The understanding of Norwegian women’s sickness absence: towards a holistic approach?

Background: In Norway men’s sickness absence from work has been stable or reduced over the last decades while women’s sickness-absence ratio has increased, but the reasons for these developments are complex and unclear. There have been considerable efforts introduced and implemented to reduce sickness absence, but they have not succeeded. One reason for this may be the insufficient knowledge about the reasons for sick leave, and especially for women’s sick leave. The aim: This article aims to examine how social factors influence sickness absence and how long-term absentees interpret and explain their ill health and sickness absence. Method: In one Norwegian county in 2010 we performed individual in-depth interviews with 20 women and ten men between the ages of 25-60 years who had been or were sick-listed for more than 30 days during the last year with a mental illness or musculoskeletal diagnoses. Results: The study illustrates how social factors influence sickness absence in different ways. The women indicated complex causes for their sickness absence, and often described an interaction between work-related and domestic-related aspects. Some accounts illustrate that their ill health might have roots in life occurrences from childhood and adolescence that have made them vulnerable to domestic and work-related strains during their adult years. The study also indicates that women, especially single mothers, seem to be especially vulnerable to domestic strains, and that these strains may lead to a paradoxical pattern of women’s sick leave: they take sick leave in order to deal with domestic strains along with the intention of prolonging their presence at the workplace in the longer term. Thus, these periods of sickness absence appear to be a necessary accompaniment of a high rate of participation among vulnerable groups in the labour market. Conclusion: Women’s ill health and sickness absence should be understood as a manifestation of an interplay of the strains found at both the workplace and the home. A successful effort to reduce sickness absence in Norway, therefore, requires a holistic perspective that accounts over the last years have their origins from my adolescence. Signals of negativity and criticism: what I did was never good enough. Some of this has been reinforced these past years, and I have to take this seriously and look forward to a better future.


Introduction
Since 1972 the total sickness absence in Norway has steadily increased and shows no signs of abating (Dale- Olsen & Markussen, 2010). Since 2001 the reduction of absenteeism has been a central, highly publicized political aim, but these efforts have not been successful (Ose et al., 2009). Moreover, an overall understanding of why these efforts have not been successful remains elusive and incomplete. There is no indication, for instance, that demographic developments or changes in the health situation in the population can explain the increase in sickness absence (Bjørngaard et al., 2009). Yet what is perhaps more striking is the difference between men and women, an issue that has only recently come to the fore of the political agenda. The statistics shows that among men sickness absence has been relatively stable (Blekesaune & Dale-Olsen, 2010;Dale-Olsen & Markussen, 2010;Nossen & Thune, 2009). Thus, women's rate of sick leave is responsible for most of the general increase. A similar gender difference in sickness absence has also been found in Sweden (Alexanderson & Norlund, 2004), and an international comparison of Canada and eight European countries has found that in most of these countries, women had higher rate of absence than men (Barmby et al., 2002). It appears that men's rates of sickness absence have been influenced by fluctuations of the labour market (Dale- Olsen & Markussen, 2010), but little is known about the factors that affect women's rates. A conclusion from a recently published review of quantitative studies on women's sick leave is that there is a general lack of knowledge about this phenomenon, and that improving our understanding can be the key to achieving reduced sickness absenteeism in Norway (Kostøl & Telle, 2011).
One reason for the lack of success in reducing sickness absence may be that the efforts have not been directed towards the main reasons for the sick leave. Specifically, the main policies aimed at reducing sickness absence have until now been directed towards the workplace only (Ose et al., 2009). This article suggests that this emphasis leads to certain misunderstandings because women's total work burden is related both to the workplace and the domestic sphere, and that both the remedial efforts and the research are deficient due to this oversight. The efforts to reduce sickness absence have only to a limited degree taken into consideration the more complex work burden that women bear, and the research has contributed little to the understanding of the complexity and women's work and life situation. If the factors remain largely unknown, then it will be naturally difficult to establish what might be the most effective remedies.
This article aims to examine the social factors contributing sick leave and how the absentees explain and interpret their ill health and sickness absence. The focus of this article is mainly on the experiences of the women and on the reasons for their sick leave. We shall refer to the sick-listed men only as a contrast.
