A prospective cohort study of deficient maternal nurturing attitudes predicting adulthood work stress independent of adulthood hostility and depressive symptoms

Abstract

Stressful childhood environments arising from deficient nurturing attitudes are hypothesized to contribute to later stress vulnerability. We examined whether deficient nurturing attitudes predict adulthood work stress. Participants were 443 women and 380 men from the prospective Cardiovascular Risk in Young Finns Study. Work stress was assessed as job strain and effort–reward imbalance in 2001 when the participants were from 24 to 39 years old. Deficient maternal nurturance (intolerance and low emotional warmth) was assessed based on mothers' reports when the participants were at the age of 3–18 years and again at the age of 6–21 years. Linear regressions showed that deficient emotional warmth in childhood predicted lower adulthood job control and higher job strain. These associations were not explained by age, gender, socioeconomic circumstances, maternal mental problems or participant hostility, and depressive symptoms. Deficient nurturing attitudes in childhood might affect sensitivity to work stress and selection into stressful work conditions in adulthood. More attention should be paid to pre-employment factors in work stress research.

Keywords::

• Childhood stress
• effort–reward imbalance
• job strain
• nurturing attitudes
• parenting
• stress sensitivity
• work stress

Introduction

Research on animals suggests that less nurturing maternal behavior in rats is associated with more pronounced hypothalamic–pituitary–adrenal axis responses to stress and higher behavioral stress responses in the offspring (Liu et al. 1997; Caldji et al. 1998). Existing research in humans also suggests that early stressful experiences may have long lasting effects on physiological stress–response systems (Luecken 1998; Bremner and Vermetten 2001; Heim et al. 2002; Evans and Kim 2007), and a recent review concluded that family environment characterized by frequent conflicts and cold, unsupportive, and neglectful relationships may predispose the individual to later stress and mental and physical health problems (Repetti et al. 2002).

In adulthood, a considerable proportion of time is spent in work settings. Work is a major source of stress and work stress has been repeatedly linked with a range of health problems, including cardiovascular disease (CVD; Kivimäki et al. 2006; Eller et al. 2009) and mental disorders (Stansfeld and Candy 2006). In spite of this, only few studies have examined pre-employment determinants of work stress (Hemmingsson and Lundberg 2006; Hintsa et al. 2006, 2007, 2008; Elovainio et al. 2007; Kivimäki et al. 2007) and to our knowledge, no previous studies are available on the association between childhood nurturance and adulthood work stress. Examining early roots of stress vulnerability is important for understanding the reasons for and consequences of work stress and for planning interventions that accurately take into account early predisposing factors.

In the current study, we examined whether deficient nurturing attitudes by the mother (intolerance towards the child and low emotional significance of the child to the mother) in the childhood predict work stress of the participants 18–21 years later in adulthood. We assessed work stress based on the two most widely used conceptual models proposing that work stress arises from a combination of high job demands and low job control (Karasek 1979) and effort–reward imbalance—a combination of low rewards received at work as compared to the level of efforts invested (Siegrist and Peter 1994). Furthermore, we examine whether the effects of maternal attitudes are explained by socioeconomic factors, mother's mental problems, or by the participants' hostility or depressive symptoms in adulthood.

Methods

Participants

The participants were derived from the ongoing prospective “Cardiovascular Risk in Young Finns” study, which began in 1980. The subjects for the original sample in 1980 (n = 3596) were selected randomly from six age cohorts (aged 3, 6, 9, 12, 15, and 18 years) in the population register of the Social Insurance Institution, a database that covers the whole population of Finland. The design of the study and the selection of the sample have been described in detail by Raitakari et al. (2008). The Young Finns cohort was enrolled into the study in 1980. The assessments of the present study were carried out in 1980, 1983 (maternal nurturing attitudes and mother's years of education) and in 2001 when the participants were aged 24, 27, 30, 33, 36, and 39 years (work stress variables and participants own education and occupational status). There were 2105 participants (58.5% of the 1980 cohort) who took part in the work stress assessments. Of these, 691 were not employed in a full-time job in 2001 or did not report employment status, and were therefore excluded. As working and possible work stress experienced by the participants at the time of maternal report of child-rearing attitudes were considered a potential source of confounding, participants who were known to work already in 1980 or in 1983 (n = 94, after other exclusions) were excluded from the analyses. There were 823 participants (380, 46.2%, men and 443, 53.8%, women) for whom full data had been obtained on all study variables in 1980, 1983, and 2001. All the analyses were conducted on these participants. The mean age of the included participants was 30.6 years (in 2001). Participants gave written informed consent, and the study was approved by Local Ethics Committees.

