History of Alcoholism and the Justice System

British Journal of Addiction (1990) 85, 883-890

RESEARCH REPORT

Discrimination on the grounds of diagnosis

MICHAEL FARRELL, MB, MRCP, MRCPsych,' & GLYN LEWIS, MA, MSc, MRCPsych,2

^Addiction Research Unit & ^General Practice Research Unit, Institute of Psychiatry, De Crespigny Park, London SE5 8AF, United Kingdom

Summary The study used the methodology of randomly allocating case vignettes to a sample of British consultant psychiatrists to assess the influence of a past diagnosis of alcohol dependence on present treatment attitudes. The case vignettes either did or did not include the previous diagnosis of alcohol dependence and the sex of the 'case' was also randomized. Psychiatrists receiving the vignette with the diagnosis of alcohol dependence were more likely to rate the patient as difficult, annoying, less in need of admission, uncomplaint, having a poor prognosis and more likely to be discharged from follow-up. There was minimal sex difference. Psychiatrist with a special interest in addictions regarded people with a past diagnosis of alcohol dependence as less difficult to manage than their non-specialist colleagues. The implications for education and treatment are discussed.

Introduction The high rates of alcohol-related problems among the medical profession might suggest that they would adopt a sympathetic approach to other individuals likewise afflicted. However, most com- mentators have remarked that physicians usually regard the alcoholic patient as an undesirable client.

Much of the older research now appears weak- ened by its preoccupation with the 'disease model' of alcoholism, lack of attempts to elicit attitudes to control subjects and the use of vague semantic differentials. For instance. Fisher et al. (1975) asked subjects to "rate alcoholic and average persons" on 16 semantic differentials including items such as 'passive-active' and 'delicate-rugged'. Their results indicated that doctors regarded alcoholics as more 'hopeless' than average persons, but this is not a surprising result. MacDonald & Patel (1975) did compare psychiatrists attitudes to alcoholics with

Correspondence to: Michael Farrell, Addiction Research Unit, 101 Denmark Hill, London SE5 8AF, United Kingdom.

Other diagnoses but used a single 'favourable- unfavourable' continuum, so illustrating that psy- chiatrists had more 'favourable' views about neu- roses than alcoholism. It is still possible that the views expressed in these studies could have been confounded, for instance, by the sex or age of the stereotypical patient which was not specified in their design.

One theme of the research on the topic has been the influence of proper training and education on attitudes towards treatment of the alcoholic (Mogar et al., 1969; Geller et al., 1989). For instance, Cartwright (1980) found that experience of working with alcoholics and clinical supervision were associ- ated with positive therapeutic attitudes. This find- ing is slightly at odds with the observation that newly qualified doctors have less hopeful attitudes than medical students (Fisher et al., 1975; Geller et al., 1989) though this may be due to the rather theoretical and idealistic attitudes of medical students to dealing with all patients, let alone alcoholics.

General psychiatrists receive the bulk of general

883

884 Michael Farrell & Glynn Lewis

practitioner referrals for alcohol problems and the aims of this study were to examine the effect of the past diagnosis of alcohol dependence on the atti- tudes to and proposed management of a depressed and potentially suicidal person. Psychiatrists are also used to dealing with more problematic patients than their medical colleagues and have received more education about alcohol problems. Strong (1989) for instance has argued that alcoholics break the 'rules' which usually govern medical consulta- tions, these rules are also broken by many psychia- tric patients. Do psychiatrists show rejecting atti- tudes to alcoholics when they are compared with other psychiatric patients?

Though one would expect psychiatrists to know more about alcohol problems than other physicians, recent work shows that most registrars in psychiatric hospitals (Farrell & David, 1988; Mitchell, 1989) do not take a drinking history though 10% of psychia- tric hospital admissions may have an 'alcoholism' diagnosis (Glass & Jackson, 1988). In a survey of senior psychiatric registrar training in the UK (Brook, 1974) only half reported having received adequate training in substance misuse.

The study used the methodology of randomly allocating case vignettes to a sample of British consultant psychiatrists. The case vignettes either did or did not include the previous diagnosis of alcohol dependence and the sex of the 'case' was also randomized. For convenience the case vignettes, with the diagnosis of alcohol dependence will be referred to as alcoholics. The vignette and methodo- logy was similar to that employed by Lewis & Appleby (1988) who investigated the previous diagnosis of personality disorder.

interested in how experience mfluenced the practice of psychiatrists and were asked to provide details about previous qualifications and experience in psychiatry and in other specialties. The real purpose of the study was not explained to the subjects.

