Part 4: Summary and conclusions 176 Chapter 13: 13.1 Summary of key findings Implications for the Patient Of key importance for the patient’s outcome are the findings regarding EE and DUI. The LEE(S) was not only able to predict symptoms, but also social functioning and quality of life at one year in this group of patients. However, the most important finding in relation to this group is that lack of emotional support rather than criticism, is the key component of EE. DUI was also found predict symptoms, social functioning and quality of life independently of EE. Therefore this study suggests that EE and DUI are independent predictors of outcome in these populations and that both should be the target of interventions. Especially since those caregivers whose relative had a longer DUI also perceived their symptoms to be more bothersome than those with shorter DUI. Another potentially important finding for this population is that in contrast to Western populations, a greater number of family members in the patient’s social network was more satisfying, and was associated with better quality of life for the patient. So, by targeting family attitudes and intervening as early as possible in the course of the illness, patients will ultimately benefit from a better illness prognosis. Social support from family members can also be encouraged to help the patient achieve a better outcome. 13.2 Caregiving in Singapore Caregiving in Singapore is seen to be very similar to caregiving in Western populations with similar concerns being held by both. However, this study highlighted some potentially important cross-cultural aspects in relation to appraisal. 177 Concepts such as, Duty, the importance of family, the impact of collectivism etc. whilst not directly measured in this study, are suggested as possibly underlying some of the factors such as appraisal and attributions and which may be useful to investigate further. Contrary to findings from their Western counterparts, neither avoidant nor active coping was related to EE in this study. It seems reasonable therefore to consider the alternative suggestion that Active Coping is having an influence on how the Caregiver appraises the situation rather than vice versa. Their appraisal is then associated with lower levels of EE. Another option is to look to an attributional model as an explanation for high EE rather than a stress and coping model. There is some support for an attributional model from this study; caregivers who perceived their relative as being in control of his or her illness also perceived the symptoms and behaviours to be more bothersome, perceived the whole situation more negatively and also tended to be higher in EE. Clearly future studies can build on these findings and examine further the role of control attributions in relation to EE, and perhaps even look at the role of causal attributions. Optimism as a personality dimension was involved to some extent in relation to the relative’s appraisals, but other personality variables need to be investigated such as that suggested earlier, shame-proneness. The evidence presented suggests that caregiving in this population cannot be approached using Western models, or Western instruments, alone. There are many cultural aspects affecting this group and these should be explored further and included in instruments for measuring optimism, coping and appraisal in particular. In addition, culturally appropriate measures already developed by Asian scholars could be used to validate future studies in psychosocial behaviours and mental health in Asian settings. 178 13.3 Summary of Cultural factors Many cultural factors were suggested as being important in this study. Firstly regarding the patients, as expected Criticism and Intrusiveness were not shown to be important factors with regards to EE and outcome. It was suggested that the importance of close family ties in this culture may be associated with this, as criticism is seen more as a personal attack and therefore relevant in an individualistic culture whilst lack of emotional support (from the family) is seen as affecting the family and group harmony more. In addition, the patient’s strong satisfaction with family support is in line with the importance of Family in this culture. Keeping social support confined to close family could also be reflecting the issue of Face. It is not clear whether the patient’s access to outside support may have been limited by the family in order to preserve their ‘Face’ . Stigma surrounding mental illness tends to fall more on the family than on the individual in Chinese families (Ow & Katz, 1999) so it could be that the family’s desire to avoid the stigma associated with mental illness and the associated loss of Face, prevented them (and therefore the patient) from mixing with others outside of the immediate family. With regards to the caregivers, the notion of Duty or Yuan was suggested as being instrumental in preventing negative appraisal leading to higher levels of EE, Wong (2000) found similar findings in a study of Chinese caregivers in Hong Kong. Collectivism was also seen as an important factor in appraisal as it has been suggested that collectivist cultures perceive caregiving as a reward, and this may have resulted in more positive appraisal in these caregivers. A further suggestion in regards to positive appraisal is that caregivers may be using the concept of ‘xiao xin yan’ to explain the illness (Yang, 2003), making it seem more normal and less threatening. If this was the case, stigma and associated loss of Face would not be an issue for these caregivers. 