Measurement of job strain and effort–reward imbalance

Job demand and effort were assessed with the same three-item scale based on the occupational stress questionnaire (OSQ; Cronbach's alpha, α = 0.6; Elo et al. 1992) developed at the Finnish Institute of Occupational Health. The OSQ has been widely used in Finland and the validity of the OSQ items has been satisfactory in studies involving a total of over 25,000 employees in various occupations (Elo et al. 1992; Elo 1994). The items used in the current study were “Do you have to hurry to get your work done?” “Does your work have phases that are too difficult?” and “Is your work mentally strenuous?” These items correspond quite closely to demands in Karasek's (1985) job content questionnaire. They also resemble efforts in Siegrist's et al. (2004) effort–reward imbalance questionnaire. Responses were obtained on a 5-point scale ranging from 1 (never) to 5 (all the time).

Job control was measured with the job content questionnaire (Karasek 1985), which includes nine items for job control (α = 0.9). Responses were given on a 5-point scale ranging from 1 (strongly disagree) to 5 (strongly agree). We constructed job strain as a linear term, a continuous job strain variable obtained from the following equation: (job demand score) − (job control score; Landsbergis et al. 1994).

Reward was also measured with a three-item scale (α = 0.6) from the OSQ. The items used in the current study were “Do you get help and support from your superior if needed?” [1 (very little) to 5 (very much)] “How do your co-workers get along with each other in the workplace?” [Their relationship range was 1 (bad, tense, resentful, etc.) to 5 (very good)] “How satisfied are you with your current employment?” [1 (very unsatisfied) to 5 (very satisfied)]. These items are similar to the reward items in the effort–reward imbalance questionnaire (Siegrist et al. 2004). Effort–reward imbalance was constructed as a continuous variable, dividing effort by reward, as recommended by Siegrist et al. (2004). A logarithmic transformation was performed to correct for skewness and kurtosis. All work stress scales were calculated with mean function. The participants were required to have answered to at least 50% of the items of each scale. Participants with more missing values were excluded.

Nurturing attitudes

Deficient nurturing attitudes (as indicated by intolerance of the mother towards the normal activity of the child, and low emotional warmth by the mother towards the child) were self-rated by the mothers in 1980 and in 1983 using a scale based on the operation family study (Makkonen et al. 1981). Intolerance (α = 0.7 and 0.7, for assessments in 1980 and 1983, respectively) was assessed with three items on a 5-point scale, e.g. “In difficult situations the child is a burden”, 1 = not at all, 5 = a great deal. Low emotional warmth (α = 0.7 and 0.8, for assessments in 1980 and 1983, respectively) was assessed with four items on a 5-point scale, e.g. “The child is significant to me”;, 1 = very significant, 5 = not significant. Mean scores of these scales were formed so that high values represent deficient nurturing attitudes, i.e. high intolerance and low emotional warmth. Here again, the participants were required to have information on at least 50% of items of each scale to be included in the study. Finally, the mean of intolerance in 1980 and 1983 and mean of low emotional warmth in 1980 and 1983 were calculated, and these scales were used in the analyses. Because these variables were slightly skewed they were transformed, intolerance with square root transformation and low emotional warmth with inverse transformation (after which low emotional warmth was reverse scaled to counterbalance the fact that the inverse transformation changed the direction of the variable). These procedures allowed for the correction of skewness resulting in the following values: 0.055 for intolerance and 0.10 for low emotional warmth.