Case histories The four case histories differed from each other in one particular. The case contained the information which a GP's letter might provide about a depressed patient. The amount of information was deliberately restricted to encourage subjects to draw inferences based on pre-existing attitudes. The first case history was as follows:

A 40-year-old man with two children aged 3 and 8 IS seen in outpatients. He complains of feeling depressed and says he has been crying on his own at home. He is worried about whether he is having a nervous breakdown and is requesting admission. He has thought of killing himself by taking an overdose of some tablets he has at home. He has taken one previous overdose, 2 years ago, and at that time he saw a psychiatrist who gave him a diagnosis of alcohol dependence. He has recently gone into debt and is concerned about how he will repay the money. He is finding it difficult to sleep.

This case was modified slightly in the following ways:

(1) Case 2: no previous diagnosis was mentioned. (2) Case 3: patient changed to female. (3) Case 4: patient female and no diagnosis was

mentioned.

Method Sample Two hundred names of psychiatrists who lived in England, Wales or Scotland were randomly selected from the 1985 membership list of the Royal College of Psychiatrists (approximately 10% of total, DHSS, 1987). Those who were described as registrars, senior registrars who were retired or were listed as being child psychiatrists were excluded from the sample (but several child psychiatrists were in- cluded in the sample because they were not listed as such). Subjects were randomly allocated one of the four brief case histories, which they were asked to read before completing and returning an accom- panying questionnaire. They were told that we were

Questionnaire The subjects were asked their sex, years experi- ence in psychiatry and to indicate any areas of special interest from the following list: neuro- psychiatry, social psychiatry, psychotherapy, child/adolescent, forensic, liaison, old age and addiction. This was used to divide the sample between those who had an interest in addictions and those who did not.

The questionnaire also included 20 items, ar- ranged in a similar manner to semantic differentials, but opposing statements rather than single words were used. Each 6-point scale was designed to elicit one aspect of the assessment or management of the case. Some of the items placed more emphasis on

Discrimination on the grounds of diagnosis 885

Table 1. Comparing the responses to 'cases' with a previous diagnosis of alcohol dependence with those given no previous diagnosis

Admission not indicated Not a suicide risk Antidepressants indicated Trying to manipulate admission Needs sickness certificate Discharge from OP follow-up Unlikely to arouse sympathy Overdose would be attention-seeking Would not like to have in one's clinic Difficult management problem Likely to annoy you Unlikely to improve Cause of debts probably controllable by patient Does not merit NHS time Unlikely to complete course of treatment Does not have mental illness Children should be on 'At Risk' register At risk of becoming dependent on you Unlikely to comply with advice/treatment Patient's condition not severe

Higher scores indicate agreement with statement on left.

Mean No

diagnosis (iV=77)

3.49 3.03 3.07 2.73 2.10 1.57 2.33 3.07 2.05 2.88 1.96 2.00 3.67 2.40 2.77 2.92 2.00 3.49 2.64 3.15

Scores Alcohol

dependence (N=67)

3.21 3.02 2.33 2.91 2.24 1.94 2.86 3.05 2.60 3.52 2.31 2.30 4.05 2.59 3.23 3.21 2.50 3.91 3.20 3.08

95% CI mean difference

-0.20, 0.76 -0.42, 0.44

0.23, 1.25 -0 .20, 0.56 -0 .35 , 0.55 -0.04, 0.78

0.16, 0.92 -0.40, 0.36

0.14, 0.96 0.21, 1.07 0.00, 0.70

-0 .03 , 0.63 0.02, 0.74

-0 .18 , 0.56 0.08, 0.84

-0.17, 0.75 0.07, 0.93

-0 .03 , 0.87 0.23, 0.89

-0.39, 0.25

r test P

0.25 0.96 0.005 0.37 0.54 0.08 0.007 0.92 0.01 0.004 0.05 0.08 0.04 0.32 0.02 0.22 0.03 0.07 0.001 0.68

practical management issues (e.g. antidepressant prescription, admission) but most asked directly about attitudes to the patient (e.g. likely to annoy, attention-seeking etc.). A full list is given in Table 1. The items were balanced so that the more rejecting ends of the items were unsystematically arranged between right and left. The items were scored so that a higher score represented responses that were regarded as more rejecting or that indicated lack of active treatment. For instance, a response at the end of the scale 'overdose would be an attention seeking act' scored 6 and a response at the end 'overdose would be genuine suicidal act' was scored 1.

