OUP user menu

Socio-economic position, family demands and reported health in working men and women

Enrique Regidor, Cruz Pascual, Luis de la Fuente, Juana M. Santos, Paloma Astasio, Paloma Ortega
DOI: http://dx.doi.org/10.1093/eurpub/ckq024 109-115 First published online: 17 March 2010

Abstract

Background: This study evaluates the extent to which domestic workload explains socio-economic differences in poor self-reported health in women and men. Methods: In total, 6284 men and women who were employed and living with a partner were selected from the 2003 Spanish Health Interview Survey. The indicators of family demands investigated were person responsible for housework, number of persons in the household and the presence of at least one child under 15 years of age in the household. The measures of socio-economic position were educational level and household income, and the measures of health status were poor perceived health and limitation of activity due to disease. Results: Household size and presence of a child under 15 in the home were not related with the measures of health status. The indicator about the person who does the housework was related with poor perceived health and with activity limitation. Specifically, the worst health status was seen in respondents who lived in homes where the partner or other family members did the housework. In general, the relation between indicators of socio-economic position and measures of health status was not modified after taking into account the person who does the housework. Conclusion: Among working people with a partner, persons who work and do their own housework do not have poorer perceived health than those living in homes where other people do the housework. This indicator of family demands does not explain the socio-economic differences in self-reported health.

  • family characteristics
  • household work
  • inequalities in health
  • multiple roles
  • self-reported health

Introduction

Socio-economic position refers to the social and economic factors that influence individuals’ position in the economic structure.1 The distribution of exposures and of resources that affect health according to the socio-economic position may differ in men and women.2,3 Socio-economic positions of men and women originate from the social relations of production linked to those operating with family and some conditions and processes in the labour marked and in the family are different in men and women. Socio-economic positions of men and women originate from the social relations of production linked to relations operating with family, and some conditions and processes of those relations in the labour marked and in the family are different in men and women. Thus, the mechanisms explaining the relation between socio-economic position and health should be studied separately in men and women.

Some authors have stated that women’s role in caring for their families must be taken into account since the incorporation of women into the workplace has not exempted them from the responsibility for housework and caring for family members.4,5 It has been suggested that the higher risk of mortality and the greater frequency of health problems in women of low socio-economic position compared with women of high socio-economic position may be due to their having to take on this double workload.6,7

Several previous studies do not support this explanation of the relation between socio-economic position and mortality, since women who play all three roles—wife, mother and worker—do not present a greater risk of mortality than women who play one or two of those roles.8–10 However, the lack of an effect on mortality does not mean there are no effects on morbidity.11–13 For example, different studies in Spain have shown that some indicators of domestic labour are related with the subjective perception of poor health and with the presence of reported chronic diseases.7,14–16

Although these findings suggest that domestic workload may help explain social inequalities in health in women, this hypothesis could not be tested in these studies. One of the studies was made in a highly selected subsample of women, and no socio-economic inequalities in health were found.14 In the other study, which used Eric O. Wright’s social classification, the association between social class and reported heath status was less evident in women than in men.16 Wright’s social classification takes into account ownership of productive assets, and control and authority relations in the workplace (control over organizational assets). The authors of the latter work noted that the relation between socio-economic position and health may not have been adequately evaluated, since women’s health is more influenced by factors related to their household roles, while Wright’s social classification reflects ownership and control over productive assets. In their opinion, the results would probably have been different if other indicators had been used, such as the educational level, which reflects health inequalities in women more adequately,5 or household social position, which is a more valid reflection of women’s socio-economic position.17 According to Arber and Klat ‘occupational class may be a less discriminating indicator of health inequalities for women than men because of women’s more fragmented employment career, while educational qualifications may capture comparable or greater inequalities for women than men’.5 On the other hand, Krieger et al. have suggested the need of incorporating measures of socio-economic position of the household in which individual reside because the social class of an individual not necessarily tells about her or his household social position.17 This is important in women given that class differences in health among women may be underestimated because categories of occupational class are based on male distributions of occupation.

To surmount this limitation of the aforementioned investigation, the objective of the present study was to evaluate the relation of educational level and level of household income with different measures of self-reported health in working men and women residing in Spain, and to assess the degree to which domestic workload explains this relation.