There are a number of studies that have emphasized the role of social factors in sickness absence. Kostøl (2010) has presented some of the popular hypotheses and a review of this literature. One of the hypotheses has been that women with children carry a double burden and are therefore more likely to take sickness absence. Quantitative studies have, however, found little or no support for this assumption (Mastekaasa & Olsen, 1998;Mastekaasa, 2000;Bratberg et al., 2002). Yet Bjørngaard et al. (2009), among others, have concluded that women probably to a larger degree experience a role conflict between work and care for children, and that women often have more physically demanding work situations. Consequently, they argued that it is urgent to maintain the gender perspective in this area of research, where studies need to look at men's and women's situation separately.
This article presents a qualitative study as an alternative approach to understanding the causes of sick leave. Qualitative studies offer an important complement to other approaches because they allow the sick listed to reflect on their situation and to express their experiences in their own words. Individual and personal understandings form the basis for the single person's actions and decisions. The qualitative data makes it possible to gain an insight into how people cope with their illness and the relation between illness and sick leave.
The next section is a short presentation of this study's theoretical orientation, and a brief review of the recent literature on the social risks for long-term sickness absence follows. The methods and the data material of the study are described before an account of the findings is given. The last section discusses these findings and on this basis reflects upon the consequences on policies and policy-making.

Theoretical perspectives
There is a general furcation in the perspectives for understanding health and illness: the medical and the sociocultural along with the psychosocial (Coutu et al., 2007). The medical perspective regards people as natural entities and tries to find exact indicators for diagnoses. This approach seeks to draw objective conclusions. In contrast, the sociocultural perspective sees people as cultural and social beings, and here the relation between the individual and society is important. Similarly, the psychosocial perspective provides insight into the individual's interactions with his or her environment. The last two perspectives focus on how individuals interpret their illness and how they cope with the situation. This study likewise also pays attention to the creation of sick-leave identities as a function of the social interaction between the sick-listed and their social networks. According to this perspective, the state of illness and disease that leads to sickness absence has a relational character, and is seen as an interactive process (Verdonk et al., 2008). To understand sickness absence, it is necessary to take the absentees' psychosocial environments into account, and it is important to clarify how absentees think about themselves and their life situation . Therefore, this study adopts a symbolic interactionist perspective which rests on three premises (Blumer, 1969:2): a) human beings act towards things on the basis of the meanings that the things have for them; b) the meanings are derived from the social interaction that one has with one's fellows; and c) these meanings are handled in and modified through an interpretative process that one uses in dealing with the things that one encounters. Included in this perspective is an explorative and holistic approach, a step-wise induction from data, and ideographic knowledge focusing upon an understanding of single cases (Alvesson & Sköldberg, 2005). For my purpose, this perspective suggests that an individual's interpretations of the sick role and the reactions from the social surroundings may influence recovery. It is important, therefore, to pay attention to how sick-leave identities are created as a function of the social interaction between those sick-listed and their surroundings, along with general attitudes towards sick-leave in society.
To understand how men and women adapt and assert themselves, Harding (1986) makes a distinction between three different levels: the individual, structural and symbolic. While the experience of men's and women's freedom of action is formed at the individual level, their adaptations are influenced by the gendered structures in society as well as by the imaginations and understandings of gender that are part of our society, that is, the symbolic level. Illness is a personal experience of suffering (Hoffmann, 2010), and this experience has consequences for how the individual adapts to the situation and what actions are taken. By focusing on gender, this study assumes that factors other than on the structural and symbolic levels also are central in men's and women's understanding of their sickness absence, and that workrelated and domestic strains may have varying degrees of significance for developing ill health.

Social risk factors for long-term sickness absence
The connection between work-related health problems and illness along with sick leave is well verified (Mehlum et al., 2006;Deckers-Sanchez et al., 2008;Tynes et al., 2008;Gravset, 2010). Work-related illness is often defined as follows: a) Individual factors or injuries, which comprise physical or ergonomic factors due to heavy physical work, awkward work postures and repetitive work, and b) social factors, which cover psychological demands, job control, social support at the workplace, and organizational factors (e.g., shift work). One important aspect of the social factors is the imbalance between the individual's resources and the demands of the job (Westin, 1994). Other studies have confirmed the significance of such psychosocial factors as job control, social support from leadership, predictability and role conflicts (Lund & Labriola, 2009;Nielsen et al., 2006). The significance of the different factors varies between men and women, especially the psychosocial factors (Lund & Labriola, 2009). Ockander and Timpka (2001) have described the complex influence of work on health. This study shows that work has different meaning for different individuals depending on their life situation and where they are in the life cycle. It is important, therefore, not to place women into one category because there are wide variations in women's relation to work and how they adapt to the labour market. This is confirmed in a Dutch study which shows that work-related fatigue is clearly related to gender, and that this condition was highest among highly educated female employees (Verdonk, 2010).