Participant socioeconomic status, childhood family income, and maternal education

Socioeconomic status (SES) of the participants was indicated by educational level and occupational group. Educational level was classified as (1) low (comprehensive school), (2) intermediate (secondary education), or (3) high (academic; graduated from a polytechnic or studying at or graduated from a university). Classification into occupational groups was based on the criteria of the Central Statistical Office of Finland. Three groups were formed: (1) manual, (2) lower non-manual, and (3) upper non-manual. Entrepreneurs, who formed a heterogeneous group of their own in the original measure, were placed in the aforementioned occupational groups according to educational level (low, intermediate, and high education corresponding to manual, lower non-manual, and upper non-manual occupational groups, respectively).

Childhood SES was assessed as it has previously predicted later stress (Hintsa et al. 2006; Evans and Kim 2007). Childhood SES was indicated by childhood family income (assessed in 1983 with an 8-point scale of annual income) and maternal education (indicated by mother's years of education in 1983).

Mother's mental problems

Mother's mental problems were assessed as they were considered to be a potential confounding factor in the current study. Mother's mental problems were self-reported by the mother in 1983 by recording whether she had been found to have a mental problem or mental disorder. The variable was coded as 0 = no, 1 = yes.

Hostility and depressive symptoms

Participants' hostility was assessed in 2001 with a mean score obtained from subscales anger, cynicism, and paranoia. The anger subscale was the seven-item irritability scale of the Hostility Inventory (sample item: “I often feel like a powder keg ready to explode”; Buss and Durkee 1957); the cynicism subscale (seven items) was derived from the Minnesota multiphasic personality inventory (“I think most people would lie to get ahead”; Comrey 1957, 1958), and the paranoid tendencies subscale from six items of the paranoid ideation subscale of the Symptom distress checklist-revised (“I think that other people talk about me behind my back”; Derogatis and Cleary 1977). Each of the items was rated on a 5-point scale ranging from 1 (very strongly disagree) to 5 (very strongly agree). We have previously reported an α of 0.87 for the hostility scale in the Young Finns data (Keltikangas-Järvinen et al. 2008).

Depressive symptoms were assessed with a revised version of Beck's depression inventory (BDI; Beck and Steer 1987; Katainen et al. 1999). In the present study, the participants were asked to rate 21 items (e.g. “I often feel sad”) on a 5-point scale ranging from 1 (strongly disagree) to 5 (strongly agree). We have previously reported an α of 0.92 for the modified BDI in the Young Finns data (Hintsanen et al. 2009).

Statistical analyses

There were no significant gender or age interactions. A single three-way interaction between gender, age, and emotional warmth on job strain (p = 0.033) was considered a chance finding. Therefore, all analyses were conducted for all participants together. These interactions were analyzed with analysis of variance (general linear model (GLM) univariate).

Attrition analyses were calculated with two-sided t-tests and χ²-tests. Means and SDs or numbers and percentages are presented for those variables where included and excluded participants differ. Participant characteristics are described by calculating means and SDs or by calculating numbers per category and percentages. Bivariate associations were calculated with Pearson correlations. Hierarchical linear regression models were constructed examining deficient maternal nurturing attitudes predicting later adulthood work stress in the participants. The first step was unadjusted. The second step was adjusted for age and gender, and the third step for age, gender, mother's years of education, childhood family income, and SES of the participant in adulthood. The fourth step was additionally adjusted for mother's mental problems and the fifth step for all mentioned variables and participant's hostility and depressive symptoms in adulthood. The variables were entered using enter method.

Interactions between depressive symptoms and maternal attitudes and between hostility and maternal attitudes on work stress were analyzed with analysis of variance (GLM univariate). There were no significant interactions between hostility and maternal attitudes and there was only one interaction between depressive symptoms and maternal attitudes (participant depressive symptoms interacting with emotional warmth on rewards at work, p = 0.017). Taking the number of interaction analyses into account, this single interaction was considered a chance finding and no further analyses were calculated.