Methods

The data were taken from the 2003 Spanish Health Survey.18 Respondents were selected by stratified multistage sampling of the non-institutionalized population. The first stage units—census sections—were grouped into strata by the size of the municipal population and were then selected with a probability proportional to the size of the population of the stratum. The second stage units—households—were selected within each census section with the same probability by systematic sampling with random start. Within each household, one person aged 16 years or older was selected to complete the questionnaire. A weight was assigned to each respondent to reflect the sample design and all analyses were carried out taking into account this weight. For the present study we selected persons aged 25–64 years who were employed at the time of the interview and who lived with a partner (with or without other persons in the household).

Two indicators of health were studied: poor perceived health and limitation of activity. Self-perceived health was measured with the following question: ‘Would you say your overall health is very good, good, fair, poor or very poor?’ A respondent was considered to have poor health when the response was fair, poor or very poor. A respondent was considered to have limitation of activity if the response to the following question was affirmative: ‘In the last 12 months, have you had any ailment, disease or health problem that limited your activity for more than 10 straight days?’

Three indicators of family demands were investigated: person responsible for doing the housework, number of persons in the household and the presence of a child under 15 years of age in the household. The first indicator was measured with the question ‘Who is mainly responsible for household tasks like cleaning, cooking, ironing, etc.?’ and the responses were grouped into four categories—respondent, partner, other family members, hired person.

The variables reflecting the socio-economic position were the educational level of the respondent and the household economic income. The non-response rate for the question on household economic income was 26%. To reduce this percentage, an income value was imputed for non-respondents using the hot-deck imputation procedure proposed by Cox and Cohen.19 Following this procedure, respondents who answered the question on income were classified according to a combination of the following variables: age (in 10-year intervals), sex, educational level and social class based on occupation. The most frequent value for income in each of these categories was obtained, and was applied to respondents with missing income data who were in the same category after grouping them according to the same variables. A previous study supports the validity of the hot-deck imputation.20

We calculated the percentage of respondents who had health problems according to family demands. We first evaluated the relation between each indicator of family demands and the measures of health by calculating the age- adjusted percentage ratio using the binomial regression.21 In each one of those relations, we evaluated the possible interaction between each measure of the socio-economic position and the indicators of family demands, since some previous studies have shown a relation between indicators of family demands and self-reported poor health only in women of low socio-economic position.7,15 In each regression model we introduced an interaction term of the indicators of family demands with education or income.

Then we evaluated the relation of family demands with socio-economic position by chi-square tests. Finally, we estimated the magnitude of the association between the measures of socio-economic position and the measures of health by calculating the age-adjusted percentage ratio. We evaluated whether any of the indicators of family demands explained part of this association by introducing family demands as an adjustment variable.

The relation between family demands and health may differ depending on a woman’s age. For example, the frequency of children under 15 in the household decreases with increasing age of the women. On the other hand, women’s incorporation into the workplace has occurred progressively, as reflected in the fact that at the beginning of this century the employment rate was 70% in women aged 25–44 years, but 40% in women aged 45–64 years.22 Consequently, separate analyses were made in these two age groups.

Results

Table 1 shows the relation between family demands and health in women. Significant age-adjusted percentage ratios were observed with the indicator about the person who does the housework, but not with the household size or with the presence of children under 15. In women aged 25–44 years, the highest age-adjusted percentage ratio of poor perception of health was observed in those who resided in homes where the partner does the housework [PR: 2.07, 95% confidence interval (CI): 1.15–3.70] and the highest age-adjusted percentage ratio for limitation of activity was observed in those who lived in homes where the housework was done by other family members (PR: 2.08, 95% CI: 1.19–3.62). In women aged 45–64 years, the highest age-adjusted percentage ratio for poor perceived health was seen in those who lived in homes where the housework is done by other family members.