Irrespective of the amount of paid work, it is common that women bear the main responsibility for domestic work. A Norwegian study reveals dramatic gender differences in the distribution of housework and care, even among the households of elite men and women (Skjeie & Teigen, 2003). Most of the elite women had partners who worked full-time or more, while 40 per cent of their male counterparts had partners who worked part-time or were housewives. A high total workload of paid and unpaid work has been found to increase the risk of negative outcomes (Krantz et al., 2005;Mellner et al., 2006).
Few studies have examined the separate impact of domestic work on women's health, but there are some notable exceptions (Staland-Nyman et al., 2007;Chandola, 2004;Glass, 1994;Griffin et al., 2002;Östlund et al., 2004). Staland-Nyman et al. (2007) have found an association between strains in domestic work measured as 'domestic job strain' and 'domestic work equity and marital satisfaction', on the one hand, and self-rated health, on the other. The importance of perceived equity in domestic work and a satisfactory relationship for self-rated health was the strongest for the dimensions that measured psychological health. Also, findings from other studies (Glass, 1994;Griffin, 2002) have revealed an association between factors related to domestic work and women's psychological well-being. These studies have found that the inequity in the division of domestic work to be a greater contributory factor to women's psychological distress than the amount of domestic work. A considerable proportion of Norwegian women are working single mothers. In 2011 22 per cent of the households in Norway with children in the ages 0-17 years were single-parent families, and 80 per cent of these were single mothers (Statistisk sentralbyrå, 2011). The dual responsibility of caring for and providing for children causes work-family conflicts for many single mothers as well as financial stress (Whitehead et al., 2000;Ugreninov, 2005). Some research has observed single mothers having higher levels of mental problems than non-single mothers (Whitehead et al., 2000;Fritzell & Burstrom, 2006). A study of single and non-single working mothers in Denmark, Sweden and Norway (Bull & Mittelmark, 2009) shows that financial stress was the most significant predictor of life satisfaction in both groups of working mothers, but also that the level of financial stress is significantly higher for single mothers. Single mothers also scored lower on life satisfaction and happiness.
Quantitative studies often use the number of children as the only variable for measuring domestic strain (Mastekaasa, 2000). The literature shows, however, that this method is insufficient and that the relation between work strain and domestic strain versus Illness and sickness absence is complex. A crucial question for women's health, then, is how studies might disentangle the respective burden of illness associated with paid work, domestic work and the combined effect of both. These studies illustrate the significance of making a distinction between different aspects of domestic work, and point to the importance of using measures that include both practical and emotional aspects. This underlines the importance of having a holistic perspective to grasp the different aspects and the complex interplay between the different factors. This present study is also concerned with identifying the variations of experiences and understandings between different groups of women related to their work situation and their situation in the domestic sphere.

Methodological approach and sampling
This is a qualitative and exploratory study. It was critical to the study that the data was collected with an open mind in order to record the informants' own thoughts and explanations in their own terms. Our study adopts the central method of symbolic interactionism of seeking to understand the phenomenon from the informants' perspectives and to describe their experiences. In addition, our analysis employed a grounded-theory approach (Glaser & Strauss, 1967;Glaser, 2010) because the inductive nature of grounded-theory methods requires openness and flexibility on the part of the researchers. In this way the themes are permitted to emerge from the informant's account (Charmaz, 2003). Our study covered four topics in all interviews, but we did not formulate the questions in advance in order to apply the flexible approach to each interview.
The interviews were tape-recorded and transcribed. Batt-Rawden and Solheim (2011) have given the first presentation of the data material, which also gives a more extensive presentation of the methods used. The analysis began with open coding in order to capture the concepts, themes, and statements that were central in their descriptions. In the next stage of selective coding, we limited the categories and showed the dimensions representing the variations within the data material. Thus we examined the data in order to identify themes, to explore categories, and to develop concepts.
