Previous research has concluded that relatives of a person with a severe mental illness (e.g. schizophrenia and bipolar disorders) experience burden [1] - [3] . The concept of burden is often operationalized into objective and subjective burden [4] , in which objective burden is the observable consequences of the patient’s illness, such as financial problems, restrictions on the relatives’ leisure activities and on employment. Subjective burden is the psychological consequences for the relatives, such as a feeling of guilt for causing the illness, a feeling of being trapped in the caregiver role and worrying about what the future will bring [5] - [7] .

This challenging life situation may influence on the relatives’ own health [2] [3] and may influence their ability to manage their caregiver role. In this context, health can be understood as the ability to cope with stress, also known as Salutogenesis. Salutogenesis is strongly connected to sense of coherence (SOC), which is defined as a global attitude that expresses the extent to which a person perceives events or the strains of life as comprehensible, manageable and meaningful [8] . A strong correlation has been reported between weak SOC, high burden and poor mental health [3] . Moreover it is considered that a strong SOC is a coping resource to protect health [9] . SOC may therefore be of importance for relatives’ ability to master the consequences of their next of kin’s mental illness.

Worldwide, there are an increasing number of people who suffer from depression, which also affects many relatives [10] [11] . The main symptoms of severe depression are lowering of mood, reduction in energy, decrease in activity and reduced capacity for concentration and enjoyment. Reduced self-esteem and self-confi- dence are often present, as are feelings of guilt or worthlessness [12] . Many patients with depression receive treatment while staying at home, but hospitalization is sometimes inevitable when a major depression seriously interferes with a person’s life, and is also one of the primary risk factors for suicide [13] .

Qualitative studies have found that relatives of patients with depression experienced a great responsibility for the depressed person, a loss of control in their everyday life, that the relationships in the family are affected and that roles are changed [14] - [16] . The importance of gaining more knowledge about the impact on health and the perceived burden in relatives of patients suffering from depression are highlighted [17] - [19] . Studies with a quantitative design, including this group of relatives with regard to burden and health, have found that relatives experience burden and that their health is affected, but these studies are characterized by small samples [17] [20] [21] , and that the relatives were recruited through outpatients clinics [17] or that the depression diagnosis was not defined [19] [22] .

As has been shown, studies have reported the consequences of being the relative of a person with a severe mental illness. Being the relative of a person with depression in general is described as being a challenge, and to the best of our knowledge, studies of being the relative of a person suffering from severe depression in need of hospitalization have been scarce. For this reason, the aim of the study was to describe burden, health and sense of coherence with regard to background variables among relatives of inpatients with severe depression. Furthermore, the aim was to investigate relatives’ burden in relation to their health and sense of coherence.

Of the 68 participating relatives, there was a predominance of women. The majority of the relatives were between the ages of 41 - 59, most were married or cohabiting and half the relatives had upper secondary school. According to their relationship with the depressed person, nearly half were adult children. In addition, less than half of the relatives lived together with the depressed person (Table 1).

The relatives’ background in relation to the mean scores of BAS, GHQ and SOC is described in Table 1. Males reported lower burden and mental distress, as well as stronger sense of coherence, than females. With regard to age groups, relatives between the ages of 40 and 59 reported the strongest sense of coherence. Considering educational level, relatives with upper secondary school reported the lowest burden and strongest sense of coherence; whereas those with compulsory comprehensive school reported highest level of mental distress and weakest sense of coherence. With respect to their relationship with the depressed person, adult children had highest burden, while spouses/cohabitants reported highest mental distress. Relatives who did not live together with the person suffering from depression reported highest burden, whereas those who lived together reported highest level of mental distress. No significant differences between the relatives’ background and their burden, sense of coherence and mental distress were found.

¹Burden Assessment Scale. Sum score ranged from19 (no burden) to 76 (a lot of burden). ²General Health Questionnaire. Sum score ranged between 0 (most favourable) and 12 (least favourable). ³Sense of Coherence scale. Sum score ranged from 13 (weaker) to 91(stronger).