Interactions between maternal education and maternal nurturing attitudes were examined with analysis of variance. No significant interactions were found. All analyses were calculated with SPSS software (version 15.0; SPSS Inc., an IBM Company Headquarters, Chicago, Illinois) using a critical alpha value of 0.05.

Results

Attrition

The differences between the original sample (n = 3596) gathered in 1980 and the final study sample (n = 823) were examined with t-test and χ²-test. Included participants were somewhat younger (mean ages 30.6 ± 4.4 vs. 31.7 ± 5.1 (SD) years, p < 0.001) and more educated (2.3 ± 0.6 vs. 2.1 ± 0.6, p < 0.001). Furthermore, they had higher occupational status (2.0 ± 0.8 vs. 1.8 ± 0.7, p < 0.001), their mothers were slightly more educated (10.2 ± 3.1 vs. 10.0 ± 3.3 years, p = 0.040), and their family had higher income in their childhood (6.0 ± 1.9 vs. 5.7 ± 2.0, p = 0.001). They also had lower hostility (2.5 ± 0.6 vs. 2.6 ± 0.6, p = 0.001), less depressive symptoms (2.0 ± 0.6 vs. 2.1 ± 0.7, p < 0.001), and their mothers reported slightly lower intolerance (1.40 ± 0.20 vs. 1.42 ± 0.20, p = 0.010) and more emotional warmth towards them ( − 0.69 ± 0.18 vs. − 0.68 ± 0.18, p = 0.049). Included participants also had higher rewards (3.8 ± 0.7 vs. 3.7 ± 0.7, p = 0.002).

Characteristics and bivariate correlations

Table I presents characteristics of the study variables. Table II presents the bivariate correlations. Intolerance and low maternal emotional warmth in childhood were correlated with higher adulthood job strain and lower job control. Intolerance and low emotional warmth were positively correlated with each other. Also, all work stress measures correlated with each other. Higher maternal intolerance correlated with lower age of the participant and participant gender, men having experienced higher maternal intolerance in their childhood. Low emotional warmth in childhood was correlated with participants' lower adulthood education and lower occupational status, and with lower maternal education. In addition, maternal intolerance in childhood correlated with lower adulthood occupational status.

Table I.  Characteristics of study variables (N = 823).

Variable (range or unit)	Mean/number	SD%
Age (years)	30.6	(4.4)
Gender
Women	443	53.8%
Men	380	46.2%
Educational level (1–3)	2.30	(0.55)
Occupational status (1–3)	2.00	(0.77)
Maternal education (years)	10.2	(3.1)
Childhood family income (1–8)	5.96	(1.90)
Mother's mental problems
No	811	98.5%
Yes	12	1.5%
Hostility (1–4.73)	2.48	(0.57)
Depressive symptoms (1–4.54)	1.98	(0.63)
Intolerance* (1–2)	1.40	(0.20)
Low emotional warmth* ( − 1–0.24)	− 0.69	(0.18)
Job demands (1–5)	2.88	(0.64)
Job control (1.22–5)	3.81	(0.76)
Rewards (1.33–5)	3.81	(0.69)
Job strain ( − 3.78–2.78)	− 0.93	(0.89)
Effort–reward imbalance* ( − 0.64–0.54)	− 0.13	(0.14)

* Transformed: intolerance with square root transformation, low emotional warmth with inverse transformation, effort–reward imbalance with logarithmic transformation.

Table II.  Bivariate correlations between the study variables (N = 823).