179 13.4 Implications for the vulnerability/stress (V/S) model In order to guide future research and aid in understanding the process of this illness, it is essential to know how useful the V/S model actually is in predicting outcome. There is support to some extent for the model as hypothesized however there is an alternative which may be more plausible. The hypothesized model stated that EE was the environmental stressor and predicted outcome, moderated by the environmental protectors: Positive Appraisal, Dispositional Optimism, Active Coping and External control attributions and by the patient’s personal protector: Social Support. This is the model discussed in Section 11 (Figure 11.4), and is supported to a great extent by the findings of this study. The effects of the environmental protectors can be seen in that as predicted they were associated with lower EE which is in turn was associated with better outcome for the patient. However, the findings not support social support as the patient’s personal protector, some other factor such as coping styles must be acting in this capacity. There is an alternative way of looking at the V/S model. EE could be conceptualized as a form of social support and as such would become an environmental protector rather than a stressor. Some other factor would take the role of environmental stressor such as stressful life events. The findings of this study can also be seen to lend support to this alternative. For instance, support was found in this study for the Relationship Perspective which suggests that social support is very difficult to separate from close relationships. The idea of low EE being a protective factor has been noted elsewhere (e.g. Leff & Vaughn, 1992) and Brown et al. (1972) found that relapse rates dropped from 22% to 9% when warmth was high. However what makes the idea of EE being a form of social support in this population is the importance of Lack of Emotional Support (LES). Low LES was conceptualized as Warmth in this study based on results of the small focus groups discussed in Appendix B, and the 180 correlations with the Warmth scale of the Chinese Parenting Scale. So there is evidence from the literature, and from the present study that warmth (i.e. low LES) could be construed as a protective factor. Clarifying the components of this model is important for this group of patients, and for research in the area of first-episode psychosis generally. Whilst some aspects will be culture specific others will generalize across cultures. Caregiving is a complex concept, and it is only by knowing and understanding how the different facets (such as personality and attitudes) work together that effective interventions can be designed and implemented. 13.5 Directions for Future Research The present study has highlighted certain very clear directions that future research should take. Firstly, the V/S model has been shown to be a useful framework, but this study suggests some modifications: instead of high EE being perceived as the stressor, Low EE is seen as a protective factor, and either DUI or possibly Stressful Life Events as the stressor. However, before these aspects can be successfully studied, some amendments need to be made to the instruments. The LEE(S) proved successful in predicting outcome and suggests that EE is an important factor for this group. However, this predictive ability was mainly driven by the sub-scale Lack of Emotional Support. The remaining sub-scales did contribute to prediction but in a very minor way. The LEE(S) therefore needs some modification. Although a stringent validation procedure was carried out on the LEE(S), there are clearly some other factors in this population that the LEE(S) is missing. The concepts of criticism, irritability and intrusiveness should be explored in a patient sample using qualitative methods to access what exactly these concepts mean for this group of patients. Indeed, it may be that there are concepts other than these that are more important, and perhaps specific to this cultural group. In addition 181 social support was not seen to be a protective factor. Perhaps locus of control, or coping, could be examined as potential protectors. As the Lack of Emotional Support sub-scale was so important, this concept should also be explored further to ensure that the LEE(S) is fully capturing the concept for this group. As to what might be underlying EE, the present study has again underlined some areas for future research. The language difference noted in certain sub-scales of the ECI suggests that for the Chinese speaking caregivers cross-cultural aspects such as collectivism are still relevant. Clearly scales cannot be imported and used in other cultures without careful validation procedures, but Singapore presents an even more complex situation. The instrument must be suitable for both the more traditional Chinese speaking population and the more westernized English speaking population. Although the ECI had previously been validated for use in