The number and percentage of relatives who reported “Some” or “A lot” of burden at item level are listed in Table 2. The four items most frequently reported were: “Worry about future” (82%), “Upset by change in patient” (58%), “Less time for friends” (55%) and “Reduced leisure time” (54%). The least frequently reported burden was: “Guilty for causing illness” (10%), “Financial problems” (12%), “Friend frictions” (13%) and “Embarrassed by patient” (16%).

The relatives were divided into three groups with respect to the total mean score of BAS. The three groups were compared for differences concerning GHQ, SOC, physical health and mental well-being, and there were significant differences between the groups. Significant differences were noted between relatives in group A and relatives in group C, with the relatives in group C reporting higher burden, more mental distress, weaker sense of coherence, and poorer physical health and mental well-being. Furthermore, the relatives in group C reported weaker sense of coherence than group B (Table 3).

¹Subjective burden; ²Objective burden.

¹Burden Assessment Scale (BAS). Sum score ranged from 19 (no burden) to 76 (a lot of burden). ²Post hoc test between pair of groups, showing p-values. ³General Health Questionnaire. Sum score ranged between 0 (most favourable) and 12 (least favourable). ⁴Sense of Coherence scale. Sum score ranged from 13 (weaker) to 91(stronger). ⁵Physical health and Mental well-being; response scale ranging from 1 (Very bad) to 5 (Very good).

The aim of the study was to describe burden, health and sense of coherence with regard to background variables among relatives of inpatients with severe depression and to investigate relatives’ burden in relation to their health and sense of coherence. This study revealed variations between relatives, with some reporting a greater burden, poorer health and a weaker sense of coherence than others.

The severity of the symptoms is often reported as the factor that places the greatest burden on the relatives of persons with severe mental illness [38] . In this study the patients were hospitalized because of depression, which indicates that the symptoms were severe; otherwise an outpatient treatment would have been given. Van Wijngaarden et al. [39] found few differences between the relatives of depressed outpatients and the relatives of persons with schizophrenia. Hence, the relatives of depressed persons more often experienced a strained atmosphere. Compared to the relatives of outpatients diagnosed with depression, the relatives of inpatients reported more extreme tiredness, sleeplessness and a lack of appetite [20] . With regard to subjective burden eight out of ten relatives reported “Worry about future”, and almost six out of ten were “Upset by change in patient”. This is supported by other studies of relatives to mentally ill persons with schizophrenia [40] [41] and dementia [42] . What the relatives worry about concerning the future may be dependent on their relationship to the patients [43] .

Regarding objective burden, more than half the relatives reported having “Less time for friends” and “Reduced leisure time” more frequently compared to another study of relatives of patients with severe mental illness [43] . Relatives who did not live in the same household as the person suffering from depression reported higher burden than those relatives who shared the household. Not living together with the person suffering from depression could require planning in terms of taking care of the person, and relatives might also worry about what happened to the person when they were apart. The adult children were most represented in this group and they may have their own family to take care of, in addition to the depressed parent, which may reduce their amount of leisure time.

Almost half of the relatives in the present study were adult children of the depressed person, and they reported the highest burden. Children, including the adult children of a parent with depression, face a demanding situation and the experience of being controlled by their parents and unable to live the life they wanted [16] . Hedman Ahlström et al. [44] reported that anxiety and concerns for the depressed parent affected children, even after they moved out. Changes in the depressed person from prior to the onset of depression often affect the interpersonal relationships and changed the roles in the family. The relatives may also worry about that their next of kin could commit suicide [16] .

The relatives’ “Worry about future” and being “Upset by change in patient” may be attributed to the lack of knowledge about depression and what they can expect from treatment and the course of treatment. Knowledge about depression is of importance when relatives are attempting to understand and engage with the depressed next of kin in an appropriate manner [16] . Gaining more knowledge about what the relatives can expect concerning the diagnosis and the course of treatment could help to make the future more predictable. The nurses play an important role in this, supporting the relatives in finding the best ways to meet the challenge and create meaning in this situation.