View this table:
Table 1

Relation of indicators of family demands with poor perceived health, and limitation of activity: women

Poor perceived healthLimitation of activity
Indicators of family demandsSample sizePercentageAge-adjusted percentage ratio (95% CI)PercentageAge-adjusted percentage ratio (95% CI)
25–44 years old
Person responsible for housework
    Hired person14211.31.0018.01.00
    Other family members12520.01.86 (1.15–3.00)37.42.08 (1.19–3.62)
    Partner3420.02.07 (1.15–3.70)18.00.98 (0.58–1.66)
    Respondent110819.31.88 (1.16–3.03)17.70.98 (0.67–1.43)
Household size
    Two30416.81.0018.51.00
    Three39514.90.82 (0.57–1.16)17.00.94 (0.67–1.31)
    Four54519.40.99 (0.70–1.39)17.40.98 (0.70–1.37)
    More than four16527.81.38 (0.93–2.04)23.71.35 (0.90–2.02)
Children under 15 years
    No44220.71.0018.21.00
    Yes96717.60.95 (0.91–1.00)18.30.97 (0.77–1.22)
45–64 years old
Person responsible for housework
    Hired person7128.91.0033.31.00
    Other family members2749.71.69 (1.01–2.85)22.60.67 (0.31–1.45)
    Partner3932.41.20 (0.67–2.15)34.81.04 (0.61–1.78)
    Respondent69631.61.14 (0.78–1.66)24.30.72 (0.51–1.03)
Household size
    Two12835.31.0024.81.00
    Three22236.61.12 (0.83–1.50)26.51.08 (0.74–1.58)
    Four27425.80.82 (0.59–1.12)24.81.02 (0.70–1.50)
    More than four20933.31.04 (0.76–1.42)25.91.06 (0.72–1.56)
Children under 15 years
    No64832.51.0024.51.00
    Yes18530.11.03 (0.79–1.34)28.91.21 (0.92–1.60)

The relation between family demands and health in men are shown in table 2. Men who lived in homes where the housework was done by other family members had the highest age-adjusted percentage ratios (PR: 2.53, 95% CI: 1.17–5.49 in men aged 25–44 years and PR: 2.42, 95% CI: 1.06–5.55 in those aged 45–64 years). This indicator of family demands was also related with the limitation of activity in men aged 45–64 years: the highest age-adjusted percentage ratio was observed in men who lived in homes where they do the housework. The size of household also showed a significant relation with poor perceived health in men aged 25–44 years: the highest age-adjusted percentage ratio was seen in men who lived in household units of more than four persons (PR: 1.65, 95% CI: 1.07–2.55).

View this table:
Table 2

Relation of indicators of family demands with poor perceived health, and limitation of activity: men

Poor perceived healthLimitation of activity
Indicators of family demandsSample sizePercentageAge-adjusted percentage ratio (95% CI)PercentageAge-adjusted percentage ratio (95% CI)
25–44 years old
Person responsible for housework
    Hired person10010.81.0019.11.00
    Other family members3527.72.53 (1.17–5.49)30.71.61 (0.85–3.04)
    Partner197614.31.37 (0.77–2.44)14.10.75 (0.49–1.14)
    Respondent1254.80.49 (0.18–1.28)15.80.86 (0.48–1.52)
Household size
    Two3219.61.0012.71.00
    Three63712.91.28 (0.86–1.91)14.81.14 (0.83–1.61)
    Four97214.71.40 (0.95–2.05)15.61.17 (0.83–1.64)
    More than four30617.61.65 (1.07–2.55)13.30.99 (0.65–1.51)
Children under 15 years
    No44213.11.0014.71.00
    Yes179414.00.97 (0.74–1.27)14.60.96 (0.75–1.24)
45–64 years old
Person responsible for housework
    Hired person10314.91.0012.91.00
    Other family members1731.52.42 (1.06–5.55)28.92.27 (0.97–5.86)
    Partner164324.01.61 (1.01–2.56)18.61.44 (0.86–2.39)
    Respondent4322.71.67 (0.82–3.41)29.42.41 (1.22–4.78)
Household size
    Two20726.41.0019.21.00
    Three41722.60.93 (0.70–1.24)22.81.22 (0.87–1.70)
    Four70322.11.00 (0.77–1.31)15.50.85 (0.61–1.19)
    More than four47925.21.11 (0.85–1.47)19.31.05 (0.75–1.47)
Children under 15 years
    No122225.61.0018.21.00
    Yes58419.40.91 (0.74–1.11)19.61.18 (0.95–1.47)

The only interaction observed was between the indicator about the person who does the housework and the educational level in women aged 25–44 years. In women with university education, the highest age-adjusted percentage ratio was seen in those who lived in homes where the housework was done by other family members while in women with less than university studies, the highest age-adjusted percentage ratio was seen in those who lived in homes where the partner does the housework.