		1.	2.	3.	4.	5.	6.	7.	8.	9.	10.	11.	12.	13.	14.	15.	16.
1.	Age (years)	1
2.	Gender*	0.053	1
3.	Education	0.002	0.051	1
4.	Occupation	0.117^†	0.039	0.613^†	1
5.	Maternal edu	− 0.126^†	− 0.067	0.246^†	0.234^†	1
6.	Income	0.039	− 0.029	0.222^†	0.221^†	0.378^†	1
7.	Mental prob^‡	0.031	− 0.050	0.007	− 0.014	− 0.012	− 0.024	1
8.	Hostility	− 0.016	0.029	− 0.138^†	− 0.174^†	− 0.092^†	− 0.101^†	− 0.026	1
9.	Depressiv	0.062	0.130^†	− 0.083^¶	− 0.111^†	− 0.060	− 0.042	− 0.063	0.672^†	1
10.	Intolerance^§	− 0.377^†	− 0.082^¶	− 0.057	− 0.086^¶	0.055	− 0.020	− 0.011	0.112^†	0.068	1
11.	Low emot^§	0.061	− 0.044	− 0.089^¶	− 0.074^¶	− 0.099^†	− 0.054	0.013	0.074^¶	0.095^†	0.332^†	1
12.	Job demands	0.112^†	0.061	0.170^†	0.171^†	0.023	0.009	0.049	0.161^†	0.233^†	− 0.025	0.008	1
13.	Job control	0.139^†	− 0.084^¶	0.275^†	0.393^†	0.164^†	0.142^†	− 0.008	− 0.253^†	− 0.289^†	− 0.109^†	− 0.120^†	0.199^†	1
14.	Rewards	− 0.035	− 0.025	0.055	0.129^†	0.062	0.104^†	− 0.062	− 0.319^†	− 0.385^†	− 0.055	− 0.048	− 0.093^†	0.458^†	1
15.	Job strain	− 0.038	0.116^†	− 0.112^†	− 0.211^†	− 0.123^†	− 0.114^†	0.042	0.332^†	0.414^†	0.075^¶	0.108^†	0.550^†	− 0.709^†	− 0.457^†	1
16.	Effort–rew^§	0.095^†	0.062	0.086^¶	0.047	− 0.026	− 0.058	0.073^¶	0.317^†	0.400^†	0.012	0.025	0.788^†	− 0.122^†	− 0.667^†	0.671^†	1

Note: maternal edu, maternal education (years); income, childhood family income; mental prob, mother's mental problems; depressive, depressive symptoms; low emot; Low emotional warmth; and effort–rew, effort–reward imbalance.

* Gender coded so that higher values represent women

^† correlation is significant at the 0.01 level (2–tailed)

^‡ higher values present having mental problems

^¶ correlation is significant at the 0.05 level (2–tailed)

^§ transformed, intolerance with square root transformation; low emotional warmth with inverse transformation; effort–reward imbalance with logarithmic transformation.

Deficient maternal nurturing attitudes in childhood predicting work stress in adulthood

Table III presents the results of linear regression analyses on deficient maternal nurturing attitudes assessed in participants' childhood/youth predicting job strain and effort–reward imbalance and their components in participants' adulthood (n = 823). Intolerance predicted lower job control (β = − 0.109, p = 0.002) and higher job strain (β = 0.075, p = 0.032) in the unadjusted model. However, the associations were attenuated to non-significant when adjustments were added to the analyses. In the age- and gender-adjusted model, intolerance also predicted rewards (β = − 0.081, p = 0.031). This association was also attenuated to non-significant when all adjustments were conducted (step 5). Low emotional warmth by the mother predicted lower job control (β = − 0.120, p = 0.001) and higher job strain (β = 0.108, p = 0.002) in the unadjusted analyses. The associations for low emotional warmth remained significant after adjustments for age, gender, educational level, occupational status, maternal education, childhood family income, maternal mental problems, and participant hostility and depression.

Table III.  Linear regression analyses of maternal intolerance and low emotional warmth predicting work stress and its components (n = 823).