Each subject was asked to complete the question- naire and then choose a diagnosis from a list of depression, anxiety, adjustment reaction, alcohol dependence, personality disorder and neurasthenia.

Analysis Repeated significance tests have been used and this will increase the likelihood of a type 1 error. The Bonferroni Criterion for this study with 20 compari- sons would be p=0.0025.

It was difficult to present the results for all the 20

items studied here. To aid presentation, on occa- sions, those items which proved to be statistically significantly associated with the previous diagnosis of alcohol dependence at the 5% level were com- bined in to a composite score, the sum of all those items. Items were not weighted. This composite score did not include the item enquiring about tricyclic antidepressant treatment, because this could be regarded as an appropriate management approach to the alcohol-dependent patient. Con- structing the composite score was done purely to illustrate results and did not form the basis of any of the conclusions that were drawn.

Results Sample characteristics One hundred and forty-four of the questionnaires were returned completed (71%) and a farther 17 (8.5%) returned uncompleted because the subjects refused to complete the questionnaire or had retired or moved. The randomization to the four groups was checked using the following variables: years psychiatric experience, sex of respondent, interest in addictions and membership of the Royal College of Physicians. There were no statistically significant differences between the groups at the 5% level.

886 Michael Farrell & Glynn Lewis

The respondents had worked in psychiatry for an average of 20.6 years (SD 7.2), 19% were women, and 16% expressed some interest in the addictions.

Previous diagnosis of alcohol dependence The two groups who were given the vignette that contained the previous diagnosis of alcohol depen-̂ dence were combined and compared with the two groups not given that diagnosis. This analysis therefore ignored the effect of the case's sex. The results of the t tests are shown in Table 1. Some of the distributions of the scores on the items were skewed from a normal distribution. Those variables were also analysed after being transformed by logarithms, but there were no substantial differences so the untransformed results are presented.

In 16 of the 20 items, those in the alcohol dependence group were given higher scores than those with no diagnosis. In particular, 'cases' once given the diagnosis of alcohol dependence were judged unlikely to complete the course of treatment or to comply with advice, would not be liked in the clinic, would not arouse sympathy and would annoy the doctor. Alcoholics debts were under their control, their children were more likely to be considered for the 'at risk' register and management was considered difficult.

Influence of patient's sex A similar analysis was performed by combining the two groups with the 'patient' of the same sex. Two of the results indicated a statistically significant difference at the 5% level, women were less likely to be admitted (male mean=3.10, female mean = 3.62; 95% Cl of difference 0.05, 0.99) and men were regarded as being more in control of their debts (male mean = 4.04, female mean = 3.64; 95% Cl of difference 0.04, 0.76). These results are difficult to interpret in view of the multiple comparisons that were made. It is clear however that the previous diagnosis of alcohol dependence had a more pro- nounced effect on the attitudes measured here than the sex of the patient.

Female alcoholics compared with male alcoholics Were female alcoholics regarded in a less or more favourable light than those of the male sex? This was examined by use of interaction terms in a two- way analysis of variance. Of the 20 items, only 2 showed a statistically significant interaction term

when each item was modelled in turn. These items were 'Does not merit NHS time' (F(l,115) = 4.21, /> = 0.04) and the item 'Not a suicide risk' (F(l,115) = 11.22,/) = 0.001). Examination of the means (Table 2) indicates that male alocholics 'were regarded as meriting less NHS time but this was not so for women alcoholics. Surprisingly, male alcohol- ics appeared to be judged less of a suicide risk while women alcoholics were thought to be at increased risk of suicide (Table 3).

Table 2. Interaction between sex of 'case' and previous diagnosis of alcohol depen-

dence: meriting NHS time

Alcohol No N dependence diagnosis

Male Female

71 72

2.69

2.50 2.21 2.61

Higher scores indicate increased agreement with the statement 'This case does not merit NHS time'.

The use of multiple significance tests in this manner will increase the likelihood of a type 1 error (false positive) and these results are presented to guide future research. A multivariate analysis of variance in which all 20 items were studied did not indicate a statistically significant interaction term between sex of vignette and the label of alcohol dependence.

