1.2.1 ‘Healthy’ and ‘successful’ ageing
3. Systematic Review
3.3.2 Study Characteristics
The 12 included articles used 10 different datasets and represented findings from a total of 35,022 individuals. Appendix 3.3 contains further details of each study’s variables, analysis techniques and results. The findings spanned five countries (the United States, United Kingdom, Finland, Sweden and Spain); representing Liberal, Scandinavian and Southern welfare regimes (refer to the previous chapter for a discussion of their specific characteristics). No studies were identified from Bismarckian or Post-Communist
9497 records identified via searches of electronic databases
Medline: 2702 Embase: 2795 PsycInfo: 1225 Web of Science: 1548
CSA Illumina: 1227
37 records identified via other sources References: 28
Citations: 9
7566 records after duplicates (N=1968) removed
80 full-text articles assessed for eligibility
7486 records excluded after title and abstract screening
68 articles excluded for not meeting study criteria
Irrelevant independent variables: 29
Irrelevant outcome: 22
Not published in scholarly journal: 8
Review articles: 5
Job, employment status or income mobility: 2
Age not specified/under 25 years: 2
12 articles selected for inclusion Cohort studies: 9
Cross-sectional studies: 3 (2 repeat cross-sectional studies)
Figure 3: Flow diagram of article identification, screening, eligibility and inclusion
regimes. There was little pattern in the associations by the type of welfare regime (Table 3.1). Five articles were specific to those aged over 50 years (Blane et al., 2004; Breeze et al., 2001; Houle, 2011; Otero-Rodriguez et al., 2011; Singh-Manoux et al., 2004), three studies included both the over 50s and under 50s in their study population (Laaksonen et al., 2007; Mäkinen et al., 2006; Marmot et al., 1998), and four studies were specific to those aged under 50 years (Huang & Sverke, 2007; Huurre et al., 2003; Johansson et al., 2007; Runyan, 1980).
Table 3.1: Studies included in the review categorised by welfare state regime type, showing the level of support for the association between life course socioeconomic position and quality of life
Bismarckian Liberal
Post-communist Scandinavian Southern
Country (N) - United Kingdom (3)
United States (3) - Finland (3)
Sweden (2) Spain (1) Support for an
association (N) - None (3)
Some (3) -
None (1) Some (2) Mixed (2)
Some (1)
Total 0 6 0 5 1
N=number of studies; None (no statistical support for an association); Some (some statistical support for an association but association may differ by gender, outcome, and life course model investigated); Mixed (results in opposite direction to that expected)
Two studies used data from the same repeat cross-sectional surveys of the Helsinki Health Study, but used different measures of socioeconomic position and analysis techniques (Laaksonen et al., 2007; Mäkinen et al., 2006). Two studies also used the Swedish Individual Development and Adaptation Cohort (Huang & Sverke, 2007;
Johansson et al., 2007); Johansson et al (2007) included a later survey wave and a different outcome. The publications by Breeze et al. (2001) and Singh-Manoux et al.
(2004) were based on the Whitehall I and II studies respectively, which included only British civil servants as the target population. Several outcomes were used, including wellbeing (using self-esteem, self-acceptance or positive psychological functioning indicators), life satisfaction, CASP-19, and the mental component summary (MCS) of SF-36. Nine studies assessed a single life course model and three assessed more than one (Table 3.2).
61
Table 3.2: Summary of articles included in the systematic review categorised by life course model First author
Year
Country Study design N
Measures of socioeconomic position Model Outcomes Results
Mäkinen
parent’s education level & childhood circumstances. Adulthood
socioeconomic position: own education level
Cumulative SF-36 MCS
No support.
father’s occupation. Own education level. Adulthood socioeconomic position: current/last occupation of household head
Cumulative Change in SF-36 MCS
Support for cumulative model – risk of decline in MCS increased linearly with increasing number of low socioeconomic positions.
father’s occupation & childhood socioeconomic circumstances. Own education level. Adulthood
socioeconomic position: employment grade
Cumulative SF-36 MCS
Support for cumulative model among men only – risk of being in lowest quintile increased linearly with increasing number of low socioeconomic positions.
Huurre
father's occupation. Adulthood socioeconomic position: own occupation
Latent
Wellbeing
Support for latent model among women only – lower childhood socioeconomic position associated with poorer wellbeing.
Marmot
parent’s education level. Adulthood socioeconomic position: own education level
Latent
Wellbeing
Some support for latent model among women who had mothers with lowest education – lower childhood
socioeconomic position associated with poorer wellbeing.
62
father’s occupation. Own education level. Adulthood socioeconomic position: current/last occupation of household head
Latent Change in
SF-36 MCS
Support for latent model – low childhood socioeconomic position associated with highest risk of decline and improvement in MCS.
parent’s education level, economic difficulties. Adulthood socioeconomic position: own education level, income &
occupation
Latent &
Pathway
SF-36 MCS
No evidence for latent model in men or women. Support for pathway model in men & women – higher adulthood socioeconomic position (via childhood socioeconomic position) associated with increased risk of lower MCS scores.
Mäkinen
parent’s education level. Adulthood socioeconomic position: own education level
Latent &
Pathway
SF-36 MCS
In women only – higher childhood socioeconomic position associated with increased risk of low MCS. Little support for pathway model in men or women.
Blane occupation & respondent’s longest held occupation. Intra-generational mobility:
respondent’s occupation aged 25 & 50 years occupation & current or last occupation of household head
Social Mobility (inter)
Change in SF-36 MCS
Support for social mobility – upwardly mobile more likely to experience change in MCS scores. No evidence for
downwardly mobile. occupation & respondent’s occupation aged around 38 years
Social Mobility (inter)
Life
satisfaction No support.
Breeze
employment grade at baseline &
employment grade at retirement
Social Mobility (intra)
SF-36 MCS Support for intra-generational effect – upwardly mobile less likely to have poor MCS score.
Houle 2011
Cohort
N=4992 Higher 100% male Intra-generational mobility: occupation aged around 36 years & 52 years
Social Mobility
(intra) Wellbeing No support for intra-generational effect when accounting for prior and current
63
United States Around 52
years
class – mobile individuals more likely to report wellbeing resembling current class than prior class.
Huang 2007 Sweden
Cohort
N=291 Average
100%
female 43 years
Intra-generational mobility:
respondent’s occupational history from ages 16 to 43 years
Social Mobility (intra)
Life
satisfaction No support.
Johansson 2007 Sweden
Cohort
N=514 Average
100%
female 49 years
Intra-generational mobility:
respondent’s occupational history from ages 16 to 43
Social Mobility (intra)
Life
satisfaction &
wellbeing
Life satisfaction: no support. Wellbeing:
some support – downwardly mobile reported lower wellbeing.
Inter=inter-generational; Intra= Intra-generational; MCS= mental component summary; N=Sample size; socioeconomic position= socioeconomic position; SF-36= short-form 36