	Demands/efforts				Job control				Rewards				Job strain				Effort–reward imbalance
	Adjusted R^2	R^2 change	Beta	p	Adjusted R^2	R^2 change	Beta	p	Adjusted R^2	R^2 change	Beta	p	Adjusted R^2	R^2 change	Beta	p	Adjusted R^2	R^2 change	Beta	p
Intolerance*
Step 1	− 0.001	0.001	− 0.025	0.467	0.011	0.012	− 0.109	0.002	0.002	0.003	− 0.055	0.116	0.004	0.006	0.075	0.032	− 0.001	< 0.001	0.012	0.727
Step 2	0.012	< 0.001	0.024	0.529	0.029	0.005	− 0.073	0.049	0.004	0.006	− 0.081	0.031	0.017	0.005	0.080	0.034	0.012	0.003	0.060	0.109
Step 3	0.043	0.001	0.035	0.342	0.179	0.003	− 0.055	0.108	0.025	0.005	− 0.077	0.039	0.065	0.004	0.072	0.049	0.021	0.004	0.066	0.077
Step 4	0.044	0.001	0.035	0.340	0.178	0.003	− 0.055	0.108	0.027	0.005	− 0.078	0.038	0.065	0.004	0.072	0.048	0.025	0.004	0.066	0.076
Step 5	0.104	< 0.001	0.005	0.887	0.234	0.001	− 0.027	0.420	0.166	0.001	− 0.032	0.358	0.211	0.001	0.027	0.432	0.190	< 0.001	0.016	0.634
Low emotional
Warmth*
Step 1	− 0.001	< 0.001	0.008	0.827	0.013	0.015	− 0.120	0.001	0.001	0.002	− 0.048	0.165	0.010	0.012	0.108	0.002	− 0.001	0.001	0.025	0.469
Step 2	0.012	< 0.001	0.003	0.921	0.042	0.018	− 0.134	< 0.001	< 0.001	0.002	− 0.048	0.173	0.025	0.014	0.117	0.001	0.009	< 0.001	0.022	0.522
Step 3	0.042	< 0.001	0.019	0.573	0.185	0.009	− 0.096	0.003	0.021	0.001	− 0.036	0.301	0.069	0.009	0.096	0.005	0.018	0.001	0.026	0.454
Step 4	0.043	< 0.001	0.019	0.579	0.184	0.009	− 0.096	0.003	0.023	0.001	− 0.036	0.307	0.070	0.009	0.096	0.005	0.022	0.001	0.026	0.462
Step 5	0.104	< 0.001	− 0.003	0.931	0.239	0.005	− 0.075	0.015	0.165	< 0.001	− 0.002	0.941	0.214	0.004	0.062	0.049	0.190	< 0.001	− 0.010	0.745

Note: step 1, no controls; step 2, controlled for age; step 3, controlled for age, educational level, occupational status, maternal education, and childhood family income; step 4, controlled for age, educational level, occupational status, maternal education, childhood family income, and maternal mental problems; and step 5, controlled for age, educational level, occupational status, maternal education, childhood family income, maternal mental problems, participants' depression and hostility.

* The mean of assessements in 1980 and 1983.

Several analyses were conducted. When the significance values were adjusted for the number of analyses (two independent variables: intolerance and low emotional warmth, and three dependent variables: demand, control, and reward), a new significance value of p = 0.008 was produced (0.05/6 = 0.008). According to this value, almost all associations of intolerance were non-significant. However, most results related to low emotional warmth remained significant. Only associations in step 5 were attenuated to marginally significant (Table II).

Table IV presents the results of linear regression analyses that included both variables assessing deficient maternal nurturing attitudes in the same analyses (n = 823). These analyses show that low emotional warmth was, in general, still a significant predictor of job control and job strain when intolerance was included into the analyses. On the contrary, the effect of intolerance was generally not significant in these analyses.

Table IV.  Linear regression analyses of intolerance and low emotional warmth simultaneously predicting work stress and its components (n = 823).

		Job control				Job strain
		Adjusted R^2	R^2 change	Beta	p	Adjusted R^2	R^2 change	Beta	p
Step 1	Intolerance*	0.018	0.005	− 0.078	0.034	0.011	0.002	0.044	0.233
	Low emotional warmth*	0.018	0.008	− 0.095	0.010	0.011	0.008	0.094	0.011
Step 2	Intolerance*	0.041	< 0.001	− 0.021	0.594	0.025	0.001	0.037	0.361
	Low emotional warmth*	0.041	0.014	− 0.126	0.001	0.025	0.009	0.104	0.006
Step 3	Intolerance*	0.185	< 0.001	− 0.018	0.620	0.069	0.001	0.039	0.329
	Low emotional warmth*	0.185	0.007	− 0.090	0.009	0.069	0.006	0.082	0.026
Step 4	Intolerance*	0.184	< 0.001	− 0.018	0.618	0.070	0.001	0.039	0.325
	Low emotional warmth*	0.184	0.007	− 0.090	0.010	0.070	0.006	0.082	0.026
Step 5	Intolerance*	0.238	< 0.001	0.004	0.908	0.213	< 0.001	0.002	0.962
	Low emotional warmth*	0.238	0.005	− 0.077	0.022	0.213	0.003	0.061	0.071