Singapore, and was successful in measuring appraisal in this group, it seems that some of the sub-scale items such as those which referred to family members setting up in their own homes, were not relevant. Also, the inclusion of a sub-scale measuring such concepts as ‘duty’ would be advantageous. Whilst coping and attributions were shown to be related to EE, there are clearly other unidentified factors involved. Some other personality styles should be explored such as shame and guiltproneness. Clearly, the cultural situation regarding these patients and caregivers in Singapore needs to be explored further and in more depth, perhaps using qualitative studies, in order to determine the extent that concepts such as collectivism and other Chinese concepts such as Yuan, may be having on the situation. As Avoidant coping in this study was not seen to be related to EE, it would be beneficial to test the hypothesis that avoidant coping leads to lower physical and psychological health in caregivers. Although this is not related to EE directly, a caregiver in low health may be less able to offer emotional support to a patient. 182 Whilst the results cannot be generalized to patients with psychosis in other hospitals in Singapore, they provide a basis from which to investigate these other populations. In addition, EE could be examined in physical illnesses such as diabetes as well as other psychiatric disorders such as depression. Finally, studies need to be conducted on both Malay and Indian families and patients to address the specific needs of these groups. Perhaps more longitudinal studies would be useful to follow the caregivers over a longer period of the patient’s illness, thus allowing for changes over time and phases of the illness. This could incorporate more qualitative work. 13.6 Implications for clinical practice 13.6.1 Methodology There were several methodological issues to be addressed in this study. The first concerned the use of the LEE as a tool for assessing expressed emotion. The second concerned the use of varied measures of outcome. The LEE was subjected to a series of stringent tests to ensure that it was valid for use in this population in both English and Mandarin. While many forms of validity were tested prior to its use (e.g. conceptual, construct), its predictive validity was confirmed only once the study was complete. The findings confirmed that the LEE is a quick and easy tool that clinicians can use to determine levels of EE in a patient’s home environment. The reliability and validity of this instrument is of great importance for several reasons. Firstly, patients are subjected to many questionnaires and psychometric tests in the course of their initial and follow-up appointments. For this reason brief tests can be very advantageous, for the clinician’s convenience and in particular, to avoid overtaxing the patient. 183 Secondly, as noted earlier, coping with a mentally ill relative can be burdensome for the caregiver. These caregivers may be feeling a range of emotions from worry and anxiety to guilt and shame. Having a dichotomous high/low EE distinction only adds to this burden for the caregiver. If a caregiver is labeled ‘high EE’, they could be made to feel responsible for the patient’s outcome. Kavenagh (1992) believes that this tendency to label caregivers could be greatly reduced if researchers used a continuous, rather than a dichotomous measure of EE, and examined the components of EE and their relation to outcome. The LEE(S) therefore allows for this continuous measure of EE rather than the dichotomous high/low distinction. Clearly the LEE(S) in its present form has shown itself to be a useful tool for measuring the concept of EE in this population and gives support for the use of self-report instruments in these cases. However, as mentioned previously, more research into the concepts for this population may increase its usefulness and predictive power. Kavenagh (1992) raises a further point which has direct relevance for the second aim regarding methodology. He states that moving from a relapse/non relapse view of outcome to one which looks at continuous outcome variables such as symptom severity and level of social functioning can also help reduce the stigma for the families. The findings from this study showed that EE and DUI were able to predict not only symptomatological outcome but also functional outcome and quality of life. 13.6.2 Interventions The findings of this study serve as reference for evidence based practice, primarily in terms of interventions for the patients and caregivers. The potential importance of cultural factors for this group of patients and their caregivers has been highlighted throughout. It is therefore important to ensure that any interventions are 184 culturally appropriate. For instance, interventions with Western families have focused on communication skills with the emphasis being on reduction of criticism. Evidence from this study suggests that in this population this would not be appropriate as criticism is not a key factor. There are several areas which have emerged throughout this study which give opportunities for intervention and these will now be examined. Firstly, regarding DUI, the finding that DUI is associated with outcome suggests that intervention at this time is crucial. The