS with our
selection strategy to start in help-seeking people. Also the transition rate to
psychosis is higher in people with an axis 1 disorder (30%) compared with those
who without an comorbid disorder (13%).

The study aims to reduce the number of transitions into psychosis in a group of
people who are in an at risk mental state. This is done with a specialised
cognitive behavioural therapy.
Other aims are:
1) Implementation of screening of all people in the age range from 18 to 35
a. that seek psychotherapeutic help at the mental health services in The
Hague and
b. that are referred to a specialised clinic on suspect of development of
psychosis with the Prodromal Questionnaire.
2) Implementation of the CAARMS in the services in The Hague and Amsterdam to
discriminate between below threshold ARMS, ARMS, and psychosis.
3) Implementation of a CBT intervention that is specifically adapted to treat
ARMS with a co-morbid disorder.
4) Test the reduction of persistence of PLEs over an 18 month period.
5) Test the transition rates of different genotypes
6) Develop materials and training packages to facilitate spread of these
services to other mental health facilities.

All consecutive referrals to the mental health services in the age from 18 to
35 years in The Hague (PsyQ) and the referrals from the mental health services
of Amsterdam to the specialised Adolescent Clinic in the age of 18 to 35 years
will be screened with the Prodromal Questionnaire. In the general population
the prevalence of one psychotic feature is about 17 percent. In the help
seeking population this prevalence will be higher and is estimated at 25
percent. The prevalence of ARMS is estimated at 3.5 percent. This is a
conservative estimate as Yung found 29 percent of the general help-seeking
population of young people aged 14 to 25 met the ARMS criteria [10]. During two
years about 6000 people in the age range from 18 to 35 years seek help at PsyQ
for a psychic problem. About 1500 people will self-report at least one
psychotic feature, 210 will be in an ARMS and have a co-morbid disorder. This
group will be selected for the intervention study. All estimates are based on
relatively small studies and the security intervals are to be regarded as quite
large. For this reason the catchment area is elaborated with Amsterdam. The
Adolescent Medical Centre has experience in screening people with an at risk
mental state as they are involved in the DUPS study.

In the selected group, the transition rate will probably be 35 percent to
psychosis over an 18 month period and about 40 percent will have persistent
PLEs. The intervention will aim top reduce the transitions to 17 percent and
the persistence as an important risk factor for later transition to 20 percent.

Furthermore, measurements have been matched with the EDIE-2 trial. A comparable
study in Britain in Manchester, Birmingham, Glasgow, Cambridge and Norfolk.
This will allow us to compare the results of the UK trial and the Dutch trial.
It would be possible for certain analysis to combine data to allow greater
statistical power. We will request permission with the sponsor of the UK study
who owns the British data. The next meeting will be end of February 2007.

Power calculation:
The only study that used a help-seeking population fond a 24 percent transition
in 12 months[11]. We will calculate power on an expected transition rate of 35
percent over 18 months. To be conservative power we expect a 50% reduction in
the transitions which is smaller than the reductions reported thus far. The
sample we need for a 2-tailed test of the proportions with an alpha of .05 and
a power of .80 is 2 x 93 for reduction of the transition rate from 35 percent
to 17 percent over an 18 month period and 2 x 82 for the persistence of PLEs.
With 20 percent drop-out over 12 months, we need to include 240 persons into
the trial. The refusal rate is expected to be very low as was the case in the
Manchester trial as people are seeking help and get an additional help offer.
Both research sites will (The Hague and Amsterdam) recruit 60 ARMS patients
each year.

Randomisation: The inclusion period will stop after 240 young people have
consented to participate in the study. These are randomised over the cognitive
behavioural intervention or treatment as usual.

Time table:
Month 0 to 2: Preparation of the sites. Training of the CAARMS raters. Training
of the therapists in the targeted CBT protocol.
Month 3 to 26: Inclusion of patients. 6 month intervention stage (targeted CBT
or TAU and 18 month since inclusions follow-up).
Months 27 to 44: Last therapies of 6 month envelop and 18 month follow-up
period.
Months 45 to 48: Analysis and report.

Intervention: The effective intervention is the Dutch translation of the
protocol as developed by French and Morrison and was demonstrated to be
effective. This is a formulation driven cognitive behavioural intervention
directed at reducing symptoms and normalising psychotic-like experiences and
preventing that an catastrophising appraisal will occur. Such an appraisal is a
next step on the road to psychosis. The control condition is treatment as usual
at PsyQ in The Hague and monitoring in Amsterdam.

N.A.