How early is early psychosis?
A review of some of
the research surrounding early intervention in psychosis.
Steve Day
1.2.01
Introduction
& Background
Relapse is a common problem for people with psychosis. There is plenty of evidence
to suggest that some relapses, although perhaps not completely preventable, can be identified & managed
effectively in their very early stages through use of medication & psycho social intervention strategies.
Studies also show that relapse in the early phase of psychosis is associated with increased probability
of further relapse & persisting symptoms. (Wiersma et al, 1998.) It isn't hard to deduce therefore that
it may be important to attempt to prevent or minimise relapse. This is particularly relevant in the so
called 'critical period' of the first 3 years of the illness where it seems social & personal disability
develop most aggressively. (Birchwood et al 1998.)
It is only in comparatively recent times that
the treatment of schizophrenia has started to move away from Max Birchwood's traditional comparison with
the last days of the British empire, i.e. "The orderly management of decline" (anon.) to something more
proactive. Notable in this respect have been the attempts to identify & codify early warning signs that
may herald the onset of psychotic illness or relapse, & to strategically intervene at an early stage
in the hope that it will prevent (or reduce the severity of) the relapse, or even illness onset. Not
only this but, successful identification & intervention may assist in minimising possible depression,
anxiety, hopelessness, humiliation, entrapment & defeat, problems that can occur when sufferers perceive
they have little or no control over their condition (Birchwood & Iqbal, 1998). Mueser et al (1992) also
report that sufferers rated as "Extremely important" the need to learn about "Early warning signs of
the illness & relapse," which Birchwood (1996, page 171) says is driven by "A perceived need for control."
Early intervention, it is urged, is also needed in the sense of early on in the course of the illness,
"We should be piling in much sooner than we do & not, as so often, 'waiting & seeing' who's illness
becomes chronic, & only then piling in with rehabilitation & extra help." (Birchwood, 1998)
Unfortunately,
(& for patients in particular), psychiatry has not always grasped this retrospectively simple concept.
Nevertheless, there has been a substantial body of research in this area. However, it is an area potentially
fraught with ethical & other problems, & of course one of these is that if there is no service available
to act appropriately once the warning signs are seen - & it probably needs something more than the blunt
tool of standard generic C.M.H.T. follow-up (or often non-follow-up!) or routine outpatient care that
is so often the norm - then the previous exercise is somewhat wasted.
Early Warning Signs - before
the first episode
Identification of early warning signs is nowhere more difficult than in the
prediction of illness in those who have not yet developed it. Perhaps this is because, as with so many
things, vulnerability to psychosis is a continuum. In Shepherd's (1989 - p34) seminal study of the natural
course of schizophrenia he observes "As the time from onset of symptoms to hospitalisation does not usually
come under psychiatric supervision our information about this period is necessarily retrospective & meagre,"
& this is a problem that remains to this day for all such studies.
There is an encouraging &
interesting study by Olin et al (1998). Here, teachers were able to correctly anticipate 35% of students
who developed schizophrenia, (chosen from the ranks of those deemed at high risk, by dint of early aetiological
factors such as family history, birth complications, maternal flu, time in care, etc.) Also those identified
by the teachers as showing markedly deviant behaviour had poorer clinical & psychiatric outcomes 10 &
25 years later. Of those in the "marked deviant behaviour" group, only 12.9% " were later considered
free of mental disorders, (as compared to 37.6% in the "normal" group, & those identified were generally
less severe.) With 38.7% going on to experience hospital admission. This sounds very useful, if a little
worrying in a Big Brotherish sort of way. Surely if teachers can spot such early warning signs with apparent
ease, should not trained mental health professionals be able to do a much better job of prediction? Yet,
when you consider this more closely, one must not forget that a hit rate of 35% is still fairly low when
considering intervening in these cases. Also, given the high risk status of the studied population, even
random selection would probably get a 'hit rate' in double figures. When taking into consideration the
percentage of the general population with "psychological disturbance" the waters are muddied still further,
especially when account is taken of the distressing events that have led to their being accorded "high
risk" status.