researchers at the department of Early Psychosis Intervention at the Institute of Mental Health (IMH), Singapore, currently have a study underway using MRI to look at structural brain changes in people with psychosis, their first degree relatives and controls. This study hopes to target those in the early stages of theillness. There are also plans to scan those at high risk of psychosis and to follow them up longitudinally. The reduction of DUI would clearly benefit those at risk of psychosis, but would also have implications for the caregivers too as longer DUI was associated with more negative perception of the patient’s symptoms and of the situation in general. Regarding EE, a finding of this study was that a family perceived as lower EE by the patient is associated with better outcome for that patient after one year. As positive appraisal seems to be key in relation to EE in this group of caregivers, this may be a good place to target interventions. Whilst interventions are generally focused on regulating distress, recently more attention has been given to generating positive emotions during stressful encounters. Moskowitz, Folkman, Collette & Vittinghoff (1996) have shown that positive affect can be increased through stimulating positive reappraisal. Affleck and Tennen (1996) have also shown that benefit finding and benefit reminding have also increased well being. Encouraging positive aspects, rather than trying to change negative aspects, may also be beneficial in increasing the family’s sense of worth. 185 An optimistic disposition was found to be beneficial in this group of caregivers, allowing them to perceive the situation more positively. For those who were low on dispositional optimism, there are interventions available which can help people to become more optimistic. Seligman (1990) believes that how people explain events to themselves determines whether they are optimistic or pessimistic. He has devised a model, the ABCDE model which attempts to address people’s beliefs and actions regarding adverse events in their lives and train them to have adaptive ways of thinking and behaving. This sort of model may be helpful for those caregivers low in Dispositional Optimism. Control attributions also presents a possible area for intervention. Education has been seen to be the most beneficial form of intervention in this regard. Interventions would generally focus on correcting misperceptions about the disease (Lopez & Wolkenstein, 1990). The department of Early Psychosis Intervention at IMH currently runs education for the caregivers, however use of the LEE(S) and the Control Attributions Scale may be able to highlight those families most in need of this type of intervention. Other potential areas for education include encouraging the use of Active Coping, allowing caregivers to identify situation specific goals. Also, given the importance of emotional support for the patients, this is an area in which caregivers could be educated. If caregivers are aware of how important it is to be supportive, they may be able to curb their tendency to withdraw from the situation, or the patient themselves, and offer more support. The coping strategies of seeking instrumental and emotional social support were seen to be linked with Positive Appraisal. Although traditionally Chinese caregivers have been reluctant to discuss this type of situation with outsiders to the family, the department of Early Psychosis Intervention has initiated a series of support groups in both English and Mandarin where family members can meet with other families and discuss areas of mutual concern. It is possible that stigma and 186 shame are not apparent when talking with others in the same situation. These groups provide an ideal way of providing the education that these families need at this time and these should be encouraged. Evidence for the utility of such an intervention comes from a study of caregivers in Hong Kong who also share the same issues with Face as Chinese caregivers in Singapore. In a comparison of two different interventions (mutual support and psychoeducation) with a control group, the mutual support intervention group’s relatives showed better functioning and less rehospitalisation compared with the other two groups (Chien & Chan, 2004). There is also a need to examine in detail the formal and informal support networks of caregivers as this can affect not only EE but also the experience of caregiving. 13.7 Limitations of the study This study used the V/S model to test the hypotheses regarding the different variables. However, the personal vulnerabilities were not included in the model as they were beyond the scope of the study, and clearly they would need to be included at some future time to fully test the model’s utility. Perceived EE was measured in this study and although some studies of perceived EE were available for comparison, the majority of the cited studies used observed EE. It is not clear to what extent the construct measured by perceived EE differs from that of an objective measure. This is clearly something which should be clarified in future work and could perhaps be done by using detailed case studies in addition to the self-report measure. Clearly in this study the patients were suffering from different degrees of a serious mental