Childhood and adulthood socio-economic position and midlife depressive and anxiety disorders

This paper investigates how childhood socio-economic positioninfluences the risk for midlife depressive and anxiety disordersat 45 years of age, assessed by the Clinical Interview Schedulein 9377 participants of the 1958 British Birth Cohort. Socio-economicposition was measured by Registrar General Social Class inchildhood and adulthood. The association of paternal manualsocio-economic position with any diagnosis at 45 years of agewas accounted for after adjustment for adult socio-economicposition. Manual socio-economic position in women at 42 yearsof age was associated with midlife depressive disorder andany diagnosis; these associations were diminished by adjustmentfor childhood psychological disorders. Effects of childhoodsocio-economic position on adult depressive disorders may be mediated through adult socio-economic position.

INTRODUCTION

TOP

INTRODUCTION

This paper examines the implications of low socio-economic positionin childhood for midlife depressive and anxiety disorders inthe 1958 British Birth Cohort. We examine the following questions:does childhood socio-economic position increase the risk formidlife disorders above and beyond adult socio-economic positionand does the relationship between adult socio-economic positionand risk of midlife disorders reflect the effects of childhoodsocio-economic position?

Method

TOP

Method

The 1958 British Birth Cohort has been followed up at ages 7,11, 16, 23, 33 and 42 years with a biomedical follow-up at45 years.¹ After exclusions for death, emigration, permanentrefusal, armed forces and long-term non-contacts, 11 971 participantsstill in contact with the study at the age of 45 years wereinvited to a nurse-led biomedical assessment that includedmeasurement of respiratory function, eyesight, hearing and a computer-assisted personal interview. The achieved sample was9377 with a response rate of 72% of the contacted sample, representing59% of the eligible sample. Ethical approval for the surveywas given by the South East Multi-Centre Research Ethics Committee.

Adult social position was based on current or most recent occupationat 42 years of age and categorised using the British RegistrarGeneral classification² as non-manual (I and II professional/managerial/technical,III other non-manual) and manual (III skilled manual, IV andV unskilled manual). Both men and women were allocated to amanual or non-manual socio-economic position on the basis oftheir own occupation. Childhood socio-economic position at 7years of age was measured by the father's occupation usingthe same classification; no male head of the family was classifiedas manual.

Depressive and anxiety disorders in the previous week were measuredby modules from the revised Clinical Interview Schedule³ administeredby a nurse using a computer-assisted personal interview at the age of 45 years. Diagnoses were derived according to standardalgorithms for ICD–10 diagnoses. Non-comorbid diagnoseswere derived for `depressive episode', `generalised anxietydisorder' and a summary measure of `any diagnosis' that includedgeneralised anxiety disorder, depressive episode, any phobia(excluding specific phobias) and panic disorder, including anycomorbid disorders.

Psychological distress at ages 7 and 11 years was measured usingthe teacher-rated Bristol Social Adjustment Guides.⁴ Internalisingand externalising behaviours at age 16 years were measured using the teacher version of the Rutter scales.⁵

Logistic regression analyses were conducted to examine associationsof socio-economic position indices in childhood and adulthoodwith diagnoses at 45 years of age. Interactions between socio-economicposition and gender were examined to test whether social gradientsdiffered between women and men in relation to diagnosis. Mediatingor moderating effects of adulthood socio-economic positionon the association between childhood socio-economic positionand midlife disorder were tested by adjustments to the modelsand the inclusion of interaction terms respectively: adulthoodassociations were similarly adjusted for childhood socio-economicposition. Multiple imputation was used to address missing datain the analyses, using the ICE program in Stata SE (version8.2). All psychological health and socio-economic position variableswere included in the imputation equations.⁶ Inverse probabilityweights were derived to address attrition.

Results

TOP

Results

Of the 9377 participants, 50% were female, 84% owned or werepurchasing their home and 88% were employed. At age 45 years,6% of the total sample met the criteria for `any diagnosis'(7.2% of women; 4.8% of men). Among participants with paternalmanual socio-economic position at age 7 years, 55.3% had attainednon-manual socio-economic position by the age of 42 years.

