Psychiatrists ' knowledge of drug induced psychosis

and vague. While the existence of an ampheta mine psychosis (paranoid and schizophreniform symptoms arising in the context of amphetamine intoxication) has been recognised for over 40 years, the role of other substances such as cannabis in the genesis of psychotic symptoms remains controversial (Thomicroft, 1990; Mathers & Ghodse, 1992; Thomas, 1993). Terms such as 'toxic psychosis', 'induced schizophreniform psychosis' and 'confusional paranoid psychosis' are used with similar but

using a specifically designed questionnaire. Knowledge about the effects of drugs of abuse and their relationship with psychosis was also examined. A substantial proportion of clinicians did not have a logical and coherent scheme for understanding 'drug-induced psychosis' and there was no consensus as to which reactions to drug taking should be included within the rubric. Diagnostic errors have potentially serious consequences, so lack of knowledge in this area needs attention and a proposed classification is given.
With the rapid increase in the use of illegal drugs among the general population, the differential diagnosis of psychotic states commonly includes 'drug induced psychosis'. This term is imprecise and vague. While the existence of an ampheta mine psychosis (paranoid and schizophreniform symptoms arising in the context of amphetamine intoxication) has been recognised for over 40 years, the role of other substances such as cannabis in the genesis of psychotic symptoms remains controversial (Thomicroft, 1990;Mathers & Ghodse, 1992;Thomas, 1993).
Terms such as 'toxic psychosis', 'induced schizophreniform psychosis' and 'confusional paranoid psychosis' are used with similar but idiosyncratic dÃ©finitions (Post, 1975;Gough & Cookson, 1984;Deveaugh-Geiss & Pandurangl, 1992). We examined the relationship between psychiatric states and substance abuse, and suggested a classification of drug induced psy chosis (Table 1) (Poole & Brabbins, 1996). The major textbooks of psychiatry give scant attention to the subject, and ICD-10 (World Health Organization, 1992) classifies these states in large, ill-defined and confusing sub-categories. In the absence of a widely accepted classification of the relationship between drug intoxication and psychosis, we have attempted to examine clin icians' dÃ©finitionsand understanding of these states as well as their knowledge of the effects of drugs of abuse in general. This study examined whether clinicians' views were internally consis tent and logical, and whether consensus beliefs could be identified.

The study
All doctors working in psychiatric services within three urban health districts were contacted by letter. They were asked to complete an anon ymous questionnaire about their age, sub-speci ality, grade and length of general adult psychiatric experience. They were asked to estimate what proportion of patients currently under their care used illegal drugs, and to indicate their assessment of their own knowledge of drug taking behaviour. An unstructured section asked for a description of their under standing of the nature and definition of 'drug induced psychosis'. This was followed by a structured section giving five definitions of drugrelated abnormal mental states; respondents were asked to indicate which they would include within their understanding of 'drug induced psychosis'. The five abnormal mental states were: 1) acute intoxication with delirium or perceptual distortions; 2) withdrawal state with delirium; 3) psychotic disorder due to substance use; 4) acute intoxication resembling psychosis; 5) relapse of pre-existing psychotic illness. The first three were drawn from ICD-10. The fourth was included as a less ambiguous definition than the first. The fifth was included for completeness. Detailed definitions were included for clarity.
There was a further structured section, in the form of a grid, with clinical states 1 to 5, as above, on one axis and a list of drugs on the other. The drugs included were opiates, cannabis, barbitu rates, benzodlazepines, ecstasy, LSD, crack/ cocaine, amphetamine, butane, alcohol and phencyclidine hydrochloride (PCP). The last was included as its use is virtually unknown In the UK. Respondents were also asked to indicate whether they had clinical experience of patients using each of the substances and whether each type of reaction could occur with each substance.