Table 3 shows the relation between indicators of socio-economic position and indicators of family demands. In general, the relation was significant, except for the presence of a child under 15 years in the household in women aged 25–44 years.

View this table:
Table 3

Sample size, percentage of respondents who do housework, who live in households of more than four persons, and who live in households with child under 15 years of age, by measures of the socio-economic position

Measures of socio-economic positionWomenMen
Sample sizeHousework done by respondentMore than four persons living in householdChild under age 15 years living in householdSample sizeHousework done by respondentMore than four persons living in householdChild under age 15 years living in household
25–44 years old
    Education
    Tertiary43463.99.065.74088.012.675.3
    Upper secondary48080.98.370.47527.610.077.4
    Lower secondary25389.516.768.85083.511.281.3
    Primary/no education24289.318.170.25683.121.686.4
    χ2 -test, P-value<0.001<0.0010.250<0.001<0.001<0.001
Monthly income
    More than €180055964.612.268.05969.711.874.5
    From €1201 to €180050687.58.770.07545.813.778.2
    From €901 to €20023087.216.066.55272.310.885.0
    Up to €90011490.514.170.23593.121.286.8
    χ2 -test, P-value<0.001<0.010.750<0.001<0.001<0.05
45–64 years old
    Education
    Tertiary19271.025.530.13294.326.339.9
    Upper secondary15077.423.727.03692.920.838.2
    Lower secondary10390.417.727.82660.725.336.8
    Primary/no education38890.327.414.98422.029.425.5
    χ2 -test, P-value<0.0010.240<0.01<0.005<0.005<0.001
Monthly income
    More than €180032377.930.624.65893.331.136.7
    From €1201 to €180025282.322.623.96082.423.933.7
    From €901 to €120010191.320.518.33451.123.024.9
    Up to €90015792.020.717.02642.126.928.9
    χ2-test, P-value<0.001<0.050.1000.100<0.005<0.001

The relation of educational level and income with the indicators of health is shown in table 4. Only the indicator about the person who does the housework was introduced in the models, because the other two indicators of family demands did not show any relation with the indicator of health. Poor perceived health in women showed a significant relation with both indicators of socio-economic position. The highest age-adjusted percentage ratio was observed in women with the lowest educational level [2.95 (95% CI: 2.10–4.15) and 2.59 (95% CI: 1.83–3.68) in women aged 25–44 and 45–64 years, respectively]. In limitation of activity, high and significant age-adjusted percentage ratios were found in women aged 25–44 years with upper secondary studies and in women aged 45–64 years who belonged to the second highest income category. In men, significant relation of both measures of socio-economic position with the poor perceived health and limitation of activity was only found in those aged 45–74 years. The highest age-adjusted percentage ratios were found in men with the lowest educational level and in men who belonged to the lowest income category.

View this table:
Table 4

Relation of education and income with poor perceived health and limitation of activity in women and in men