The study is based upon a sample of long-term sick-listed persons' own explanations and understandings of their sickness absence due to mental problems or musculoskeletal problems, the two most frequent groups of diagnoses among people on sick leave in Norway. In 2010 among the men, musculoskeletal problems caused 46.9 per cent of the days on sick leave, and mental problems 17.5 per cent (NAV, 2011). Among the women, 37.4 per cent were sick-listed because of musculoskeletal problems, and 20.3 per cent because of mental problems. With both of these diagnoses it is often difficult to find objective symptoms, and often the background to these problems is complex. In addition there often is an overlap between these two groups of diagnoses. A survey in Sweden shows that two-thirds suffer both musculoskeletal pain and mental problems (Eriksson et al., 2008).
The informants were contacted according to the instructions set by the project SOFAC in accordance with the Regional Ethical Committee (REK). The Department for Statistical Analysis at the Norwegian Labour and Welfare Service (NAV) selected a sample of 260 women and 260 men diagnosed with mental illness (ICPC-codes: P76, P02, A04) or musculoskeletal illnesses (ICPC-codes: L03, L84, L02, L18, A01, L92). The sample included people between 20-60 years old, who were or had been long-term sick-listed for at least 30 days during the previous year. To reduce travelling time, we limited the sample to long-term sick-listed individuals in ten municipalities in one Norwegian county. The purpose was to find a broad representation of age span (20-60 years old) and men and women in different life situations.
A letter was sent to the sample with information about the study. To comply with the rules concerning anonymity, NAV sent the letter and the informants themselves were asked to contact the researchers about participating in the study. In total, there were 30 informants, ten men and 20 women, who made contact and were interviewed. Our initial aim was to gather a sample of 20 women and 20 men. However, the lower number of men was not critical due to the exploratory nature of this study.
The low response rate might have been connected to the procedure of selection, which relied heavily on the informants' initiative. Owing to ethical guidelines, NAV were required to distribute letters that provided the necessary information to potential informants, and then it was up to the informants to initiate the first contact with the researchers. Another reason is that the group asked to participate is composed of people who suffer from serious illness, and being sick-listed they might lack the motivation or the strength, or both, to participate. We do not know the reasons why so many declined to participate in this study, but there were certain revealing characteristic of the group that was willing to share their experiences. While most of them have had long-term and repeated experiences of being sick-listed for several years, at the time of the interview they had recovered and they often were willing to participate because they had experiences from their period of sickness absence that they wanted to convey.

Characteristics of the sick-listed women
The interviewed women were between 25-60 years, and two-thirds of the group were in their forties (8) and fifties (7). Only three were in their thirties and two in their twenties. For most of the women the sickness had developed over a period of years, and they had been sick-listed two or more times before. They are a group with long and varied experiences with being sicklisted. Among the women nine of 20 had diagnoses related to mental problems, and 11 had diagnoses related to musculoskeletal problems. However, some of the women who were diagnosed with mental problems had musculoskeletal problems earlier, and vice versa. This shows that there often is an overlap between the two diagnoses, and therefore we do not make a systematic distinction between the two groups of diagnoses in the presentation of the data.
The women were relatively well educated, with 12 of them having attained higher education (at least three years at college or university), four vocational training, and four basic schooling. At the time of the interviews 12 of the women were married or cohabitants. Of these, eight had lived in a stable marriage for years, and four were divorced and had established a new relationship. All of the women had children, and ten had been or were in practice single mothers for their children. In this last group the fathers lived far away and did not participate in the daily lives of their children.

Results: Women's holistic understandings of their sickness absence
This study confirms that the reasons for people's sickness absence are complex. Our study shows that individuals often interpret their illness and sick leave as a function of the social interaction between the sick-listed and their social surroundings.
Among the important elements of this interaction are the gendered structures in Norwegian society. The interpretation presented here is how these structures, which enforce the expectations towards men and women in society, influence these individuals' adaptations. The reasons that the respondents gave for their sickness absence included work strain, domestic strain and the interplay between the two. Another important aspect among some of the women was the incidents of strain in their life histories in their childhood or adolescence.
In the following I shall first present the dimensions of work strain, and then the interaction between work strain, domestic strain and relational strain in their life histories that seem to have an influence on their ill health and sickness absence.