Manual social position at 7 years of age was associated witha small increased risk for any diagnosis at 45 years of age(OR=1.26, 95% CI 1.02–1.57) that was eliminated afteradjustment for adult socio-economic position (online TableDS1). Manual socio-economic position at 42 years of age wasassociated with a significantly increased risk of any diagnosisand depressive episode in women but not in men (Table 1).No association was found between adult socio-economic positionand generalised anxiety disorder. Adjustment for socio-economicposition in childhood made only minimal change to the risksfor any diagnosis or depressive episode associated with adult socio-economic position. After additional adjustment for childhood psychological disorders these associations were diminishedin size but remained substantial in magnitude. Analysis ofcomplete data demonstrated the same pattern of results, albeitwith weaker associations; adjustment for childhood psychologicaldisorders removed the significant effect of manual midlifesocio-economic position on any diagnosis (online Tables DS2and DS3).

View this table:
[in this window]
[in a new window]

Table 1 Associations between social position at 42 years of age and Clinical Interview Schedule diagnoses at 45 years of age (n=9377)

Discussion

TOP

Discussion

This is one of very few studies examining links between childhood socio-economic position and midlife depressive and anxietydisorders.

Sample attrition was the main limitation although imputationand weighting was designed to minimise this. Missing data onless advantaged socio-economic position was associated withprevious poorer mental health. Generalisability is limitedby studying one age group, subject to specific cohort influences; nevertheless, the associations between socio-economic positionand psychiatric diagnoses are similar in nationally representativesamples. Registrar General Social Class is not always sensitiveto recent changes in occupational structure, may be less meaningfulin women and may not take account of partner occupational status.⁷

The loss of the small association between manual socio-economicposition in childhood and any diagnosis at midlife after adjustmentfor adult socio-economic position suggests that the effectsof childhood socio-economic position are largely mediated throughadult socio-economic position, through pathways that lead fromdisadvantaged childhoods to disadvantaged adulthoods.

The effects of less advantaged socio-economic position in adulthoodare not primarily the result of childhood socio-economic positionbecause adjustment for less advantaged socio-economic positionin childhood did not account for the association of adult socio-economicposition and midlife disorders. Any diagnosis and depressiveepisode were strongly influenced by adult social disadvantageas has been shown in other birth cohort studies^8,9 and cross-sectional studies.^10,11 In addition, decline in socio-economic statushas also been linked to increased risk of depression.¹² The stronger association in women than men was unexpected; associationsbetween socio-economic position and common mental disorderswere weaker in women than men in a national UK survey.¹⁰ Adult socio-economic position has a greater influence on depressiveepisode than on generalised anxiety disorder in this cohortpossibly because anxiety disorders may vary less by occupationallybased social class than by material disadvantage. It is alsopossible that non-manual participants are more comfortablein reporting anxiety symptoms than manual participants, thus obscuring an effect of social disadvantage. Although the associationof adult socio-economic position with any diagnosis and depressionis not explained by childhood socio-economic position, theassociation diminished slightly after adjustment for childhoodpsychological disorders. This may relate to the effect of childhoodpsychological disorders on adult socio-economic position; furtherwork is needed to investigate this.

ACKNOWLEDGMENTS

TOP

ACKNOWLEDGMENTS

The biomedical examination and related statistical analyseswere funded by Medical Research Council grant G0000934, awardedunder the Health of the Public initiative. C.C. is supportedby an Engineering and Physical Sciences Research Fellowship.B.R. is supported by Research Fellowships 148948 and 366758and by Program Grant 179805 from the National Health and Medical Research Council of Australia. Research at the Institute ofChild Health and Great Ormond Street Hospital for ChildrenNHS Trust benefits from research and development funding receivedfrom the NHS Executive.

REFERENCES

TOP