Indicators of socio-economic positionPoor perceived healthLimitation of activity
PercentagePercentage ratio (95% CI)PercentagePercentage ratio (95% CI)
Model 1Model 2Model 1Model 2
25–44 years old
Women
Education
    Tertiary10.01.001.0015.51.001.00
    Upper secondary19.41.97 (1.41–2.76)1.93 (1.37–2.72)20.91.35 (1.02–1.79)1.34 (1.01–1.79)
    Lower secondary20.72.05 (1.41–2.97)2.01 (1.37–2.95)18.01.16 (0.83–1.64)1.15 (0.81–1.64)
    Primary/no education30.22.95 (2.10–4.15)2.90 (2.03–4.14)18.31.19 (0.84–1.68)1.17 (0.82–1.65)
Income
    More than €180015.21.001.0017.91.001.00
    From €1201 to €180020.01.34 (0.86–2.04)1.24 (0.81–1.89)19.31.08 (0.84–1.39)1.09 (0.84–1.42)
    From €901 to €120022.81.48 (1.09–2.02)1.37 (1.00–1.88)17.30.97 (0.69–1.35)0.93 (0.66–1.32)
    Up to €90019.81.37 (1.05–1.78)1.27 (0.97–1.66)17.30.97 (0.62–1.69)1.00 (0.64–1.56)
Men
Education
    Tertiary12.01.001.0014.11.001.00
    Upper secondary13.71.17 (0.85–1.61)1.10 (0.80–1.53)14.61.05 (0.78–1.41)0.90 (0.68–1.18)
    Lower secondary15.51.31 (0.94–1.83)1.19 (0.84–1.68)14.91.07 (0.78–1.47)1.04 (0.78–1.38)
    Primary/no education13.71.12 (0.80–1.56)1.02 (0.72–1.44)14.91.06 (0.78–1.44)1.05 (0.79–1.53)
Income
    More than €180010.81.001.0015.31.01.00
    From €1201 to €180013.51.24 (0.93–1.67)1.19 (0.88–1.61)13.30.88 (0.68–1.15)1.09 (0.80–1.49)
    From €901 to €120014.41.32 (0.97–1.81)1.26 (0.91–1.74)15.10.97 (0.74–1.28)1.12 (0.80–1.17)
    Up to €90018.61.74 (1.27–2.39)1.63 (1.18–2.26)15.20.99 (0.73–1.34)1.10 (0.79–1.53)
45–64 years old
Women
Education
    Tertiary15.81.001.0013.01.001.00
    Upper secondary29.81.85 (1.23–2.79)1.90 (1.26–2.88)18.21.12 (0.78–1.60)1.19 (0.82–1.71)
    Lower secondary26.31.68 (1.06–2.66)1.79 (1.12–2.85)13.00.77 (0.48–1.23)0.85 (0.53–1.34)
    Primary/no education42.42.59 (1.83–3.68)2.77 (1.92–3.98)20.01.07 (0.80–1.45)1.19 (0.86–1.65)
Income
    More than €180022.01.001.0010.81.001.00
    From €1201 to €180039.41.73 (1.34–2.24)1.72 (1.32–2.23)21.81.59 (1.20–2.11)1.68 (1.26–2.23)
    From €901 to €120034.21.50 (1.07–2.12)1.52 (1.08–2.15)18.51.27 (0.85–1.90)1.37 (0.91–2.05)
    Up to €90039.21.67 (1.24–2.24)1.72 (1.32–2.23)22.11.28 (0.90–1.81)1.39 (0.97–1.99)
Men
Education
    Tertiary11.51.001.0011.31.001.00
    Upper secondary17.41.62 (1.11–2.35)1.61 (1.10–2.35)20.61.87 (1.30–2.70)1.89 (1.30–2.74)
    Lower secondary22.21.96 (1.37–2.79)1.94 (1.33–2.84)16.51.48 (0.98–2.23)1.52 (1.00–2.31)
    Primary/no education31.42.67 (1.95–3.67)2.66 (1.92–3.68)21.41.88 (1.35–2.62)1.91 (1.35–2.71)
Income
    More than €180018.41.001.0017.11.001.00
    From €1201 to €180020.11.14 (0.90–1.44)1.11 (0.87–1.40)17.81.06 (0.83–1.36)1.04 (0.81–1.34)
    From €901 to €120029.21.51 (1.19–1.91)1.45 (1.14–1.85)18.51.06 (0.80–1.42)1.03 (0.77–1.39)
    Up to €90035.91.86 (1.48–2.35)1.80 (1.41–2.29)24.51.41 (1.07–1.86)1.37 (1.03–1.81)
  • Model 1: age-adjusted percentage ratio. Model 2: percentage ratio adjusted for age and for person responsible for housework.

Adjusting for the person who does the housework did not modify the magnitude of the percentage ratios in both sexes, except for the relation between income and poor perceived health in women aged 25–44 years, which declined by about 30% after adjusting for the person who does the housework. Adjusting for the size of household in men aged 25–44 years did not modify the magnitude of the association either.

Discussion

Two indicators of family demands—number of persons in the household and the presence of a child under 15—did not show a relation with the health measures analysed. The indicator about the person who does the housework was related with poor perceived health and, in some groups, with the limitation of activity. The worst health was observed in respondents who lived in homes where the housework was done by the partner or by other family members. The relation between the indicators of socio-economic position and the measures of health status was not modified after taking into account the indicator about the person who does the housework.