Table 3. Interaction between sex of 'case' and previous diagnosis of alcohol depen-

dence: suicide risk

Alcohol No N dependence diagnosis

Male Female

71 72

3.22

2.83

2.62

3.44

Higher scores indicate increased agreement with the statement 'Not a suicide risk'.

Special interest in the addictions The hypothesis that psychiatrists who specialize in the addictions would have less critical attitudes was examined in a similar way. There was only one interaction term that was significant 'Difficult management problem' (F( 1,115) = 4.96, /) = 0.03) (Table 4). Alcoholics were regarded as less difficult to manage amongst those indicating an interest in the addictions, while those without a specialist

Discrimination on the grounds of diagnosis 887

interest regarded alcoholics as being more difficult sis. However, examining the individual means

to manage.

Table 4. Interaction between previous diagnosis of alcohol dependence and interest in the addictions:

difficulty of patient

Alcohol No N dependence diagnosis

Interest in addictions

No interest in addictions

117

23

2.57

3.77

3.11

2.87

Higher scores indicate icreased agreement with the statement 'Difficult management problem'

The possibility of an interaction term between interest in the addictions and previous diagnosis of alcohol dependence can be illustrated by computing a composite score, the sum of all the variables in Table 1 which showed a statistically significant (5% level) difference between the alcohol dependence group and the remainder. These results are shown in Fig. 1. Overall there is a suggestion that attitudes to alcoholics are less rejecting in those expressing an interest in the addictions. As mentioned above it is difficult to draw too many conclusions from this analysis.

Diagnosis At the end of the questionnaire the psychiatrists were asked to indicate their own preferred diagnosis (only a single choice was allowed) and the fre- quency of these are given in Table 5. It can be seen that the diagnosis of alcohol dependence was only made in the group given the previous diagnosis of alcohol dependence. In addition, the diagnosis of personality disorder was more likely to be made in the alcohol group. Lewis & Appleby (1988) have already argued that personality disorder is a pejora- tive label. Depression was the preferred diagnosis in the majority of cases.

The effect of previous diagnosis of alcohol dependence could be mediated by the diagnoses made by the psychiatrist. Two-way analyses of variance were performed on each of the 20 variables to examine whether the effect of previous diagnosis of alcohol dependence was independent of the effect of the psychiatrists' own diagnosis. Overall, there did appear to be a reduction in the size of effect after adjustment for the psychiatrists' own diagno-

illustrated that even when the respondents made a diagnosis of depression there was still an effect of

3.0

2.8

2.6 -

2.4 -

No Diagnosis Alcohol Dependence

Figure 1. Illustrating the interaction between past diagnosis of alcohol dependence and a special interest in addictions. The right-hand axis is the mean score on the composite variable. It is the sum of all items in Table 1 that were significant at the 5% level. Higher scores indicate more rejecting attitudes. (0) Special interest in addictions; (C)

other interests.

previous diagnosis of alcohol dependence. In addi- tion, the alcohol dependence and personality dis- order diagnoses were associated with more critical attitudes. These results are illustrated in Table 5 using a composite variable calculated by summing all those variables in Table 1 which were statistically significant at the 5% level.

Discussion This study has demonstrated that psychiatrists viewed people with a previous diagnosis of alcohol dependence as uncompliant, not accepting advice and having a poor prognosis. The responses to the questionnaire indicated alcoholics were judged in less need of admission to hospital, less likely to be prescribed antidepressants and were more likely to be discharged from follow-up. They were more annoying and less likely to arouse sympathy. There was no evidence that the sex of the case vignette substantially altered these rejecting attitudes.

Surprisingly, the male problem drinkers were regarded as having a lower suicide risk, while female 'cases' given a previous diagnosis of alcohol depen-

888 Michael Farrell & Glynn Lewis

Table 5. Diagnoses made by the responding psychiatrists in the experimental groups and illustration of the influence of psychiatrists' diagnosis on the composite variable

Depression Alcohol dependence Personality disorder Adjustment reactions Anxiety Neurasthenia

Total

Frequency of diagnoses

Alcohol dependence

35 (56%) 11 (18%) 8 (13%) 5 (8%) 2 (3%) 1 (2%)

62

No diagnosis

55 (74%) 0

3 (4%) 10 (13%) 6 (8%)

0 74

Mean score on composite

Alcohol dependence

2.65 2.96 3.24 3.00 2.65

variable

No diagnosis

2.27

3.20 2.52 2.50

dence were rated as higher suicide risks. Since alcoholics have a higher suicide rate, the lack of overall difference between the two groups (Table 1) may indicate a lack of concern for the safety of alcoholics. This is unlikely to be due to lack of knowledge.