Note: in these analyses intolerance and low emotional warmth are placed in the same analyses. Step 1, no controls; step 2, controlled for age; step 3, controlled for age, educational level, occupational status, maternal education, and childhood family income; step 4, controlled for age, educational level, occupational status, maternal education, childhood family income, and maternal mental problems; and step 5, controlled for age, educational level, occupational status, maternal education, childhood family income, maternal mental problems, participants' depression, and hostility.

* The mean of assessments in 1980 and 1983.

Discussion

Our main finding was that low emotional warmth by the mother in the childhood of the participant predicted the participant's lower job control and higher job strain 18–21 years later in adulthood. These findings were not dependent on age, gender, childhood or adulthood socioeconomic conditions, maternal mental problems, participant hostility, or depressive symptoms. Furthermore, when intolerance and low emotional warmth were placed into the same model, low emotional warmth still predicted job control despite the significant correlation between these two predictors.

Much of the work stress literature focuses on work characteristics as a source of stress. Less attention has been paid to individual factors and even less to early experiences related to stress vulnerability. Our findings show that adulthood work stress may be predicted by variations in childhood environment. Our present findings are in line with previous research indicating that even ordinary variations in maternal nurturing are related to stress reactivity of the child (Hane and Fox 2006). Furthermore, our findings are in line with previous research suggesting that childhood early experiences may lead to long lasting stress vulnerability and dysregulation of physiological stress systems (Luecken 1998; Bremner and Vermetten 2001; Heim et al. 2002; Evans and Kim 2007).

The deficient nurturing attitudes by the mother explained approximately 0.5–1.5% of the variation in the offspring's later work stress (job control and job strain). As such, these proportions are rather small; they should not be expected to have a large effect on the life of an individual. However, at the population level such an effect is still important. Furthermore, we assessed only one aspect of childhood environment, deficient maternal nurturing attitudes. There are many other potential aspects (e.g. deficient paternal nurturing attitudes) that may as well affect later work stress, and their combined effect is likely to be more substantial.

Overall, we were able to explain only a small amount of variance in work stress (ranging from 10 to 24% in the final step, step 5) with the variables included in our study. This is to be expected because other factors (e.g. organizational factors and economic situation) also contribute.

In the current study, work stress was assessed by self-reported questionnaires on work characteristics. Therefore, we have information on work characteristics as the participants perceived them. Both objective work characteristics and individual stress vulnerability are likely to affect these perceptions. A stressful childhood environment may predispose the individual to stress and, therefore, to perceiving the characteristics of the work environment as more negative. Selection to more stressful jobs because of individual stress vulnerability is also possible, that is, stress vulnerability induced by early experiences may lead to selection so that participants with stressful early experiences end up in more stressful jobs through lower educational attainment. In the present study, the association between low emotional warmth and work stress remained after educational level and occupational group were taken into account.

In addition to affecting physiological stress vulnerability, deficient nurturance may have indirect stress-inducing effects. For example, child rearing may affect development of social skills (Steelman et al. 2002) which are important in the current working life where team work and personal networks are becoming increasingly significant. Therefore, deficient social skills are a likely source of stress.