illness. It could have been the case that the patient’s psychopathology actually affected the way that the patient perceived the levels of EE in the family. For instance, a patient with more disturbed thinking may have viewed 187 the family as more hostile etc than they actually were. The comparison of the FRI scores between patients and caregivers suggests that this was not the case as there were no significant differences between them but it is still a consideration for several reasons. Firstly, if the patient had severe psychopathology which resulted in a more negative view of the family atmosphere, this severe psychopathology could also have contributed to poorer outcome in one year, and the association between EE and outcome would be spurious. However, the regressions analyses conducted between baseline outcome variables and one year outcome variables suggest that this is not the case. Secondly, if the patient’s view of the family was not accurate, this would have implications for interventions which would be aimed unnecessarily at apparently high EE families. Taking this patient’s perspective has so many advantages (outlined earlier) that it is felt important that perceived EE should be measured rather than objective EE. Further, by only interviewing patients when declared stable by the clinician would hopefully avoid patients in their more disturbed states. A further limitation regards the use of path analysis. Ideally, full structural equation modeling (SEM) would have been preferable. The use of SEM would have allowed for examination of latent rather than observed variables. When this study was initiated the Early Psychosis programme at IMH had only just become operational and whilst there was some information on projected numbers of participants in the programme, there was little data on projected numbers of defaulters, refusers, etc. Ultimately the numbers of patients and caregivers available did not support the use of full SEM. A common failing of studies on EE is that only one caregiver per family participates. This can also be seen to be the case here. Whilst the patient’s perception was of the family atmosphere as a whole, the patient responded in theory to one particular family member. Although every effort was made to interview that 188 same family member this was not always possible. Often the patient would not identify one particular member, other times that particular family member was not available or not willing. It is possible that different family members may have held different attributions or attitudes which would have affected the family atmosphere. This could perhaps be avoided in future studies by asking the patient not to respond to one particular family member but instead to give an overall impression of the family atmosphere. As only 48% of the available patients participated, results cannot be generalized to the patient population of IMH as a whole. However, the findings provide a good basis from which to proceed. There was also no way of knowing whether the caregivers who refused to participate were in any way different from those who agreed to participate. Perhaps issues of Face were relevant here, with caregivers not wishing to lose Face by discussing the illness with an outsider. Whilst factors such as medication were considered not to be confounding the results, there are other therapeutic factors which may have had a confounding effect. Some patients did undergo counseling from the psychologists and there could have been some effect from this support. A final point is that with hindsight more qualitative work could have been conducted with the caregivers. As the study progressed it became apparent that some of the caregivers were very keen to talk about their experiences and their situation, albeit in a very informal way. This qualitative data could have elaborated on the in-depth interviews and helped to design future caregiver instruments. 13.8 Conclusion DUI and EE (as measured by the LEE(S)) are both seen as independent predictors of a variety of outcome measures in this population. 189 Caregivers are seen as using both Avoidant and Active Coping. However, a model which shows Active Coping being associated with EE, mediated by the caregiver’s Positive Experience of Caregiving seems to best describe the situation for these caregivers. In addition, caregivers who hold more internal attributions of control are more likely to be high EE which impacts negatively on the patient’s outcome. The potential importance of cultural factors cannot be over emphasized for both patients and caregivers. In most areas studied, cross-cultural factors were identified, in particular the role of Lack of Emotional Support as the equivalent of criticism in this culture. These factors need to be incorporated into future research and interventions. Although researchers claim that Singapore is becoming more and more westernized (e.g. Kiong et al., 1992) there is evidence from this study that many traditional Chinese concepts and beliefs still exist. The findings of this study have direct relevance for the patients of IMH in Singapore, and could potentially have relevance for other patients with psychosis in Singapore. However, they may also add to the body of knowledge regarding EE and Caregiving generally. With regards to EE, it has long been felt (Jenkins & Karno, 1992) that it is only with cross-cultural research that the mechanisms underpinning EE can be fully understood. For instance, by knowing that avoidant coping is not related to EE in this population, it alerts researchers in Western cultures that perhaps coping is not the key factor in regards to the low/high EE distinction. Other Asian populations could also benefit from the findings of this study. Whilst there is huge variation amongst different Asian cultures, there are also some similarities. These findings may open different avenues of enquiry for other Asian countries to explore. 