Some of the relatives in this study reported poorer health than university students in a Norwegian study [29] and a population sample in a Swedish study [34] . Almost four out of ten relatives were spouses/cohabitants of the depressed person reported a high GHQ mean score. A Norwegian population study [19] reported significantly higher levels of symptoms of depression and anxiety and a lower degree of subjective well-being among spouses of persons suffering from anxiety and depression disorders compared to spouses to persons not suffering from these mental disorders. Depression is found to influence communication between spouses in a negative manner [45] and to provide care for the depressed spouse in addition to holding down a job and taking care of other family members may be a challenge [16] .

Sense of coherence is found to be a resource with an impact on how people deal with stressful life events and how they mobilize their resources in a health-promoting manner [46] . The SOC mean score among the relatives in this study is somewhat higher than in a Norwegian sample of relatives of persons with severe mental illnesses [3] . On the other hand the relatives in the present study showed a weaker SOC compared to a Swedish population study [34] . Critical life events may decrease the level of SOC [47] and a weak SOC represents a health risk, while a stronger SOC is related to better health [46] . Antonovsky [8] describes General Resistance Resources (GRR), e.g. social networks, social support, self-esteem, healthy behaviour and knowledge, all of which are factors of importance when handling difficult life situations. GRR may help the relatives to handle the situation.

Nurses have a responsibility to promote health [48] , which requires commitment and a desire to try to understand the uniqueness in the situation of the relative. When this study revealed that relatives differ in their extent of perceived burden, there is probably no single intervention that could address this heterogeneity. When nurses have knowledge about which factors are of importance to help cope with demanding life situations, they may be able to identify relatives at risk of developing their own health problems.

The present study has both strengths and limitations that have to be noted. One strength is the clear inclusion criteria with reference to the relatives of inpatients and a further diagnosis of depression according to ICD-10. This stands in contrast to other quantitative studies, which primarily refer to the relatives of patients with depressive disorders/depression in general [19] [22] .

The response rate in this study was 75% which may be seen as a strength [24] . The first author visited all the wards in order to provide information about the study and have regular contact by phone and e-mail with the health personnel recruited to remind them and to answer any questions concerning the recruitment.

During the recruitment period, a total of 223 patients were admitted with depression diagnoses; of these, 125 patients were asked for permission to ask one of their relatives to participate in the study. Those patients may have been excluded because of not having the capacity to give informed consent, having relatives under the age of 18 and/or not being able to speak and understand Norwegian. Another reason for exclusion was when patients only had formal caregivers or had conflictual relationships. The data collection was carried out over a 24-month period from 2010 to 2012. Two festive seasons and two summers are included in the data collection period, and some patients may have been overlooked or been forgotten to be asked because of periods with temporary staff. The sample size was relatively small and was recruited from only one hospital trust in Norway; therefore, generalizability must be interpreted with caution.

The recruitment procedure was comprehensive in that the patients were asked for permission to ask their closest relatives to participate in this study, which was a requirement from the Regional Committee for Medical and Health Research Ethics. Furthermore, no information about the patients was allowed. Lack of information about the patients’ symptoms and the duration of the depression may be seen as a limitation. However, what we know is that the depression was serious enough to require hospitalization.

Both the SOC and GHQ instruments are well known and frequently tested for validity and reliability in multiple samples. The Cronbach’s alpha coefficients in the present study showed a good reliability concerning SOC, GHQ and BAS. The BAS was chosen because it measures both objective and subjective burden. The Norwegian version of BAS was pilot tested by a panel of researchers in mental health.

The relatives of inpatients with severe depression reported burden to a various degree. Subjective burden was more frequently reported than objective burden. The relatives with higher level of burden reported more mental distress, poorer health and weaker sense of coherence than those with a lower level of burden. Further research should focus on identification of factors predicting burden and health of relatives of inpatients with severe depression. There is also a need to investigate the relatives’ experiences of support from the psychiatric specialist health services.

We wish to extend our sincere thanks to the relatives who participated in this study, as well as thanks to the staff of the nine units who recruited the relatives for this study. We gratefully thank statistician Jari Appelgren for his statistical advice.