Findings
Of 118 questionnaires despatched 66 were returned (55%). There was no significant differ ence in response rate between different grades of doctor. Although 47% of respondents indicated that 'some', 'many' or 'most' of their patients took drugs (53% indicated 'few' or 'no' patients), only 29% felt that their knowledge of drug taking behaviour was 'good' or 'extensive' (71% re sponded 'moderate', 'slight' or 'no' knowledge), liiere was no significant difference in these responses between different grades of doctor, by date of qualification or length of adult psychiatry experience.
In the unstructured section requiring a defini tion of drug induced psychosis, 27% of responses were internally logical and narrowly defined so as to exclude other disorders (e.g. pre-existing functional psychosis); 17% were internally logical but overinclusive, explicitly overlapping with other psychotic disorders or embracing non-psychotic disorders; 56% were not internally logical (e.g. tautological responses such as "a drug Induced psychosis is a psychosis induced by drugs").
In the structured section, withdrawal states and relapse of an existing functional psychosis caused by drug use were regarded as being self-evldenuy outside any reasonable dÃ©finition of a drug Induced psychosis. One or both of these was Included within the definition by 32%. There was no predominant pattern of responses, and virtually all combinations of different states appeared.
Within the 'clinical state by drug' grid, some responses were regarded as 'certain', at least in terms of the existing literature. Examples In cluded "amphetamines cause psychotic symp toms in the context of intoxication", "intoxication with opiates does not resemble psychosis". Other responses were regarded as 'uncertain', which is to say either absent from the existing literature, or subject to controversy, for example "cannabis can provoke relapse in existing functional psy chosis". The numbers of 'certain' responses answered correctly, incorrectly or as not known, were analysed by grade of doctor. Consultants, senior registrars and registrars answered cor rectly in approximately 75% of 'certain' re sponses, while senior house officers and nontraining grades fared less well. However, there was considerable variation within different re sponses. Some 'certain' cells were answered correctly by most respondents, for example "acute intoxication with delirium or perceptual distortions occurs with LSD", and "acute intox ication with amphetamines can resemble psycho sis". Other 'certain' cells provided a poorer rate of correct responses, for example only 39% of respondents recognised that acute Intoxication with butane can resemble psychosis.
A consensus view (arbitrarily defined as more than 60% of respondents giving a particular answer) was identified In only a handful of 'uncertain' cells. These were: an opiate with drawal state with delirium; cannabis Induces relapse of a pre-existing psychotic illness; benzodiazeplnes do not cause relapse of a pre-existing psychotic illness; acute Intoxication with delirium or perceptual distortions occurs with ecstasy; acute intoxication with ecstasy resembles psy chosis; LSD causes a psychotic disorder; LSD causes relapse of an existing psychotic illness; alcohol causes psychotic disorder; alcohol causes acute intoxication with delirium or perceptual distortions. With regard to PCP, in three cells, a third of respondents felt confident enough to make a definite response.
The 'no clinical experience' cells showed marked differences between substances, all doc tors having clinical experience of benzodiazepine or alcohol abuse, whereas 80% had no clinical experience of PCP abuse.

Comment
The methodology can be criticised. In common with most questionnaire studies, the response rate was relatively low and the degree of thought put into responses probably varied widely. Re sponses regarding PCP suggested that guess work played a part In a substantial proportion of responses. The instrument was a relatively crude way of examining clinicians' beliefs about a psychiatric syndrome though the inclusion of both structured and unstructured sections to some extent protected the data from instrument artefacts. The design of the study was not suitable for statistical analysis of individual items.
Despite these caveats, the results suggest that a substantial proportion of clinicians did not have a logical and coherent scheme for understanding 'drug induced psychosis' and other adverse psychiatric reactions to drug taking. Furthermore there was no consensus as to which reactions to drug taking should be included in the diagnosis 'drug induced psychosis'. Psychiatrists' know ledge of drug taking behaviour and the effects of drugs was poor by their own estimation. On objective testing, some substances were well understood, others less so. These findings applied equally to experienced and inexperienced clin icians.
Clinical beliefs are known to be illogical in some fundamental areas (Harper, 1994). The use of tautological definitions of 'drug induced psycho sis' is worrying as it would tend to lead to a diagnosis of 'drug induced psychosis' in prefer ence to functional psychosis when drug taking and psychosis coincide. We have pointed out elsewhere that a causal link is by no means certain in all circumstances (Poole & Brabblns, 1994). The diagnostic error of mis-attribution easily arises, and may be relatively common. Elsewhere we have described the difficulties in evaluating the scientific evidence in this area and have proposed a classification of these states (Table 1) (Poole & Brabbins, 1996). drugs causing schizophreniform symptoms in affective disorder) 3. Chronic hallucinosis induced by substance abuse (e.g. alcoholic hallucinosis and LSD flashbacks) 4. Drug induced relapse of functional psychosis (drugs precipitating relapse of an existing illness. They may also exacerbate symptoms in an ongoing illness) 5. Withdrawal states (e.g. delirium tremens with alcohol and a similar state with benzodiazepines and barbiturates) 6. Other reactions: a) Intoxications with clouding of consciousness (e.g. barbiturates causing delirium) b) Post-intoxication depression (e.g. following stimulants) c) Panic attacks (e.g. with hallucinogens and stimulants) Psychiatric training emphasises the treatment of opiate dependence. It would appear that nondependent use of drugs is an area of ignorance among many clinicians which needs to be cor rected both within training schemes and continu ing professional development programmes.