Several studies in Spain have found that the prevalence of poor perceived health in working women increases with the number of persons in the household.7–15 This relation has also been seen in men, although the magnitude of the estimates was not significant.14–15 Based on these findings, the authors of these studies state that household size is a good indicator of the total domestic workload for married women. They support their argument by referring to a study showing that reported stress was more frequent in women in larger households,23 even though the authors did not find in their own studies that household size was related with either psychosomatic symptoms14 or mental health.7,15

In our study no relationship was found between household size and poor perceived health in women. On the other hand, whereas in the above-mentioned studies the frequency of poor perceived health increased with the number of persons in the household, studies in the United States and Spain have found that the risk of mortality in men and women aged 25–64 years decreases with increasing number of persons in the household.24,25 The findings of the present study and of the investigations on mortality suggest there is considerable uncertainty about the significance of the relationship between household size and health.

The presence of a child under 15 in the home has also been used as an indicator of domestic workload due to its possible association with reduced psychological well–being.26 However, studies in England and Japan have observed opposite results: among women who live with a partner, those with young children have better indicators of health than those who do not have children living at home.27,28 In our study we found no relationship between the presence of young children in the home and the indicators of health status.

In relation with the findings of the previous studies carried out in Spain, the results of our study were unexpected, since the worst health was not seen in women who do their own housework. In the first place, the prevalence of poor perceived health in women aged 25–44 years was very similar in women who live in homes where the housework is done by other family members, the partner or the women themselves—although women whose partner does the housework showed a slightly higher prevalence ratio. In the second place, the highest prevalence of poor perceived health in women aged 45–64 years was seen in those who live in homes where the housework is done by other family members. Although it is not possible to rule out the possibility that women who do housework are more healthy than the rest of the women, the prevalence of poor perceived health was similar in women who live in homes were the housework is done by the partner and in women who live in homes were the housework is done by themselves.

In men the highest prevalence of poor perceived health and of limitation of activity was observed in those who live in homes where the housework is done by other family members. These findings in men and the results in women aged 45–64 years suggest that these persons have a high frequency of health problems and may receive help from family members for this reason. However, the same analyses were made with persons who did not have any of the 16 chronic diseases studied, with similar results. Therefore, two possible explanations can be suggested: the sick bias would not be responsible for the findings because the same results were found in healthy people, or else the sick bias would be responsible for the finding due to other different chronic diseases than those included in the questionnaire.

Finally, the indicators of family demands that were related with some measures of health status did not explain the association between the socio-economic position and self-reported health. It was only found that the magnitude of the relation between income and poor perceived health decreased by around 30% in women aged 25–44 years after taking into account the person responsible for doing the housework.

Data limitations

This study is based on a sample of working people. Since working people are likely healthier than the general population it can not be ruled out that a healthy worker effect may be responsible for the lack of association between indicators of family demands and indicators of health in many cases. Likewise, working people may be quite different in many socio-economic aspects than non-working people, so the relation of the indicators of family demands and health could be different in people living with a partner where only one or neither of them works.

This was a cross-sectional study; therefore, it may be that the presence of health problems is what determines who does the housework. The results were identical, however, in the analyses made of persons with no chronic problems. Nonetheless, this finding does not rule out the possibility that problems of a psychosocial nature could explain the higher frequency of poor health in respondents who live in homes where the housework is done by the partner. In fact, majority of the chronic problems studied refer to physical problems and two of them refer to the presence of depression and the presence of other mental illness in general. However, it is possible that those women show psychosocial problems associated with occupational and work-related stressors.

In summary, neither household size nor the presence of children under 15 years in the home was associated with reported health. Furthermore, working persons who do their own housework did not have poorer perceived health than those living in homes where the housework is done by other persons. The indicators of socio-economic position showed a relation with poor perceived health, but this relation was not explained after the statistical adjustment by the indicator about the person who does the housework.

Funding

Fondo de Investigaciones Sanitarias (grant no PI060115).

Conflicts of interest: None declared.

Key points

  • Some indicators of domestic work frequently used in previous investigations did not show a relation with the measures of reported health studied.

  • The indicator that reflects the person who does the housework was related with poor perceived health. Neither women nor men who did their own housework had the poorest perceived health.

  • The highest frequency of poor perceived health was seen in women and men who lived in homes where the household tasks were done by the partner or other family members.

  • The relation between the socio-economic position and poor perceived health was not explained by the indicator that reflects the person who does the household tasks.

References

View Abstract