Vulnerability to work strain
All of the men in our study were full-time workers, and they all described their sickness absence entirely or mainly to be connected to factors at work. The women also had a strong identity of being working women, but some had parttime jobs, and during their life span they had adapted their work to obligations towards their families. Accordingly the amount that the women worked varied. The most common work-related dimension mentioned by the women was illhealth and signs of wear after years of pressure. Care workers are an especially exposed or vulnerable group in this respect. Lise, a woman working in a nursing home, expressed the following: … I think the sickness absence is a result of the expectations placed upon us. It is about being women in typically women's jobs. The job is demanding both physical and mentallyyou are expected to cope with everything, to do it all and to have time for everything. By increasing the number of workers I think the sick leave would have decreased. And then people would have had a much better life. It is only women who would have accepted this situation. … During the day we are not entitled to a break. We have our meals together with the patients and if they need help we are accessible.
Other dimensions of work-related strain that the women mentioned were workplace conflicts or inefficient leadership and accidents or injuries. Another work-related reason for sick leave was the boundlessness of their jobs, which entail expectations of being accessible for extended work time. For some of the women, this condition led to feelings of inadequacy and of having limited control in their work situation.

The complexity of strain: The interaction between work-related strain and domestic strain
However, these work-related explanations the women give for their sickness absence are often not the only explanation of their sickness absence. Their histories of ill health and sick leave often involve descriptions of the struggle to deal with both domestic strains and strains at work, and in some cases also the strain in their life histories.
Some women attributed the total burden of all these three elements as the explanation of their sick leave. One example is Gunn, who had been sicklisted several times in recent years. She is married, has two small children and works in a nursing home. She said that the total work burden represented too much care work. But the background for her sick leave was more complex. In addition she explained that her sickness absence also was related to her deteriorating relationship with her husband, her parents' recent divorce and her unsupportive mother during her adolescence and childhood. She expressed the expectations that: ... you have to be strong, you will not say anything. If you aren't managing things well, you have to say that everything is ok. The origin of these attitudes comes from my adolescence. My mother was often negative. I think that many of my problems that have been reinforced over the last years have their origins from my adolescence. Signals of negativity and criticism: what I did was never good enough. Some of this has been reinforced these past years, and I have to take this seriously and look forward to a better future.
When the men in a few cases mentioned aspects of domestic life as influencing their sick leave, they connected it to unexpected events in the families, like a sudden death or family members having had serious accidents or serious financial problems. The dimensions mentioned by the women were very different. One aspect was the workload and challenges related to helping and supporting their children or other family members. Other aspects were the feelings of having the overall responsibility along with a lack of practical and social or emotional support. Our data indicates that the women still have the role as the head of the domestic life in the families. The burden of this role becomes particularly pronounced in problematic situations. Anna, for example, who was divorced and had been a single mother for her two children, was living with a man with two children his own. Her son had been addicted to heroin for some years and the daughter of her partner developed serious mental problems, and after some time committed suicide. In these circumstances she described her role in the family as follows: Somebody had to keep the family together, so I compromised my own needs … I had to take care of my partner and my son. In my situation it has been impossible to take care of myself, even if somebody had asked me to take a break. I have been doing this continually for a long time … Dealing with the strains at work and in the domestic sphere was a challenge for all the women, but it was particularly expressed by women living alone with their children. A 'single mother' is normally defined as a woman who lives with children who are neither married nor living with a partner. In our study most of the single mothers are divorced and living alone with their children. The fathers are absent for different reasons, most often because they live too far away and are therefore not able to participate in the daily lives of their children. Women married to men who work abroad for long periods may be in a similar situation, since their husbands are absent most of the time. The single mothers related the lack of space for themselves and the sparse opportunities for recreation. Their jobs together with the additional demands in the domestic sphere took all their time and strength. The children's extracurricular activities could be particularly draining. When children participate in sports or other activities, there are strong expectations towards the parents to contribute and show their support The single mothers reported little understanding from the other parents for their situation as an only parent. Without support from family and friends this life situation is demanding. Dagny described her life situation like this: You want to manage and you have a responsibility for your job. I work with clients and they need help. I was working 80 per cent [of the amount of a full-time job] until the summer and I clenched my teeth. And I worked and cooked for my childrenthat was the life I lived. A very boring life, but it had to be like this. In the autumn I asked to increase my work to 100 per cent because I needed the money. Then I could not be sicklisted immediately after that, so I had to stay in work even if it turned out to be too much for me, but I managed until February.
