Whereas administering the Camberwell Family Interview requires no skills beyond those possessed by most competent clinicians, scoring these scales from tape-recordings of conversations between interviewers and relatives is very difficult indeed and requires special (and, from personal experience, very tedious) training. This is because the intonation and emotional expression of the relative are crucial. For example, a statement such as ‘He wandered around town in a confused state and was picked up by the police’ can bea straightforward description of events if expressed in a level tone of voice but can be a highly critical comment if expressed with sufficient venom. Not surprisingly, some researchers have attempted to develop alternative measures to the CFI, based, for example, on questionnaires or analyses of short segments of speech, but these have never been entirely successful.31
The results of research on schizophrenia patients conducted in the years since Brown’s initial studies have been very consistent. In 1994, Paul Bebbington and Elizabeth Kuipers of the Institute of Psychiatry in London obtained individual data from 1346 families who had been studied in twenty-five separate investigations, of which only three had yielded negative results.32 Across the studies, about half of the families investigated were judged to include a high EE relative. The overall relapse rate during the first year after patients had returned to these families was 54.4 per cent whereas the rate for low EE families was 21.8 per cent. However, as always, other factors complicate this apparently simple picture.
Figure 16.6 shows data from a famous study of the effects of EE by
Figure 16.6 Relapse rates for schizophrenia patients living with low and high EE relatives, showing the moderating effects of face-to-face contact time with the relatives and receipt of neuroleptic medication (from Vaughn and Leff, 1976).
British researchers Christine Vaughn and Julian Leff, published in 1976, from which it is obvious that patients living with high EE relatives are especially likely to relapse if they spend a lot of time in contact with those relatives.33 Vaughn and Leff’s findings also indicate an interaction between expressed emotion and patients’ medication. Over the nine months of the study, patients taking anti-psychotic drugs received some protection from their high EE relatives. Although the data seem to indicate that those living with low EE relatives received no benefit from their drugs, this finding was not supported when the patients were followed up over a longer period. Vaughn and Leff concluded that medication protected patients from all kinds of stress, and not just the stress of living with a critical or over-controlling relative.34
Vaughn and Leff’s research led to a second important qualification about the impact of expressed emotion. Studying people who had experienced admission for depression as well as those with a diagnosis of schizophrenia, they found that the depressed patients were even more vulnerable to the adverse effects of stressful relationships than the schizophrenia patients. In fact, it is now clear that Brown’s discovery was not a discovery about schizophrenia at all, but a discovery about people with psychiatric complaints in general. The impact of high EE spouses on the well-being of recovering depressed patients has since been the focus of a series of investigations by Jill Hooley, a British-trained psychologist based at Harvard University, in the USA.35 David Miklowitz and his colleagues at the University of California in Los Angeles have extended EE research to patients with a diagnosis of bipolar disorder. Once again, it has been shown that living with a high EE relative increases the risk of future episodes of illness, in this case of depression and mania.36
Surprisingly, researchers have almost completely neglected the psychological mechanisms responsible for the impact of high expressed emotion. Although it seems obvious that exposure to constant criticism and over-protective (in extreme cases, infantilizing) attitudes will result in emotional distress and a loss of self-esteem, few attempts have been made to investigate these kinds of effects, or to reveal how they might lead to an exacerbation of psychotic symptoms.
A few researchers have attempted to measure the immediate emotional impact of exposure to a high EE relative. British psychologists Nick Tarrier and Graham Turpin have demonstrated that schizophrenia patients show unusually high levels of emotional arousal (as measured by changes in the electrical conductivity of their skin) when meeting a high EE relative.37 In a study conducted by the UCLA schizophrenia research group, it was also found that high expressed emotion relatives can provoke transient increases in patients’ symptoms. Patients in the study often expressed unusual thoughts immediately after hearing a critical remark from a high EE parent.38
Christine Barrowclough and Nick Tarrier have recently carried out the first study in which the impact of expressed emotion on self-esteem has been measured directly. They used an interview measure of self-esteem, which allowed them to distinguish between positive and negative beliefs about the self. They found that a high level of criticism by a close relative was a strong predictor of negative beliefs about the self, and that these negative beliefs in turn predicted positive symp-toms.39 This finding is, of course, consistent with other evidence about the role of the self in psychosis that we have encountered in earlier chapters.