190 [...]... study that many traditional Chinese concepts and beliefs still exist The findings of this study have direct relevance for the patients of IMH in Singapore, and could potentially have relevance for other patients with psychosis in Singapore However, they may also add to the body of knowledge regarding EE and Caregiving generally With regards to EE, it has long been felt (Jenkins & Karno, 1992) that it... actions regarding adverse events in their lives and train them to have adaptive ways of thinking and behaving This sort of model may be helpful for those caregivers low in Dispositional Optimism Control attributions also presents a possible area for intervention Education has been seen to be the most beneficial form of intervention in this regard Interventions would generally focus on correcting misperceptions... seen as independent predictors of a variety of outcome measures in this population 189 Caregivers are seen as using both Avoidant and Active Coping However, a model which shows Active Coping being associated with EE, mediated by the caregiver’s Positive Experience of Caregiving seems to best describe the situation for these caregivers In addition, caregivers who hold more internal attributions of control... way of providing the education that these families need at this time and these should be encouraged Evidence for the utility of such an intervention comes from a study of caregivers in Hong Kong who also share the same issues with Face as Chinese caregivers in Singapore In a comparison of two different interventions (mutual support and psychoeducation) with a control group, the mutual support intervention... potential importance of cultural factors cannot be over emphasized for both patients and caregivers In most areas studied, cross-cultural factors were identified, in particular the role of Lack of Emotional Support as the equivalent of criticism in this culture These factors need to be incorporated into future research and interventions Although researchers claim that Singapore is becoming more and more... about the disease (Lopez & Wolkenstein, 1990) The department of Early Psychosis Intervention at IMH currently runs education for the caregivers, however use of the LEE(S) and the Control Attributions Scale may be able to highlight those families most in need of this type of intervention Other potential areas for education include encouraging the use of Active Coping, allowing caregivers to identify situation... cross-cultural research that the mechanisms underpinning EE can be fully understood For instance, by knowing that avoidant coping is not related to EE in this population, it alerts researchers in Western cultures that perhaps coping is not the key factor in regards to the low/high EE distinction Other Asian populations could also benefit from the findings of this study Whilst there is huge variation amongst... traditionally Chinese caregivers have been reluctant to discuss this type of situation with outsiders to the family, the department of Early Psychosis Intervention has initiated a series of support groups in both English and Mandarin where family members can meet with other families and discuss areas of mutual concern It is possible that stigma and 186 shame are not apparent when talking with others in the... relatives showed better functioning and less rehospitalisation compared with the other two groups (Chien & Chan, 20 04) There is also a need to examine in detail the formal and informal support networks of caregivers as this can affect not only EE but also the experience of caregiving 13.7 Limitations of the study This study used the V/S model to test the hypotheses regarding the different variables However,... extent the construct measured by perceived EE differs from that of an objective measure This is clearly something which should be clarified in future work and could perhaps be done by using detailed case studies in addition to the self-report measure Clearly in this study the patients were suffering from different degrees of a serious mental illness It could have been the case that the patient’s psychopathology . a better outcome. 13.2 Caregiving in Singapore Caregiving in Singapore is seen to be very similar to caregiving in Western populations with similar concerns being held by both. However,. notion of Duty or Yuan was suggested as being instrumental in preventing negative appraisal leading to higher levels of EE, Wong (2000) found similar findings in a study of Chinese caregivers in. benefit finding and benefit reminding have also increased well being. Encouraging positive aspects, rather than trying to change negative aspects, may also be beneficial in increasing the family’s