In this statement it is clear that Dagny's economic situation is an additional strain. This was also the case with the other single mothers. Their financial situation sometimes forces them back to work too early. Women who worked part-time told stories about how they pushed themselves to take on a higher percentage of work than they could manage. The motivations for this were their need of a higher salary, and, if they earned a higher salary, they would be entitled to a higher sickness benefit. For most of them this was only a short-term solution.
An additional challenge for some of the women was the strain related to sick children or children in need of extra care. Children's illnesses and social problems are a challenge for parents, and in our study most of the women in this situation were single mothers. The children's problems included drug abuse, Attention-Deficit/Hyperactivity Disorder (ADHD), mental problems, sexual abuse, myalgic encephalomyelitis (ME), and eating disorders or selfharm. Children with drug-abuse or mental problems sometimes exhibit threatening behaviours that make their mothers anxious. The women described the heavy burden of being a single parent to children with health problems. Most of them had lived with these problems for years. What is more, the mothers felt that their responsibility for caring for their troubled children extended into their children's adulthood, when they are normally expected to live their own independent lives. For most of the women dealing with these problems was a lonely task. Only a few of the single mothers received support from a social network; the others had a tendency to isolate themselves.
The workplace was for many of them a place where they could forget their domestic worries, a place where they could be accepted and gain recognition. But the total burden forced them to take on either part-time jobs or to have periods of sickness absence.

The workplace as a recreational arena
Domestic problems may in some cases legitimize sick leave, but this does not give them a break from their most acute problems. Rather, the consequence of being sick-listed may be that they have to stay at home where the roots of their problems are situated, and this gives them no chance to have some relief from the problems.
If the roots of the problems leading to ill-health are mainly work-related, the sickness absence period is a chance to have some distance from the place and persons where the problems are situated. For some of the men in our sample this gave them an opportunity to participate more in the daily life in their families. They received the time and opportunity to take their children to kindergarten and to participate in leisure activities, in addition to helping with the daily work in the household. Their children and wives very much appreciated these opportunities, and the men regarded them as a positive aspect and an added value of their sickness-absence period.
For the women in our study, this was not the case. Since most of the women already had most of the responsibility for the everyday life of their families, the out-of-work situation was not a respite from their daily lives. On the contrary, some of the women expressed that they would have preferred to have been sick-listed from their burdens in the domestic sphere. For some of the women in this situation, going to their workplace was actually a time-out from their problems at home. For those who had a job they were coping with their tasks and had good relations with their colleagues and managers. In short, the workplace was an important arena for respect and recognition. One example was Anita, who dealt continually with domestic strain because her two children were drug abusers. She isolated herself and was depressed, but she strove to maintain her dignity. In this situation the nursing home where she worked was the best place to go: 'At my job I am known, respected, accepted and yesthere I feel I am at the top.' Anita is an example of those women who would actually fight to stay at work and to avoid sick leave. Staying at home made her depressed, and she looked forward to the day when she could work again. She had even begun planning to look for alternative forms of work if she could not return to the nursing home because of her musculoskeletal problems. The study showed that many women visited their workplace frequently during their sickness-absence period, which indicates that the women had a high degree of job satisfaction and positive interpersonal relations at their workplace. In our study care workers were well represented. In spite of the substantial pressures of the job they enjoyed being together with their patients and colleagues, and they missed this contact when they were sick-listed.

Discussion
This study has shown that to understand sickness absence from the perspectives of sick-listed women, it is necessary to take into consideration their psychosocial surroundings and the complexity of their life situation. From this perspective illness and sickness absence can be understood as manifestations of the strain in people's lives. The study has pointed out different aspects of women's work situation and domestic life that may cause illness and sickness absence. Certainly, this study is limited; the sample is small and generalizations must be made with caution. We do not claim to have unveiled all aspects that may have an influence, but the data gives some insight into complex developmental processes behind their sickness absence. In particular, the study is a reminder about the different aspects of women's domestic situation that may have an influence upon sickness leave.
The findings of this study are discrepant to those of some earlier quantitative studies (Mastekaasa, 2000, Mastekaasa & Olsen, 1998) that have not found support for an increase in sickness absence among women with high domestic burden. One explanation may be the way that domestic burden is defined, where the number of children is often used as an indicator of domestic burden. The present study finds that number of children as a measure of domestic strain is insufficient. A number of other factors also appear to influence domestic strain, including the presence of children who are sick or have problems that demand a great deal of attention and the adverse effects of having a troubled relationship to their partner. Other studies have also confirmed this observation. Östlund et al. (2004) have identified three dimensions of domestic strain that influence sickness absence: the division of domestic work, the division of responsibility for domestic life and the amount of social-emotional support at home. Staland-Nyman et al. (2007) have found correlations between domestic work strain, domestic work equity, marital satisfaction and self-related health.