The nature of high expressed emotion
Some investigators have wondered how the high EE attitudes translate into the actual behaviour of relatives in the presence of a patient. In an attempt to answer this question, researchers at UCLA watched schizophrenia patients and their parents as they discussed family problems. They found that high EE parents, in comparison with low EE parents, made more critical and intrusive comments to their ill sons and daughters.40 David Miklowitz described this type of interaction as negative affective style. Subsequent studies have shown that, when measured as an alternative to EE, negative affective style is a strong predictor of relapse.41
Other researchers have wondered why some relatives exhibit high expressed emotion, whereas others do not. One approach to this problem has involved looking in more detail at the beliefs associated with high EE. In earlier chapters I have emphasized the role that patients’ attributions (causal beliefs) play in their complaints, but, of course, psychotic patients are not the only people who make inferences about the causes of events. The attributions of schizophrenia patients’ relatives have been investigated in studies conducted by Chris Brewin in London42 and by Christine Barrowclough in Manchester.43 Perhaps unsurprisingly, those relatives who were judged to be high EE, especially those who were hostile or made many critical comments, tended to explain the distressing behaviour of their mentally ill sons and daughters in terms of causes that were internal to the patient and controllable, for example, saying that they were ‘lazy’ or ‘difficult’, as if they could choose to be otherwise.
Other investigators have wondered whether parents’ EE status might be related to the stress involved in caring for their mentally ill children. Of course, the experience of seeing an adult son or daughter suffering is almost always extremely distressing. This distress is often compounded by the problems involved in attempting to help the son or daughter cope with the demands of everyday life, while at the same time trying to obtain help from psychiatric services that are under-resourced and sometimes unsympathetic to parents’ needs. When it is remembered that these kinds of difficulties can continue for many years, often at a time of life when most parents can expect to be released from the responsibility of caring for their children, it is not surprising that many parents in this position experience high levels of depression and anxiety.44
In a recent study conducted by Elizabeth Kuipers and David Raune at the Institute of Psychiatry, it was found that high EE parents experienced the burden of caring for their mentally ill sons and daughters more severely than low EE parents. However, in the same study, it was found that they often attempted to cope with this burden by ‘disengaging’ from it, for example by avoiding discussing problems with their children, or by using alcohol to relieve their distress.45 This finding raises the possibility that high expressed emotion arises partly as
a consequence of the parents’ difficulties in coping with relationships in general.
An earlier study conducted by American psychologists Diana Diamond and Jeri Doane provides some support for this idea.46 Parents of severely disturbed adolescents and young adults (some of whom were psychotic) were interviewed about their relationships with their children, and also about their relationships with their own mothers and fathers. The more inadequate the attachment relationship between the mothers in the study and their own mothers, the more likely they were to show a negative affective style towards their children. As a consequence of not receiving adequate emotional support in childhood, it seemed that they did not have the emotional resources necessary to offer effective support to their own distressed children. This finding has since been replicated by a group of British investigators.47
Further caveats
Overall, studies of expressed emotion add up to one of the few unequivocally successful stories in modern psychiatric research. The findings from these studies clearly demonstrate that the behaviour of psychotic patients cannot be fully understood without understanding the social environment in which they live. However, two further qualifications about the expressed emotion findings are warranted. First, we must recognize that interactions between patients and their relatives are determined in part by both parties. It is dangerous to assume that high EE parents are somehow exclusively ‘to blame’ for the adverse affects that they have on their children. Patients themselves contribute to these kinds of negative emotional interactions. This was demonstrated in the UCLA study that examined the immediate impact of criticism on patients’ symptoms. Careful analysis revealed that critical comments by parents were often triggered by unusual or psychotic comments made by their children. Indeed, high EE behaviour appears to be part of a cycle of escalating negative emotion, in which the bizarre or unusual behaviour of the patient triggers critical comments, which in turn lead to more bizarre and unusual behaviour.
The second qualification concerns researchers’ focus on stressful relationships with parents. Jill Hooley’s work with people suffering from depression has mostly examined expressed emotion in spouses, so it is clear that husbands and wives can also have a negative impact on the well-being of people who are recovering from a psychiatric crisis. More recently, researchers have wondered whether mental health professionals can also be capable of high expressed emotion attitudes.
Studies carried out by Elizabeth Kuipers and Estelle Moore at the Institute of Psychiatry in London have demonstrated that expressed emotion can be reliably measured in psychiatric care staff.48 In a comparison of two hostels for recovering patients, one of which was staffed by high EE carers, it was found that patients were more likely to leave the high EE hostel, presumably because it did not provide a comfortable living environment.49 A similar study in Los Angeles found that patients living in a residential care home staffed by high EE carers had a poorer quality of life and worse symptoms than patients living in a low EE environment.50 Clearly, mental health professionals cannot afford to be complacent about their therapeutic impact on their patients. Indeed, it seems likely that there are high EE psychiatrists, high EE nurses and (although I do not like to admit it) high EE clinical psychologists who would probably be better employed doing something different.
