Are psychiatrists sexist ? A study of bias in the assessment of psychiatric emergencies

In order to investigate bias in history taking among psychiatric trainees, a retrospective study of casenotes was undertaken in an emergency psychiatric clinic in a teaching district. Two hundred and twentyseven consecutive new patient assessments were assessed for quality of alcohol, substance use and forensic histories. Trainees were more likely to take alcohol, substance use and forensic histories from men, and more likely to take substance use histories from younger patients. It is concluded that trainees make sexist and ageist assumptions when they assess patients. Thereis a need for the education of doctors in this area.

It is well known that the incidence of certain psychiatric disorders varies according to gender (Oppenheimer, 1991), while more recently it has become apparent that racial differences in the incidence of psychiatric syndromes exist (Harrison et al 1988).Also risk factors for psychiatric illnesses are subject to variations in incidence according to gender and ethnic group.For example, men as a group drink more alcohol than women (Dight, 1976), and certain types of drug use are more common in particular ethnic groups (Lillie-Blanton et al 1993).
Does such knowledge about populations lead us to make assumptions when we assess individuals?Lewis et al (1990) investigated this theoretically by asking consultant psychiatrists to consider a case vignette, altering the sex and race of the 'patient' between clinicians.They found that the race and the sex of the 'patient' influenced clinical predictions and attitudes of the clinicians towards the 'patient'.
In a more clinically orientated study, Redman et al (1991) found that Australian general practitioners and interns rate women more psychologically disturbed than men given an equivalent level of disturbance as measured by the General Health Questionnaire (GHQ).More women than men were rated as psychologically disturbed when they were non disturbed according to the GHQ.
We undertook a retrospective case-notes study to investigate whether trainee psychiatrists make such assumptions about individuals of a particular ethnic group, gender, or particular age group in clinical practice.The setting chosen was the emergency assessment of new patients.The variables were the recording of substance use.alcohol and forensic histories, chosen because of their relevance to emergency psychiatry and because assumptions may be made about alcohol and substance use for the reasons described above.

The study
We assessed retrospectively all the available case notes of new patients seen in the Hackney Hospital Emergency Clinic (EC) in the first 165 days of 1991 and the same period in 1992.The EC provides an emergency psychiatric service for the City and Hackney Health District on weekdays between 9 am and 5pm.Referral are accepted from any source including self referral.Patients are assessed by senior house officers (SHOs) or registrars in psychiatry.Table 1 shows the information recorded from each set of case notes.
We compared whether the presence or absence of alcohol, substance use and forensic histories varied significantly according to the age, gender or recorded ethnic group of the patient, using x2 tests of association.Data analysis was performed using a SMART spreadsheet and standard statistical tables.

Quality of history taking
Of the new patient assessments studied, an alcohol history was recorded in 150 cases (66%), but in 33 (15% of total) it was nonquantitative.Evidence of other substance use was noted in 130 histories (57%), but was nonquantitative in 26 (11% of total).Forensic history was present in only 56 cases (25%).

Source of bias
Patient gender.Alcohol, substance use and forensic histories were all significantly more likely to be recorded from men than women (Table 2).Patient ethnic group.No significant association was found between the patient's ethnic origin and the recording of either substance use, alcohol or forensic histories.
Patient age.The sample was divided into two groups, one younger and one older than the median age of 29.The recording of alcohol and forensic histories did not differ significantly between the two groups, but a substance use history was recorded in a significantly greater proportion of the younger than the older group (Table 2).

Comment
Our findings suggest that in the emergency assessment of new psychiatric patients, clinicians are biased as regards the gender of the patient, perhaps assuming that women drink less, take fewer drugs, or commit less crimes.They imply that the sexist assumptions found under experimental conditions by Lewis et al (1990) also operate in clinical practice.They further imply that there may be ageist assumptions about drug taking habits.
History taking did not appear to be biased according to ethnic origin.However, in view of the lack of ethnic data for 31% of the sample, this particular finding should be interpreted with caution.
Patterns of substance and alcohol use are different in men and women, and in people of different ages.However, extrapolating from the known behaviour of different populations invites stereotyped assumptions about an individual which may be false.This in turn can lead to diagnostic errors, for example, missing drug induced psychoses or depression secondary to alcohol use.The present study suggests that such an error is more likely to happen in women and those over the age of 29, because of the assumptions of their doctors.This is of particular concern in the light of statistics from the Office of Population Censuses and Surveys which suggest that Bios in the assessment of psychiatric emergencies 8% of women drink more than the recommended maximum of 14 units of alcohol per week, and a further 1% drink more than 35 units per week, a theoretically dangerous level (Goddard & Ikin, 1988).It has been said that "deeply ingrained sexism characterises much psychiatric practice" (Allen, 1986).The present study suggests that both sexist and ageist stereotypes affect clinical practice and may undermine the management of female and older patients.
Given the importance of substance criminal activity and alcohol in emergency psychiatric practice, how do doctors practising in this field compare in their history taking skills with doctors practising in other areas?We know of no previous surveys of the taking of substance use or forensic histories.However, there have been surveys of alcohol history taking by housemen, mostly of routine surgical patients (Barrison et al 1980), by psychiatric registrars admitting patients to hospital (Parrel & David, 1988), and by doctors admitting the elderly with an acute medical condition (Naik & Jones, 1994).The results of these studies are compared with the present study in Table 3, which suggests that patients presenting with psychiatric emergencies are indeed the most likely to have a quantitative alcohol history recorded, with elderly medical cases the least likely (again suggesting ageist assumptions).It is also striking how frequently, an alcohol history is unrecorded in all groups.
This failure of doctors to record important historical information is a cause for considerable concern.Sustained educational efforts are still required to increase the awareness of medical staff to the importance of substance use and alcohol histories in all patient groups, in order to overcome sexist and ageist clinical assumptions as demonstrated in our study.

Table 1 .
Items recorded from each set of case notes (29%) as black (includes 'Afro-Caribbean', 'African' and 'West Indian'), and 10 (6%) as Asian.Twelve (8%) were in other groups (includes 'Greek', Turkish' and 'Kurdish').The ethnic groupings are based on those used by the clinicians in the case notes and are not directly comparable with Office of Population Censuses and Surveys (OPCS) categories.

Table 2 .
Comparison of history taking for gender and age: significant findings

Table 3 .
Comparison of studies of alcohol history takingAlcohol history-number where recorded (%)