In Chile, with a population of 16 million people, the prevalence of schizophrenia ranges between 1.4-4.6 persons for every hundred thousand people, with the annual incidence rate being 12 new cases for every hundred thousand inhabitants . The prevalence of this disorder is higher in the Arica-Parinacota region and in the capital, Santiago . Further, a National Health Survey carried out in 2004 recorded 917,939 Chileans (5.7% of the population) as being moderately/severely disabled, with 68.71% of these being taken care of by relatives .
At the beginning of the 1990s, and together with the return to democracy in Chile, mental health reforms began to be implemented and the country witnessed a progressive increase in government investment in the area, the development of psychiatric and mental health service networks, and the creation of user groups. Although these changes led to the emergence of a community-centred model of care, reform has been partial and unequal across the country. It was partial because these developments did not translate into adequate social services and benefits, despite the meaningful advances made in terms of healthcare. And it was unequal since large differences remain between different regions of Chile, with a greater need in the regions . In this regard, research conducted in the country's second largest city reported lower levels of caregiver burden, in contrast to the intense burden found among caregivers in the extreme north of Chile. This may be due to the fact that the mental health system in the north of the country has minimal resources for patient and family care [4, 5].
At all events the closing of mental institutions has enhanced the study of mental disorders from a family perspective. That is, research has focused on emotional and affective states not only in the patient's psycho-social environment but also in his/her home [6, 7]. It is in this context that burden needs to be understood. Dillehay and Sandys  defined caregiver burden as a psychological state that appears as a combination of physical and emotional work, social pressure, and financial restrictions which are consequences of taking care of a patient. This construct has been understood as any potentially verifiable and observable disturbance in the life of the caregiver that is caused by the patient's disease and which appears as a result of supervision and stimulation tasks [9, 10]. The psychological distress perceived by caregivers is related to multiple factors, such as the caregiver's personality, the patient's psychological symptoms, and the availability of social support and financial resources [5, 11–13].
From this perspective, the caregiver is conceptualized as an individual belonging to the patient's informal support system (a relative or friend) who takes responsibility for the main care tasks and who is perceived within the family as the person in charge of the patient, without being financially rewarded for this work . This person is usually referred to as the 'main caregiver', the one who dedicates most of his/her time (number of hours per day) to caring for the patient . With patients who have a severe mental disorder the role of main caregiver has usually been taken by relatives. As such patients cannot perform many of their personal and social duties caregivers have to stand in for them. Furthermore, they need continuous assistance with daily life, as well as frequent help in containing disturbed behaviours related to their disorder [4, 16–18].
One study with a Latin American sample  found that relatives of patients with schizophrenia showed high levels of burden as a result of the care task. The financial problems, the restriction of spare time, and the patient's future, were of considerable concern to caregivers. In this context it is helpful to distinguish between objective and subjective burden. The former includes any disturbance in the family environment due to the patient's disorder and which is potentially verifiable and observable. Operatively, objective burden refers to disturbances in work, social life, spare time, housework, finances and interactions both within and outside the family [19, 20]. By contrast, subjective burden refers to the feeling of being subjected to a heavy and oppressive duty, and to the subjective reaction of caregivers with respect to the tasks they have to perform .
Another important aspect relates to social networks. Research has demonstrated that caregivers are more likely to face restrictions in their social activities, thereby reducing their own social networks ; they may therefore remain isolated in their homes with few social contacts. Stigma about mental illness is also still frequent in many families and can contribute to social isolation [23, 24]. Indeed, guilt and shame seem to be related to discrimination as a form of social avoidance . Three large studies reported that between a fifth and a third of family members showed concern and maintained distant relationships with the rest of the family and friends because of the mentally disturbed relative [26–28]. By contrast, research has found that multiple roles among caregivers can reduce stress by expanding their available resources . For example, working outside the home seems to be a good shield against burden as it can provide access to social networks in the form of interpersonal interactions and social support . It has been shown that caregivers of people with a mental disability who spend more hours working outside the home experience significantly less stress in comparison with caregivers who do not .
In summary, these studies suggest that burden levels a) can change over time, b) are influenced by the nature of the patient's symptoms, and c) will probably not decrease without specific interventions .