The Slings and Arrows of Outrageous Fortune
When recovering patients live with a high EE relative, spouse or care worker they are exposed to a constant and persisting form of stress. Other stressful events, which occur more suddenly and unpredictably, are known as life events in the jargon of modern psychiatric research. These may include separation from a loved one, losing a job, being forced to move home, being arrested by the police, or any number of other calamities. Again, the idea that such events can have a negative impact on mental health is hardly new – Kraepelin noted that periods in which his patients’ symptoms remitted often came to an end when they experienced major life changes. However, despite the intuitive appeal of this observation, it has been surprisingly difficult to establish a clear link between specific life events and episodes of illness.
Part of the problem lies in the formidable challenges facing researchers. For practical reasons, most studies have been retrospective, and have attempted to assess what has happened to patients in the weeks or months preceding a breakdown. However, memories are not always reliable, patients’ interpretations of events may be affected by their difficulties and, in any case, the importance of different kinds of events may vary between individuals. In practice, researchers have attempted to overcome these problems in one of two ways.51 Some have attempted to draw up questionnaires listing life events that have previously been judged as stressful by a panel of ordinary people. Participants have then been asked to indicate those events on the questionnaire that they have experienced within a particular period. The second, more sophisticated approach was developed by George Brown and his colleagues in London and involves interviewing patients, using a Life Events and Difficulties Schedule (LEDS), to obtain a detailed account of their recent experiences. The patients’ descriptions of events are then presented to a panel of independent judges who score them on various rating scales, for example measuring their severity from the standpoint of the patient.
A further difficulty has been that of teasing out the causal relationship between an event and the emergence of symptoms. Although it is easy to assume that stressful life events must precipitate complaints, sometimes the relationship is the other way round. For example, a patient may become so paranoid that she falls out with her employers and loses her job. George Brown and his colleague Tirril Harris therefore argued that life events should be rated according to whether or not they appear to be independent of symptoms. This is usually done by assigning each recorded event to one of three categories: independent, possibly independent and dependent. In general, researchers have only counted the first two of these categories when exploring the link between life events and increases in episodes of illness.
Using this approach, in 1978 Brown and Harris reported the results of a landmark study, which they had carried out in Camberwell, an economically deprived suburb of London. In by far the majority of cases of depression they detected in a group of women living in the community, they found that significant life events preceded the onset of symptoms by only a few weeks. However, they found that many women who experienced life events did not become depressed, so other factors were obviously important. Women who were living in stressful social circumstances (who were unemployed, had small children at home or lacked supportive relationships), or who had lost their mothers at an early age, were especially vulnerable to reacting badly to an adverse event.52 In general, later studies have confirmed Brown and Harris’s finding of a clear link between negative life events and depression. Indeed, in twenty studies using Brown’s Life Events and Difficulties Schedule, estimates of the number of depressive episodes closely preceded by negative life events have varied between 67 per cent and 90 per cent.53 These studies have also shown that the kinds of social factors that leave people vulnerable to the impact of life events may vary with location. (In a study conducted in the Outer Hebrides, for example, Brown found that people who were poorly integrated with the local community were especially vulnerable.)54
Not surprisingly, some researchers have tried to discover whether there is a similar association between life events and episodes of psychosis. The earliest study of this kind was conducted by Brown and his colleague Jim Birley, which was published in 1968.55 They reported that 46 per cent of a group of schizophrenia patients had suffered a life event in the three weeks preceding a relapse, and therefore concluded that life events could trigger the onset of psychotic symptoms. However, the results of subsequent studies have not always been consistent with these findings. Whereas some found a similar link between life events and relapse, others did not. A study by Paul Bebbington, conducted in Camberwell in the early 1990s, compared the life events of fifty-one patients who had expe
rienced a relapse with the life events experienced by a group of ordinary people living in the community, and found that the patients had experienced an excess of independent life events in the three months prior to becoming ill.56 Another recent study, carried out by Steven Hirsch and his colleagues, also in London, involved the regular assessment of patients who received either anti-psychotic medication or a placebo. In this study, no effect of life events was observed when the four weeks preceding a relapse were examined, but an effect was found when life events over the previous year were calculated.57 One possible implication of this finding is that psychotic breakdowns sometimes occur after stressful experiences have accumulated over a long period of time.
In a 1995 review of similar studies conducted with bipolar patients, American psychologists Sheri Johnson and John Roberts were able to identify fifteen studies that had used questionnaire measures and ten that had used interview methods (usually the LEDS).58 Although the findings were not entirely consistent, the overall picture suggested that life events can trigger episodes of mania as well as depression. Johnson’s own study of sixty-seven bipolar patients, published two years later, found that those whose episodes had been triggered by negative life events took, on average, three times longer to recover than those whose episodes began without severe life events.59
Madness Explained Page 49