Yung & McGorry (1997) in their survey of this field & relating in particular to
their experiences establishing a P.A.C.E. Clinic (Personal Assistance & Crisis Evaluation) in Australia,
state "The literature ..... Indicates that prodromal psychopathology is extremely diverse & includes
many non-specific symptoms such as depression & anxiety." The oft used research tool of D.S.M.-III-R
(American Psychiatric Association, 1987) cites the following as prodromal symptoms of schizophrenia:
1. Marked social isolation or withdrawal
2. Marked impairment in role functioning as wage
earner, student or home maker
3. Markedly peculiar behavior e.g.. Collecting garbage, talking to
self in public, hoarding food
4. Marked impairment in personal hygiene & grooming
5. Blunted
or inappropriate affect
6.. Digressive, vague, over elaborate, or circumstantial speech or poverty
of speech, or poverty of content of speech
7. Odd beliefs or magical thinking, influencing behavior
& inconsistent with cultural norms, e.g.. Superstitiousness, belief in clairvoyance, telepathy, 'sixth
sense', 'others can feel my feelings'', overvalued ideas, ideas of reference
8. Unusual perceptual
experiences, e.g.. Recurrent illusions, sensing the presence of a force or person not actually present
9. Marked lack of initiative, interests or energy
While useful, these criteria are sadly somewhat
vague & open to interpretation. This is highlighted in a study by McGorry et al (1995). Here the prevalence
of these symptoms was assessed in a large (2525) cohort of high school students. It was found that individual
symptoms were highly prevalent in the sample & the prevalence of DSM-III-R prodromes ranged from 10 -
15% to 50%. For example symptoms of:
7. Magical Ideation experienced by 51%
of the sample
8. Unusual Perceptual Experiences 45.6%
2. Impaired Role Function
41%
9. Anergia
39%
The authors admit to methodological weaknesses but despite this it clearly suggests
that classic prodromal features are extremely prevalent in adolescents & because of this, unlikely to
be specific enough to predict incipient psychosis. Their conclusion was quite naturally that more accurate
predictors than DSM-III-R criteria were needed.
This rather changes the nomenclature. With such
a large potential pool of prospectively ill people, one could not honestly call such signs & symptoms
a "prodrome" or "early warning signs" when most of them will probably not go on to become psychotic.
Yung & McGorry (1997) therefore instead proposed the term "Precursor syndrome", or "At risk mental
state" which makes this difference clear.
Both these & other researchers have attempted to refine
this prediction by adding other predictive factors to enhance accuracy. These have included family history
of psychosis & age. Yung & McGorry (1997) also stated that they were investigating the possibility of
using other early warning signs as predictors, such as attentional & other cognitive deficits, neurological
soft signs & structural brain abnormalities. However, despite their best attempts they were only able
to find rates of onset of psychosis of 21% - 40%..
Following this up in a later paper McGorry
et al (2000), aswell as noting the limitations of D.S.M.-III-R as a criteria, & its specificity to schizophrenia,
also notes (as others have found), the consequent poor inter-rater & test-re test reliability. Could
this be why DSM-III-R's replacement - DSM-IV (American Psychiatric Association, 1994) chose not to record
specific prodromal criteria for schizophrenia?
In their sample of 200 young people who had had
a first episode psychosis McGorry et al's (2000) study tries to overcome these difficulties by using
some fancy statistical footwork (Receiver Operating Characteristic [ROC] & Quality Receiver Operating
Characteristic [QROC] ) whose main function seems to be to refine recorded data, using the DSM-III-R
criteria, & increase the measure's specificity, while reducing potential false positives. All were admitted
as inpatients (some potential bias here perhaps if those not ill enough for admission are excluded from
the sample) & people with intellectual disability, pervasive developmental disorder, organic disorder
[Hmmmm,
some potential debate here too I think, what of those who believe there is an organic or
genetic component to psychosis], poor English language or were not aged between 14 & 46 were excluded.
