June 20, 2016

expert reaction to study on socioeconomic status and brain tumours

Socioeconomic status and risk of developing brain tumours is investigated in a paper published in the Journal of Epidemiology and Community Health with the authors reporting a higher risk of glioma for those with measures of higher socioeconomic status. A Before the Headlines analysis accompanied these comments.

Dr Jane Green, Clinical Epidemiologist at the University of Oxford, said:

“As the authors say, the fact that higher socioeconomic status has been linked to higher risk of brain tumours is not new. This paper is based on large numbers with good follow-up, and it is good to look at more than one measure – in this case, education, income and occupational status – because they tend to measure different aspects of socioeconomic status, which can be hard to define.

“The results generally confirm the links between higher socioeconomic status and risk of brain tumours. But I would interpret apparent differences between the risks in men and women and for the different tumour types with caution – the way the risks are reported here makes it hard to say which risk differences might be real.

“Even though this was a large study, by the time people had been divided into several income, occupational, and educational groups, the number of cases in some subgroups was small, meaning that the estimates of risk were not very precise (as shown by wide confidence intervals). Nor were risks between men and women, or by type, compared formally with each other [there are statistical tests to do this]: sometimes a risk will be ‘significant’ in one group and ‘non-significant’ in another just because the size of the group is different, not because the two risks are different from each other.

“For example, risk related to high university education for meningioma was (RR) 1.12 (95%CI 0.94-1.33) in men – ‘non-significant’, based on 180 cases – and 1.16 (1.04-1.29) in women ”significant’, based on over 500 cases. These two relative risks are probably not ‘really’ different- in fact they are clearly similar- and it is misleading to report one as associated and the other not, on the basis of their individual statistical significance.

“What this study cannot do (and the authors note this) is look at why there may be differences in risk by socioeconomic group. Of course, no one thinks that income or education themselves directly affect disease risk. This study used disease register information and had no data on individual characteristics or lifestyle factors which might explain the links observed (and do appear to explain similar links seen between socioeconomic status and risk of other diseases, such as some cancers).

“The authors mention smoking; two additional factors which might be interesting in this context are height and, in women, hormone replacement therapy. Risk of brain tumours (as of most cancers) is higher in taller people, and taller people tend to be richer and more educated. HRT increases risk of brain tumours, and its use tends to vary by socioeconomic group.”

Prof. Sir David Spiegelhalter, Winton Professor of the Public Understanding of Risk at the University of Cambridge, said:

“In spite of my degrees, I find these results deeply reassuring. For example, in each group of 3,000 men of the lowest educational level, we would expect five to be diagnosed with a glioma over 18 years. In 3,000 men with the highest educational level, we expect six gliomas.

“This is a classic example of where ‘big data’ can find results that are of ‘statistical’ but not of practical significance.

“The authors acknowledge the findings could be due to biases in detection, and in any case the study was really about socio-economic position in society.

“In spite of the strong suggestion in the title of the press release, this does not provide any evidence of a causal relationship with university education, and the relationship with occupation is weak. It should be a great relief.”

Declared interests

None declared

‘Socioeconomic position and the risk of brain tumour: a Swedish national population-based cohort study’ by Amal Khanolkar et al. published in the Journal of Epidemiology and Community Health on Monday 20 June 2016.