OUP user menu

‘How are you?’: what do you mean?

Thierry Lang, Cyrille Delpierre
DOI: http://dx.doi.org/10.1093/eurpub/ckp083 353 First published online: 23 June 2009

For Anthelme Brillat-Savarin, a French writer and distinguished amateur of food, ‘What is health? It is chocolate!’ (‘Qu’est-ce que la santé? C’est du chocolat!’).1 Interviewing Henry David Thoreau, the problem of how to measure health would be solved as follows: ‘Measure your health by your sympathy with morning and spring’.2 For Balzac, at the end of the XIXth Century, a woman over 30 years of age could expect little from life and should consider herself as elderly, or at least, she had to persuade herself that this was so.3 What answer would these persons have given to the question, ‘How is your health in general?’, which is included in the OECD Health Data? Obviously, in the first examples cited health would have been governed by very specific conditions and the expectations of Balzac's heroine, Julie d’Aiglemont, would have been very low.

For >30 years, there has been a huge increase in publications concerning self-perceived health (SPH). Discussing perception and observation, Sen stated that self-reported morbidity has severe limitations and can be extremely misleading.4 Recent epidemiological reports invite us to raise this question again: what are we looking for when exploring perception of health?

SPH is indeed a very privileged indicator if we aim to respond to the WHO definition of health. The answer to the question ‘How is your health in general?’ is supposed to summarize perception of one's health far beyond the absence of illness, and to cover the social, physical and mental aspects of health. As it appears theoretically to be based on a wide, multidimensional definition of health, this measure of health is also very cheap. You do not need a sophisticated tool to measure it, it can be answered in a few seconds, little explanation is needed and the answer can be collected on the phone. The high predictive validity of self-reported health with respect to mortality has been repeatedly observed.5 As a result, SPH has been widely used in different international projects.6

The OECD advises caution in using SPH, for at least two reasons. The first is technical: in some countries, the response scale is asymmetric, skewed on the positive side (excellent, very good, good, fair and poor), whereas in others it is symmetric (very good, good, fair, poor and very poor).7 The other limitation pointed out by the OECD relates to the cultural backgrounds and traits. The latter observation has been further documented in the epidemiological literature. When looking at social inequalities in health, some results have been puzzling. Social inequalities in mortality were found to be high in France; in contrast, they were much lower when SPH was used. SPH predicts mortality. This has been confirmed, but prediction varies with the social status of the groups studied. Furthermore, prediction is better in upper classes in some countries and lower classes in others.8 Evidence from sociological works suggests that persons tend to rate their health in comparison to their social peers, according to their expectations.5 So the same ‘burden of disease’ may impact a person's perceived health differently according to social category.5 Moreover, this phenomenon appears to vary between countries, genders, according to age and probably to the disease involved. In other words, while a disease differs in impact on the SPH of men and women of different educational levels and social categories, these variations are themselves not uniformly observed in different countries.8

Although using perception of health has the advantage of incorporating an individual view of health and illness, such use can be seriously misleading where public policy is concerned. Sen stated that using this tool, the impoverished state of Bihar would have better health than Kerala, and better even than the USA.4 Recent reports suggest that using perception of health within a given country5,8 would also be misleading and would result in underestimation of social health inequalities, with severe consequences on health policies. Should we have second thoughts before using SPH for planning and assessing public health policies? The question deserves to remain open for debate …


View Abstract