As regards attitudes, these have been defined as "predispositions to associate certain kinds of stimuli with certain kinds of responses" . Such responses are classified as affective (assessing feelings as pleasant or unpleasant), cognitive (concerning beliefs, opinions, and ideas about the attitude object) and behavioural (concerning behavioural intentions or action predispositions). This leads to a three-component attitude model, in which attitudes constitute a hypothetical construct mediating between the observable preceding stimulus and the subsequent behaviour. The present research is based on this multi-component model.
With regard to the attitudes of relatives of patients with schizophrenia it should be noted that the social stigma attached to mental disorders contributes to feelings of frustration and anger. Families are forced to acknowledge the stark reality of having a member with schizophrenia and to mourn the loss of unfulfilled expectations. Moreover, as a result of the chronic stress associated with the task of caring, the family may experience a series of marital conflicts between parents or differences in relation to the other siblings . Thus, it is common for families to have emotional responses such as anxiety, fear, guilt, stigma, frustration, anger, sadness and so on. Furthermore, these family conflicts contribute to the stress experienced by its members [34–36]. Families may also develop other behavioural responses, such as adapting to the situation as if it were normal, resorting to prayer, finding meaning in the patient's communications, ignoring the patient's behaviour, or taking on additional responsibilities, and this can sometimes lead to overprotection.
Research on relatives' attitudes towards patients with schizophrenia began in 1959, when Brown and colleagues studied the effect of a patient returning home after hospitalization . They found that patients who returned to live with their parents or partners had a higher number of relapses than did those returning to another type of living environment. Subsequent research thus began to evaluate the relationship between caregiver attitudes, patients' social functioning, caregivers' quality of life, caregiver burden, and the effectiveness of family intervention [38–40].
The assessment of patients who returned home after hospitalization led to two main findings: a) close emotional ties are not always beneficial for patients; and b) the continuing close contact between the patient and the family may be associated with relapse. Interest in the notion of close affective ties saw research being conducted into expressed emotion (EE) in the family. This showed that in families of patients with schizophrenia that are characterized by high EE, the rate of relapse is higher (48%) than in those with a low EE (6%) . This research argued that although the aetiology of schizophrenia remains unknown, it is possible to talk about factors that precipitate relapse and which are linked to pharmacological and social influences (such as the family's EE). Initial work in this area reported that high levels of EE in the family were repeatedly found to predict rates of relapse in patients with schizophrenia nine months after discharge . A stressful family environment is an important risk factor, and the risk of relapse is about four times higher in patients returning to a family environment with high levels of EE. By contrast, low EE caregivers show greater acceptance towards their relative with schizophrenia, and this is related to lower levels of distress in both patient and caregivers .
Given the above it appears that the attitudes underlying the construct of EE, namely the criticism, hostility and emotional over-involvement (EOI) shown by relatives, have a significant impact on the course of the disorder, as well as on caregiver burden. However, research suggests that these attitudes may differ across ethno-cultural groups. Several studies found that Mexican Americans usually show lower rates of EE than do European Americans [44–47]. This research emphasizes that Mexican American caregivers are less critical, less hostile and show more warmth towards their relatives with schizophrenia than do Anglo-American caregivers. Other studies have suggested that ethno-cultural differences also affect the nature of the association between EE and relapse in patients with schizophrenia and the health status of caregivers [48, 49]. Thus, while criticism is considered the main contributor to patient relapse and caregiver burden in Anglo-American samples, EOI seems to be the main predictor of worse health outcomes and burden in Mexican American caregivers . Latin Americans generally show high levels of engagement with their ill relatives and usually understand and empathize with their problems. However, Breitborde and colleagues  suggested that only moderate levels of EOI were associated with better outcomes, both in patients and caregivers. When EOI is too high or too low, patient relapse and caregiver burden rates increase again. All these studies highlight the need to take into account ethno-cultural aspects when studying caregivers' experiences.