Methodology A previous similar study demonstrated how a past diagnosis of personality disorder led to pejorative attitudes and the methodological issues of using this method have been discussed there (Lewis & Ap- pleby, 1988).

The experimental nature of this method is an important advantage. Critical variables in the 'cases' can be manipulated independently of all potential confounders and subjects can be randomly allocated to the different conditions. Though some questions, for instance the prescribing of antidepressants, may indicate appropriate management for an alcoholic patient many of the other questions tap attitudes and beliefs that most would regard as unfavourable to the alcoholic.

This study was designed to elicit the attitudes to a short clinical vignette. The actual mangement of patients cannot be inferred from responses to this questionnaire. However, attitudes to patients are an important area of study, particularly when consider- ing the management of emotional issues.

Reasons for negative attitudes There is general agreement that the management of problems drinkers can be a difficult business, but this property is shared with many other psychiatric disorders. Four possible explanations will be dis- cussed here.

Personal and social prejudice. It is not surprising to find that many psychiatrists share a common social prejudice about substance abusers. It has been demonstrated that an 'alcoholic' carrying a bottle is less likely to be helped than the same subject carrying a stick and presumed disabled. Weiner (1980) has argued that this is because people are more likely to attribute responsibility to the actions of alcoholics. Such theories of causal attribution were supported in this study by the observation that alcoholics were regarded as being more in control of their debts than others. Doctors cannot divorce themselves from the culture in which they live but one hopes that such popularly held beliefs will not adversely affect the treatment of problem drinkers.

Doctors are pressured by competing clinical demands and such views may lead him/her to think that other patients are more deserving of precious clincal time. A study of medical students asked them to write a referral letter for a patient with a peptic ulcer with two earlier diagnoses of alcoholism. Four-fifths of them omitted to mention the history of alcohol problems in their letter and only one in seven recommended follow-up care for the alcohol problem (Flaherty & Flaherty, 1983).

Diagnostic labelling and stigmatization. This study has provided some evidence that alcohol depen- dence is a stigmatizing diagnosis. This is not a good argument, on its own, to abandon this diagnosis considering its importance and usefulness in the management of substance abuse. Lewis & Appleby (1988) suggested that the critical attitudes towards personality disorder argued towards abandoning that concept. However, the important issue is whether the adverse effects of labelling are outweighed by the helpful and useful results of the

Discrimination on the grounds of diagnosis 889

label. The advantages of the dependence syndrome in guiding managment and research are consider- able. It is difficult to provide equivalent support for the diagtiosis of personality disorder. One weakness of many of the explanations of pejorative attitudes towards problem drinkers is the assumption that they only approach doctors with alcohol problems. Drinkers are at great risk of presenting with alcohol related physical or psychiatric problems and in this study's case vignette, someone with a past diagnosis of alcohol dependence was presenting with depres- sive symptoms. It would appear to be particularly unhelpful if the past diagnosis of alcohol depen- dence were to adversely affect medical treatment for different though related conditons.

Training. In the present study there were weak suggestions that the group of doctors with a special interest in addictions were somewhat less rejecting in their ratings. This could be because training in the addictions leads to a less judgemental approach, or that doctors with a less judgemental approach are attracted to work in the area of addictions. This association is consistent with other literature that indicates positive attitudes are associated with more knowledge of and experience with alcoholics (Cartwright, 1980; Geller et al., 1989).

In this study the psychiatrists regarded the alcoholic patient as more difficult. 'Difficulty' can be interpreted as reflecting professional uncertainty where there is an absence of clear guidelines and little confidence that one's therapeutic approach is appropriate. The results of this study are consistent with the view that psychiatric education does not provide sufficient training in the addictions and this results in psychiatrists who lack confidence in treating alcoholics.

Therapeutic pessimism. Many doctors expect poor compliance and high drop-out rates from problem drinkers but no research has been done to compare this with other persisting problems like anxiety or depression. Thus therapeutic pessimism may result in negative attitudes to patients which may discour- age patients on the brink of positive change (Edwards, 1987).

Among professional helpers liking for clients is consistently correlated with the belief that the client has improved or will improve (Doherty 1971; Willis, 1978).