Diagnosis was made (initially from a computer driven algorithm, with secondary validation by a psychiatrist
or psychologist) using the Royal Park Multi-diagnostic Instrument for Psychosis (RPMIP)}{\fs24 }{\fs24
(McGorry et al 1990a & b) which they state has been found to be valid & reliable & uses multiple sources
of information. The diagnoses generated were as follows:
Schizophrenia
30.5%
Schizophreniform
24 24.5%
Schizo-affective disorder 10%
Delusional disorder 24 6.5%
Bi-polar affective
disorder with psychotic features 13.5%
Depression with psychotic features
24 8%
Brief reactive psychosis 0.5%
Induced psychosis
0.5%
Psychotic disorder NOS
6%
Data was collected from these people & also from an informant.
This data was also used to reconstruct the period of illness prior to hospital admission. Again, good
attention seems to have been paid to strict symptom definition, & inter-rater reliability. They defined
prodrome onset as "the point at which the first perceptible change from pre morbid functioning was detectable,"
& ending as "the point at which frank psychotic symptoms emerged".
There then followed some complex
statistical & procedural manoeuvring to calculate numbers of true & false positives & negatives, standard
error, & significance among other things. This was highly complex, & although used separately by others
in the past, seems to have all been used together in an attempt to be thorough by McGorry et al..
The results showed that using the DSM-III-R prodromal
variables alone, found the presence of more
than 3 items was most efficient at predicting schizophrenia (not terribly helpful when we recall the
results of the previously discussed high school study). However, using a combination of DSM-III-R variables,
prodrome duration & pre morbid functioning was found to be more efficient, (the presence of more than
6 of these variables being predictive of schizophrenia.). Using a technique with conjunctive decision-making,
(DSM-III-R prodromal symptoms for 185+ days, with social isolation/withdrawl or poor pre-morbid social
adjustment/work history & poor pre morbid adjustment, total P.A.S. Score 51.60+) was more effective
aswell, although they found that the duration variables were of most importance & using these alone was
similarly predictive & obviously simpler.
Overall, the team were able to (re)prove that DSM-III-R
criteria have only modest predictive power, that duration of pre-psychotic reduced functioning is a better
predictor, especially when defining the prodrome with DSM-III-R criteria.
The team's approach
was exacting & impressive, but despite this they were still not able to find any clear predictive "cut
point" that would enable clear prediction of developing schizophrenia. This was despite the fact that
all their subjects were "true positives" in that they had all become ill. They did to a small extent
"improve the diagnostic efficiency & predictive power of prodromal symptoms of schizophrenia", but probably
not in a meaningful, clinically useful way, beyond what is already generally known.
Limitations
of this Study
As expressed above, a problem in this type of study is that by necessity it is
retrospective on people who have already been diagnosed with psychosis. The authors admit therefore,
a major problem is that there is no examination of people who did not go on to become ill, potentially
meeting the same criteria as those studied that did. This reduces any potential pre-illness prediction
further still. It would seem that these predictive factors can be no more than general "early warning
signs", rather than diagnostic criteria calling for intervention. This is therefore quite different to
the establishment, monitoring & use of early warning signs in people who have already developed psychosis,
where such signs (as evinced by the Early Signs Scale, (Birchwood, 1992) & development of individualised
relapse signatures (Birchwood et al, 1998)), can be a signal for prophylactic medical & psycho social
interventions.
McGorry et al (2000) agree that such predictive factors as they found can not
be taken to the general population, especially given the fairly low incidence of schizophrenia as a whole.
It would therefore be a grossly impractical procedure.
Clinical experience also suggests that
trying to define a clear, time limited prodrome may be a flawed idea. Many clients I see & have seen
have demonstrated differences from their peers from an early age, common factors being time in special
schools as a child, intra-familial difficulties & stressors, & abuse. This would also be in line with
the widely accepted stress vulnerability model (Nuechterlein & Dawson, 1984), & would suggest that there
are very many other variables in play.