Another study reviewing the relationship between family attitudes and the social functioning of patients with schizophrenia  found a significant correlation between an empathic attitude and the patient's social and occupational functioning. Caregivers who show more tolerant, non-intrusive and supportive attitudes towards patients help them to achieve better social functioning. Thus, more empathic caregivers can adjust their behaviour according to the patient's emotional state and needs, which protects them from extreme reactions that could eventually trigger a relapse. Furthermore, these empathic attitudes promote the generation of constructive ideas and facilitate the resolution of problems, which can help identify social opportunities. The authors of this study suggest that improving caregiver attitudes must be a part of family intervention programmes. Recently, Gutiérrez-Maldonado and colleagues  assessed the efficacy of a family psycho-educational programme designed to change attitudes and health perceptions in relatives of patients with schizophrenia. The results showed that the programme was effective in modifying caregiver attitudes, although it had no effect on their health perceptions. The programme also increased levels of satisfaction regarding the course of the patient's disorder .
Based on these outcomes, greater emphasis is now placed on psycho-educational interventions, since they not only provide training, emotional reinforcement, information and social support, but also improve relatives' attitudes towards schizophrenia, which in turn impacts on the caregiver's quality of life [51, 54–56]. Dixon and colleagues  found that a family intervention programme led to significantly greater family, community and health service empowerment, and also reduced displeasure and concern about the family member with a mental disorder. Group-based interventions enable families to share experiences with others in similar situations, which can provide comfort and facilitate the expression of feelings about the disorder, thereby improving coping skills. It has been suggested that such intervention groups also increase the motivation of family members involved. Finally, it should be noted that psycho-educational programmes do not only provide information, but also reinforce the idea of respect for families and encourage them to consider themselves as co-therapists in the process. In this way, the therapeutic team and the family can develop a less polarized and less stressful relationship, and even more reluctant family members become more willing to cooperate, thereby reducing the burden on both parties .
Multifamily interventions have also considered the attitudes of caregivers. Weissman and colleagues  found a significant relationship between caregiver emotions and supportive behaviours toward the patient, and when caregivers showed a more sympathetic attitude towards patients they also made a greater effort to help them cope with their condition. These authors also conducted a qualitative analysis of the beliefs, values and behaviours of family members in order to gain a better understanding of their low EE. They found three main categories of attributes: a) the majority of caregivers assumed that the patient had a legitimate mental disorder; b) external environmental stressors were regarded as the cause of the condition or its worsening; and c) many caregivers implicated God in their attributions about the disorder and indicated that religion had become a source of hope about the situation. The authors concluded that the promotion of positive and favourable emotions among caregivers could be a key step in terms of establishing a low EE environment. At all events, the limited expression of negative affect reported in this sample should be seen as reflecting the Latin American cultural tendency towards social desirability, which together with a suppression of negative affect may be a factor that contributes to favourable outcomes in the patient .
Despite all the above findings there is still a gap between outcomes and service implementation. Moreover, this difference is more marked in developing countries that are characterized by different cultural values and a scarcity of available resources [59, 60]. Given the importance of this issue, which is considered a priority area by policy makers (including in Chile), and the lack of research in different cultural contexts, the aim of the present study was to assess attitudes and burden among caregivers of patients with schizophrenia, in this case, patients who were users of public mental health services in Arica, Chile. The first step involved developing the Attitudes Towards Schizophrenia Questionnaire for Relatives, which covers three attitudes components: emotional, cognitive and behavioural. We then took into account that previous studies have found that the schizophrenic relative's symptoms affect both caregiver attitudes and perceived burden [61, 62]. Furthermore, several authors consider that contextual variables, such as demographic characteristics of caregivers and relatives with schizophrenia, may modulate the effect of patients' symptoms on the emotional state of caregivers [6, 63]. In light of this, a second aim was to assess differences in the attitudes of caregivers according to the social and demographic characteristics of both caregivers and patients. Specifically, with regard to patient characteristics, we expected that symptom severity, the number of hospitalizations and the receipt of a disability pension would affect caregiver attitudes. As regards caregiver characteristics, we expected that attitudes would be influenced by working outside the home (employment). Finally, a third aim of the study was to assess the relationship between caregiver attitudes and the burden they experience. We expected that more negative attitudes would lead to greater perceived burden.