The present study shows that a sufficient understanding of the causes of women's sickness absence requires a holistic approach. In a few cases the causes of their sickness absence are related solely to either domestic or to work strains, but most often sick leave is an outcome of the interplay between both sources of strain. It is, therefore, insufficient to study work-related or domestic sickness absence in isolation. Women's ill-health and sickness absence must be understood as an interaction between domestic and workrelated strains. The women's interpretations of their ill health and sickness absence also indicate that the inclusion of their life histories is needed. In addition to life strains in adulthood, life strains experienced during childhood and adolescence also contribute to greater vulnerability to ill health. Bjørngaard et al. (2009) have concluded that women probably to a larger degree than men experience a role conflict between being a caregiver of children and a worker. The present study partly confirms this conclusion, but it is necessary to add that work burden and the sense of having total responsibility, as well as the emotional aspects, are important. Single mothers and mothers who have partners who are away from home for longer periods are especially vulnerable.
Our study shows that single mothers and mothers with sick children or with family members in need of care often experienced a heavy burden of responsibility and feelings of underachievement. These factors make them vulnerable groups for ill-health and sick leave. The quality of life of the single mothers can be reduced by both economic strain and the total load of work.
The studies of Östlund et al. (2004) and Staland-Nyman et al. (2007) emphasize the importance of using measures that also include emotional content of domestic work. Our study illustrates that a sufficient understanding of the factors contributing to sick leave requires the inclusion of emotional dimensions in both the workplace and the domestic sphere.
The women in our sample had a strong identity both as mothers and workers. The single mothers experienced the conflicts between these identities especially intensely. According to Harding (1986) these identities are formed through interactions between symbolic and structural factors. At the symbolic level the expectation that women are to take care of the children is greater than what is expected of men. At the structural level there are strong expectations for women in our societyincluding single mothersto participate in the workforce. This pressure comes to expression through the work fare policy. These factors have consequences on women's freedom of action and their adaptations in their daily lives.
In the last decade, the reduction of the sickness absence has been a main political aim in Norway. The efforts to reduce the level of sickness absence have been mainly directed towards the workplace and work-related sick leave. Some of the efforts in certain businesses have been reported to have been successful, but the efforts have not resulted in a significant reduction of the national level of sick leave (Ose et al., 2009). To find successful actions for reducing sickness absence, it is important to have as precise knowledge as possible of the factors contributing to ill health. Our study shows that domestic strain and the interplay between work strain and domestic strain may be important factors that contribute to women's sickness absence. This study also advises that efforts to reduce sickness absence would do well to take this into consideration. One important question that arises from this consideration is how it might be possible to help relieve women with a high burden of domestic strain. Therefore, to reduce sickness absence particularly among women, it may be important for employers, doctors, and social workers to have strategies which meet the absentee's need for help not only at the workplace but also in the domestic sphere. Today some households turn to private firms in order to reduce the burden of domestic work (e.g., house cleaning). However, this solution is beyond the resources of low-income families, who often have problems making ends meet. For some of these families, support from the public services in the form of domestic help could be a solution for keeping women in the work force. However, women may be in need of different kinds of help and support, and it is necessary to take the individual's specific needs into consideration.

Conclusion
The present study illustrates the importance of both gender-specific approaches and qualitative methods in grasping the scope and complexity of the conditions that contribute to women's sickness absence. In particular, this study suggests the importance of a holistic approach. The efforts to reduce women's sickness absence must take into consideration their domestic strain, and especially the situation of single mothers and women who bear the responsibility of caring for grown-up children who are sick or unable to take care of themselves as well as elderly people in the family in need of care. Since the factors that contribute most to ill health for some women are situated in the domestic sphere, it is necessary to consider what efforts can be done to mitigate their domestic burdens.
Both quantitative and qualitative research is needed to explore further the different aspects of domestic strain, the interaction between domestic and work-related strain, and the long-term consequences of the double burden that women carry. Only on the basis of this knowledge will it be possible to formulate the appropriate efforts that will effectively contribute to a reduction in women's sickness absence.