The Scottish Mental Survey 1932 linked to the Midspan studies: a prospective investigation of childhood intelligence and future health

Abstract

The Scottish mental survey of 1932 (SMS1932) recorded mental ability test scores for nearly all children born in 1921 and at school in Scotland on 1 June 1932. The Collaborative and Renfrew/Paisley studies, two of the Midspan studies, obtained health and social data by questionnaire and a physical examination in the 1970s. Some Midspan participants were born in 1921 and may also have taken part in the SMS1932, so there was a possibility that their mental ability data from childhood would be available. The 1921 born Midspan participants were matched with the computerized SMS1932 database, and in total, 1032 of 1251 people (82.5%) were matched successfully. Of those matched, 938 (90.9%) had a mental ability test score recorded. The mean score of the matched sample was 37.2 (standard deviation (SD) 13.9) out of a possible score of 76. The mean (SD) for the boys and girls was 38.3 (14.2) and 35.7 (13.9), respectively. This compared with 38.6 (15.7) and 37.2 (14.3) for boys and girls in all of Scotland. Graded relationships were found between mental ability in childhood, and social class and deprivation category of residence in adulthood. Being in a higher social class or in a more affluent deprivation category was associated with higher childhood mental ability scores, and the scores reduced with increasing deprivation. Future plans for the matched data include examining associations between childhood mental ability and other childhood and adult risk factors for disease in adulthood, and modelling childhood mental ability, alongside other factors available in the Midspan database, as a risk factor for specific illnesses, admission to hospital and mortality.

Introduction

The basis for associations between social position and health status is of considerable research and policy interest. Much social epidemiology has used different socio-economic indicators as though they were interchangeable, and rather few studies have examined associations between different indicators and outcomes in the same database. Only analyses of this latter kind can address the different interpretations of the social patterning of health-related behaviours and disease outcomes that have been advanced. With respect to education and health, observed associations have been attributed to the direct effects of education. Explanations have included the acquisition of knowledge regarding health-damaging behaviours, the ability to optimize use of health services, the development of time preferences favourable to health maintenance, an increasing willingness to invest in human capital, and the promotion of the psychological attributes of high self-esteem and self-efficacy.1., 2., 3. Alternatively, the associations could reflect education indexing socio-economic circumstances in childhood and both indexing and influencing socio-economic circumstances in adulthood. In the case of the health effects of occupational hazards in unskilled manual jobs, these may depend directly on work conditions, and reducing socio-economic differentials consequent on such exposures requires changes to work environments. In the Collaborative study, cigarette smoking was more strongly related to adulthood occupational social class than to age at leaving full-time education.⁴ This suggested that the social environment in adulthood maintains, or initiates, smoking behaviour. Age at leaving full-time education is, in this formulation, mainly related to smoking behaviour through its determination of occupational level in adulthood. However, age at leaving full-time education may have different associations than intelligence (as measured by standard psychometric tests) to health-related behaviours such as smoking and excessive alcohol use. No studies to our knowledge have been able to examine intelligence and a wide range of lifetime socio-economic measures in relation to health-related behaviours (such as smoking) and their physical concomitants (such as lung function and cholesterol levels).

We now describe how two established research teams have combined with the aim of studying risk factors for specific diseases and mortality. The first team is responsible for the Midspan prospective cohort studies conducted in the west of Scotland on adult men and women in the 1970s. These studies have generated novel findings on the use of social, lifestyle and health factors to predict mortality,4., 5., 6., 7., 8. hospital admissions,9., 10. and specific disease outcomes such as cardiorespiratory illness,¹¹ stroke¹² and cancer.¹³ The participants have been followed-up for mortality and cancer since inception, and have recently been linked with the Scottish morbidity records (SMR) system, enabling data on physical and mental illness to be investigated.

The second team is responsible for follow-up studies of the Scottish mental survey 1932 (SMS1932). The SMS1932 contains mental ability test scores for children born in 1921 and at school in Scotland on 1 June 1932. Follow-up studies of the SMS1932 have so far traced and examined psychosocial factors and health in participants of the SMS1932 in Aberdeen and Lothian. Intelligence quotient (IQ) at 11 years of age proved moderately stable to 77 years of age,¹⁴ related to birth weight,¹⁵ health status at 77 years of age,¹⁶ and lifetime psychiatric contact,¹⁷ and was a risk factor for mortality,¹⁸ and for senile but not presenile dementia in later life.¹⁹

The age ranges of the Midspan studies were such that they included men and women born in 1921. Some of the participants could therefore have mental ability data available from 11 years of age. By collaborating and sharing data, the SMS1932 and Midspan research teams anticipate improved understanding of risk factors for diseases and mortality. This paper describes the process and outcomes of the matching exercise.