An alternative suggested is of locating & investigating
high risk groups, then using DSM-III-R & duration criteria as predictors. Such groups could include those
with family history of the illness "& ultimately biological markers." (McGorry et al, 2000). It is suggested
that some people have a biological lesion from birth, which predisposes them to psychosis & are therefore
"Doomed from the womb" This approach is currently being investigated by E. Johnstone in the Edinburgh
High Risk Study [not yet published] where preliminary results are that there is an approximately 50%
transfer to psychotic illness in people with 2 first degree relatives with psychosis + prodromal symptoms.
However, even this comparatively high figure still leaves many potential "false positives" & is therefore
probably more useful as a rule for heightened suspicion rather than direct treatment. Byrne et al (1999)
have also shown in preliminary results of the controlled Edinburgh High Risk Study, that high risk individuals
(based on family history) performed more poorly than controls on all tests of intellectual function &
on aspects of executive function & memory. But to suggest that we therefore give up, go home & do nothing
is not on either & even more extreme.
Ethics
This also leads us into the ethical mine
field. Even if the predictive efficacy can be further improved, there are still likely to be false positives,
probably many. If we then propose to intervene in some way, (otherwise what is the point of trying to
identify those at risk of developing psychosis?) We may create more problems than we solve. On the other
hand it is clearly recognised that reducing duration of untreated psychosis (DUP) is an important prognostic
factor & could have significant benefit on progress & outcome & taking this approach maximises any chance
of "catching them early".
These issues are addressed well in Yung & McGorry's
earlier (1997)
study. Despite some progress, the stigma of mental illness is still pervasive & disabling. If we, by
offering treatment, label people as a (potential) psychiatric case, we allow others (family, friends,
employer, society) to also so label them, with all the myriad drawbacks & serious social disability that
this entails. Other problem areas could be ignored following such a labelling. Most clinicians will recall
the attempts of families of young patients, confusing illness with 'normal' (!!) adolescent behaviour.
Many will also recall the unwillingness of some General Practitioners to seriously consider physical
health problems in people with mental illness. We should also remember the continuing influence of life
& health insurance underwriters & the potential negative effects on practical life. We could increase
the stressors for a person potentially to the extent of making a breakdown more likely.
Not
only this, but what of the person's own internal labelling & self stigma. Poor self esteem is well recognised
as a huge problem in psychosis. The creation of unnecessary fear & distress is likely. We all know of
the frequency of secondary morbidity in psychosis, so perhaps even to the extent of production of secondary
morbidity without the primary illness! This is without even considering the other ethical dilemmas of
prescribing prophylactic neuroleptic
medication. Such medication is potentially damaging & disabling
in its own right, & not licensed for such use anyway at present. With a preponderance of potential "false
positives" it would also be hard to test & distinguish who had had illness prevented by treatment & who
would not have got ill anyway.
It seems that to go too far down this road is likely to be excessively
damaging & paternalistic, unless we can refine the paradigm to a greater extent than is likely to be
possible at present. There are of course also the resource implications. Even basic screening of high
risk groups would be costly, & given the standard non-attendance rates of approx. 50% at psychiatric
appointments & the traditional engagement difficulties of this client group it would be hard to implement
& harder to justify. The assertive outreach approach has its problems regarding ethics & civil liberty
as it is without consideration of outreach to those who are still (relatively) well .