According to several reviews, there is a relationship between psychosocial working conditions and CVD (Hemingway and Marmot 1999; Belkic et al. 2004; Kivimäki et al. 2006; Eller et al. 2009). This is clearly established among men and particularly so among men below the age of 55 years. In the more recent studies, psychological demands seem to dominate in the associations (Eller et al. 2009). The relationship between job strain and CVD among men seems to correspond to an etiological fraction in the order of 3–7%, and in men below the age of 55 years the etiological fraction seems to be still higher (7–16%; Karasek and Theorell 1990). This means that the childhood nurturing factor that has been shown in the present study can explain only part of the job strain–CVD relationship. Accordingly, our results do not necessarily reduce the importance of improvement of the psychosocial work environment although it brings a new dimension into the discussion.

Limitations and methodological considerations

In interpreting the current results, several limitations should be taken into account. Deficient nurturing attitudes were self-rated by the mothers and may, therefore, be affected by social desirability. Furthermore, mother's nurturing attitudes may to some extent reflect the characteristics of the child, for example, the child's temperamental difficulty. Child-rearing practices may be evoked by the characteristics and the behavior of the child. Another limitation is that our data included only maternal nurturing attitudes. At the time of the data collection in the early 1980s, the role of the fathers was not perceived to be as important as it is at present.

As maternal nurturance is assumed to affect general stress sensitivity, it could be expected to be associated with all work stress variables and not just some as was the case in the current study. Job demands and the variables included in the effort–reward imbalance model were not assessed with original measures, which may have led to somewhat poorer measurement, which in turn could explain this phenomenon. However, in our previous studies, we have found associations between the measures used in the current study and cardiac health (Hintsanen et al. 2005, 2007) which gives credence to our measures.

The effort–reward imbalance model predicts that a personal coping characteristic, over-commitment, leads to high effort and to maintaining of effort even in low reward conditions (Siegrist 1998). According to the model, high effort is maintained because being in control is self-rewarding for persons with over-commitment. In the long run, however, exhaustion may follow (Siegrist 1998), and this might lead to increased feelings of effort and imbalance between effort and rewards. In the present study, we did not assess over-commitment and could not, therefore, examine its effects.

In the current study, several separate analyses were conducted. When the significance level was adjusted to account for the number of analyses, maternal intolerance was in general not a significant predictor of work stress. Furthermore, the effect of low maternal emotional warmth was attenuated to marginal significance in the final step of the regression analyses. This indicates that the current findings should be interpreted cautiously and future studies are needed to confirm our findings.

We took maternal mental problems into account. However, we did not have a specific measure because it included mental problems of all levels without more specific categorization. Furthermore, the variance was rather low. When available, use of a more fine-grained measure is recommended.

Because of the limitations of our study, causal inferences are to be made with utmost caution, especially as there is a lack of previous prospective studies examining effects of deficient nurturing attitudes on stress vulnerability. However, despite the limitations, it is noteworthy that the effect of deficient maternal nurturing attitudes remained even after taking into account the potential confounders and mediators, i.e. childhood and adulthood socioeconomic factors, maternal mental problems, and participant hostility and depressive symptoms. It should also be noted that our study examined ordinary variation in maternal nurturance, i.e. we did not select participants with extremely low levels of nurturance. The magnitude of the association might be larger if groups of such low levels of nurturance were examined.

In this prospective longitudinal study, we were able to examine how deficient maternal nurturing attitudes assessed in the childhood of the participants affected participants' work stress in adulthood. To our knowledge, no previously published studies on this subject exist. Overall, few previous studies have been able to examine the effect of childhood environment on adulthood stress vulnerability prospectively over such a long follow-up period.

Conclusions

According to our results, deficient maternal nurturing attitudes in one's childhood might predict adulthood work stress 18–21 years later, although this finding needs to be confirmed in future studies. Our findings imply that more attention should be paid to individual variance in stress vulnerability in work stress research.

Acknowledgments

This study was supported by the Academy of Finland (grants 111056, and 124399 for LK-J; grants 117604, 124332, and 124271 for MK), the BUPA foundation (MK), the Signe and Ane Gyllenberg Foundation (LK-J and MH), and Yrjö Jahnsson Foundation (MH), Niilo Helander Foundation (MH), and Emil Aaltonen foundation (MH).

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.