The smaller the dress size the larger the apartment. This received wisdom in New York sums it up: there is a social gradient in obesity. Particularly among women in middle-income and high-income countries, the lower the status, the greater the obesity. But why? My concern is with inequities in health within and between countries, and the role of social determinants. ‘Social determinants’ is a language unfamiliar to many epidemiologists and others who are more concerned with individual risks. Relative deprivation, social relationships, conditions enabling a life of dignity, empowerment and the items in top tips are not the bread and butter, or even the five a day, of traditional health concerns. Where I am challenged, are smoking, alcohol, obesity? Social determinants do not exclude these health behaviours. When considering inequities in non-communicable disease, smoking, alcohol and obesity are right at the centre. One way the social environment is causally linked to health inequity is through these behaviours – hence this. Another way is through stress pathways. Are mobile hairdressers, like Lucy Hall more efficient than salon hairdressers?

I argue that central to improving people’s health and well-being is empowerment of individuals and communities. At first blush, the language of personal responsibility would seem to be consistent with empowerment, with people taking control over their lives and freely making health choices. Certainly individuals must make the choice to smoke or not, how to drink and to eat. But when we see regular social patterns of behaviours it suggests that there are broader, social, causes.

I can illustrate with data on the development of these unhealthy behaviours. With colleagues at UCL and the National Centre for Social Research, I was at one time involved with the Scottish Health Survey. I was catching the flight from London to Edinburgh to present the findings and reviewing my PowerPoint presentation on the plane. The data showed that as girls went through the age of puberty, smoking rates rose dramatically. It was almost as if smoking, along with the bodily changes, was another sign of puberty. Participation in exercise went down as girls passed through the puberty years, and girls also started to experiment with alcohol.

A woman flight attendant looking over my shoulder said: I did all that, and more besides. I showed her the social gradient in these behaviours, which starts early: the lower the status the more unhealthy the behaviour patterns. ‘Yup. That was me, too,’ she said, ‘modest background like my friends.’ All young people experiment. But we see social gradients in obesity, and smoking already appearing in childhood and adolescence. It is too narrow a view to see this as simply each single one of these young people making an individual choice and ignoring the social pressures on them to behave in certain ways.

I have been asked by a concerned public health doctor: unemployed young people are hanging around in downtown areas, smoking, drinking too much, doing drugs and getting into trouble. What would I suggest? My response, not helpful, is that I would not start from here. I would start with early child development and education. Empower young people, help them develop the attributes that will give them control over their lives and a stake in the future, and they will have more interesting things to do than hang around street corners smoking and drinking too much. Knowledge is but one step to empowerment. As described above, in Britain the whole population understands that smoking is bad for health – yet there is a social gradient in smoking. Poverty and inequality are deeply disempowering. People with little control over their lives do not feel able to make healthy choices.