Intervention
Strategies
Yung & McGorry (1997) take a practical view. In their work at the Personal Assistance
& Crisis Evaluation (PACE) Clinic, which aims to identify & treat prodromal individuals before illness
onset, interventions are targeted at "help seeking symptomatic individuals .... Manifest[ing] significant
levels of symptomatology & disability." Therefore, they are in most cases young people with particular
difficulties wanting treatment. They also state that for such referred patients, transfer rates to active
psychosis in a short time are high (though they do not reveal how high). As regards communication of
increased risk state to patients, many, particularly those with ill parents, & people with Brief Limited
Intermittent Psychosis (BLIPS) are already aware of this risk & are anxious to discuss it. Therefore,
staff will discuss this with them, while also trying to offer possible solutions, & communicating the
message that with intervention & monitoring of symptoms it may be possible to avert or ameliorate development
of psychosis. The attempt to instil the sense of hope & control that is the common denominator in much
of the early psychosis field.
It seems likely that few would disagree when they talk of offering
psychological & social treatments to such a client group, especially when such symptoms have been long
lasting (in their study average time from symptom onset to help seeking was 86.6 weeks, & time from first
help seeking to attendance at PACE Clinic was 41.4 weeks). So, bearing in mind the earlier discussions
on duration, symptoms are of fairly long-standing, & even if the person is a "false positive" is likely
to be deserving of help. However, they state that they do not use neuroleptic medication in the absence
of psychosis, although Faloon (1992, cited in Yung & McGorry, 1997) has explored this & found low dose
neuroleptics of benefit for people with prodromal symptoms. (There are few extant examples of this approach,
however, which reflects its difficulties.) Yet, it is an area that, in these circumstances is probably
ripe for further investigation, particularly considering the penchant of a few GPs for reactive prescription
of neuroleptics for anxiety, insomnia, agitation et al..
Treatments likely to be of most benefit
are those suited to presenting symptoms (& likely to be suited to the general population through health
education too - not to mention many mental health professionals!.). For example stress management, substance
misuse support & education, problem solving skills & other coping strategies. The need to improve access
to help, which is always related to duration of untreated psychosis in studies of pathways to care (there
are often multiple failed attempts to access help) is another consideration around treatment requiring
thought. The area of social skills training & support is also of particular relevance.
Conclusion
It is the stress vulnerability model that gives us probably the best road map for navigating this
difficult course. From what we have seen, the client group is too heterogeneous & the risk factors too
myriad & diverse to create a clear predictive tool that will enable us to identify usable early warning
signs. Schizophrenia is still more a syndrome than an illness for this to work in any watertight way,
& therefore prediction & treatment is likely to remain more an art (albeit one with good rules & guidelines)
than a science, relying on interpretation & practice by skilled & trained clinicians.
This is
however by no means a call for inaction. Practicable strategies aimed at raising public awareness to
reduce duration of untreated illness as demonstrated impressively in Norway (Larsen et al, 2000) - where
a public education campaign in one area, together with mobile community treatment teams was able to reduce
DUP from an average 20 months to 3 months with consequences on speed of recovery & rate of relapse -
will only help those people at risk. (Unfortunately this study also falls into the false positive trap
with the risk of pulling people on the edges into treatment & potentially boosting results artificially.)
Not only this, but such activities can only help to reduce stigma & fear.
Implications for Clinical
Practice
1. Maintain high level of suspicion of potential psychosis, particularly where DSM-III-R
prodrome criteria are met & are of long duration. This should be communicated to primary health care
team staff. Possible need for referral guidelines encouraging this low threshold.
2. Health promotion
about mental health service, stress & coping strategies
3. The importance of raising level of
public awareness education on diagnosis & treatment of psychosis, together with need to reduce duration
of untreated illness.
4. That contact (especially first contact) with mental health services should
be easy & minimally stigmatising. (Attention to "normal" setting & to team ethos/atmosphere.) Youth &
user focus likely to be of benefit
5. Monitor pathways to care to seek improvement in access
6. Flexibility of service response. It doesn't have to be all or nothing. Ambiguous cases can often be
monitored without difficulty.