Implications There is a high prevalence of psychiatric disorders among problem drinkers (Ross et al., 1988) with 20% of men and 27% of women suffering severe depression, and an elevated parasuicide rate in both sexes. This study has suggested that the previous diagnosis of alcohol dependence has an adverse effect on psychiatrists' attitudes to a depressed patient.

A possible way to address this problem is through the training provided at an undergraduate and postgraduate level in substance misuse. Surveys have shown that British medical schools allocate 4-8 h to formal teaching in substance misuse (Glass, 1988). There are not figures available for postgra- duate psychiatric training but it is suggested that substance misuse ranks as a similar low status subject at this level.

One of the basic responsibilities of a doctor is to provide a non-judgemental, humane and sympa- thetic ear to the patient and to make management decisions unbiased by prejudice. Is the diagnosis of alcohol dependence grounds for discrimination in medical care?

Acknowledgements We wish to thank all those who completed and returned questionnaires. We would like to thank Professor Edwards, Graham Dunn and Troy Cooper for help and advice and Karen Lovell for secretarial assistance. This project was funded by the Addiction Research Unit. Michael Farrell is supported by Action on Addiction. Glyn Lewis is currently supported by the Department of Health.

References BROOK, P. (1974) The postgraduate education and train-

ing of consultant psychiatrists, British Joumai of Psychiatry, 124, pp. 109-124.

CARTWRIGHT, A. K. J. (1980) The attitudes of helping agents towards the alcoholic client: the influence of experience, support, training and self-esteem, British Joumai of Addiction, 75, pp. 413-431.

DOHERTY, E. G . (1971) Social attraction and choice among psychiatric patients and staff: a review, Joumai of Health and Social Behaviour, 12, pp. 279-290.

EDWARDS, J. G. (1987) The Treatment of Drinking Problems, 2nd edn (Oxford, Blackwell Scientific).

EARRELL, M . P . & DAVID, A. S. (1988) Do psychiatric registrars take a proper drinking history? British Medical Joumai, 290, pp. 395-396.

FISHER, J., MASON, R. & KEELEY, K. (1975) Physicians and alcoholics: the effect of medical training on attitudes

890 Michael Farrell & Glynn Lewis

towards alcoholics, Journal of Studies of Alcohol, 36, pp. 949-955.

FLAHERTY, J. & FLAHERTY, E . (1983) Medical students performance in reporting alcohol related problems. Journal of Studies of Alcohol, 44, pp.1083-1087.

GELLER, G. , LEVINE, D . M . , MAMON, J. A., MOORE, R. D., BONE, L. R. & STOKES, E. J. (1989) Knowledge, attitudes and reported practices of medical students and House staff regarding the diagnosis and treatment of alcoholism. Journal of the American Medical Associa- tion, 261, pp. 3115-3120.

GLASS, I. B. (1989) Undergraduate training in substance abuse in the United Kingdom, British Journal of Addiction, 84, pp. 197-202.

GLASS, I. B. & JACKSON, P. (1988) Maudsley Hospital survey: prevalence of alcohol problems and other psychiatric disorders in a hospital population, British Journal of Addiction, 83, pp. 1105-1111.

LEWIS, G. & APPLEBY, L . (1988) Psychiatrists responses to personality disorder, British Journal of Psychiatry, 153, pp. 44-49.

.MACDONALD, H. B. & PATEL, A. R, (1975) Attitudes

towards alcoholism, British Medical Journal, 2, pp. 430-431,

MITCHELL, C . (1989) Alcohol histories taken by Psychia- tric registrars: a comparative study in a non-teaching hospital, British Journal of Addiction, 84, pp. 438-439,

MOGAR, R. E., HELM, S. T . , SNEDEKER, M . R. & WILSON, W. M. (1969) Staff attitudes towards the alcoholic faiieat. Archives of General Psychiatry, 21, pp. 449-454.

Ross, H. E., GLASER, F. B., STIASNY, S. (1988) Sex differences in the prevalence of psychiatric disorders in patients with alcohol and drug problems, British Journal of Addiction, 83, pp. 1179-1192.

STRONG, P. M. (1980) Doctors and dirty work—the case of alcoholism, Sociology of Health and Illness, 2, pp 25-47.

WEINER,B. (1980) A cognitive (attribution)—emotion- action model of motivated behaviour: an analysis oi judgements of help giving. Journal of Personality & Social Psychology!, 39, pp. 186-200.

WILLS, T . (1978) Perception of clients by professional helpers., Psychological Bulletin, 85., pp. 968-1000