7. Collaborative work with other agencies.
(Based on the
IRIS guidelines)
REFERENCES:
American Psychiatric Association, (1987) Diagnostic
& statistical manual of mental disorders. 3rd revised edition
American Psychiatric Association,
(1994) Diagnostic & statistical manual of mental disorders. 4th edition
Birchwood, M, (1992)
Early Intervention in schizophrenia: theoretical background & clinical strategies. British Journal
of Clinical Psychology. 1992 31, 257-278
Birchwood, M, Todd, P, Jackson, C. (1998) Early intervention
in psychosis. The critical period hypothesis. British journal of psychiatry, vol. 172 supplement 33
Birchwood, M., & Iqbal, Z (1998) Depression & suicidal thinking in psychosis: a cognitive approach
in outcome & innovation. In Wykes, T., Tarrier, N., & Lewis, S. (eds.), Psychological treatment of
schizophrenia.
Birchwood, M., (1998) Quoted in Jackson, C., Catching them young,, Mental Health
Care, 1998, 1, 9
Byrne, M., Hodges, A., Grant, E., Owens D.C., & Johnstone, E.C. (1999) Neuropsychological
assessment of young people at high genetic risk for developing schizophrenia compared with controls:
preliminary findings of the Edinburgh High Risk Study Psychological Medicine 1999 Sept; 29(5): 1161-73
Falloon, I.R.H., (1992) Early intervention for first episodes of schizophrenia: a preliminary exploration.
Psychiatry 1992, 55; 4-15 (cited in Yung & McGorry, 1997)
Larsen, T.K., Johannessen, J.O., McGlashan,
T., Horneland,M., Mardal, S., & Vaglum, P. (2000) Can duration of untreated psychosis be reduced?
In Birchwood,M., Fowler, D., & Jackson, C. (eds.) Early intervention in psychosis, a guide to concepts,
evidence & interventions , John Wiley & sons, Chichester
McGorry, P.D., Copolov, D.L., Singh,
B.S., (1990a) The Royal Park multi-diagnostic instrument for psychosis: part I: rationale & review.
Schizophrenia Bulletin 16, 501-515}
McGorry, P.D., Singh, B.S., Copolov, D.L., Kaplan, I., Dossetor,
C.R., Van Riel, R.J., (1990b) The Royal Park }{\fs24 multi-diagnostic instrument for psychosis:
part II: development, reliability & validity. Schizophrenia Bulletin, 16, 517-536
McGorry, P.D.,
McFarlane, C., Patton, G.C., Bell, R., Hibbert, M.E., Jackson, H.J., Bowes, G. (1995) The prevalence
of }{\fs24 prodromal features of schizophrenia in adolescence: a preliminary survey. Acta Psychiatrica
Scandinavica Oct. 1995, 92(4):241-9
McGorry, P.D., McKenzie, D., Jackson, H.J., Waddell, F.,
& Curry, C. (2000) Can we improve the diagnostic efficiency & predictive power of prodromal symptoms
for schizophrenia? Schizophrenia Research 42 (2000) 91-100}
Nuechterlein, K.H., & Dawson,
M.E. (1984) A heuristic vulnerability/stress model of schizophrenic episodes. Schizophrenia Bulletin
1984 10, 2, 300-312
Olin, S.S., Mednick, S.A., Cannon, T., Jacobsen, B., Parnas, J., Schulsinger,
F, & Schulsinger, H. (1998) School teacher ratings predictive of psychiatric outcome 25 years later.
British Journal of Psychiatry (1998) 172 }
(suppl. 33} 7-13
Shepherd, M, Watt, D, Falloon,
I & Smeeton, N (1989) The natural history of schizophrenia: a five-year follow up in a representative
sample of schizophrenics. Psychological medicine, Monograph Supplement 15.
Wiersma, D., Nienhaus,
F..J., Sloof C., & Giel, R. (1998) Natural course of schizophrenic disorders: a 15 year follow-up
of a Dutch incidence cohort. Schizophrenia Bulletin, 24, 78-85
Yung, A.R., McGorry, P.D (1997)
Is pre-psychotic intervention realistic in schizophrenia & related disorders? Australian & New Zealand
Journal of Psychiatry 1997; 31: 799-805}
|
|