Re: Evidence for an "epidemic" of myopia.

In the Annals January 2004 theme issue on Myopia, Park and Congdon 1 have disputed the reality of the increases in prevalence of myopia in East Asia for several reasons. In particular, they criticise the 2 best documented sets of cohort data, from Taiwan 2-6 and Singapore, 7-10 on methodological grounds. In the case of the studies from Taiwan, they argue that there are methodological uncertainties that undermine the validity of the studies. In the same issue of the Annals, Lin et al 3 outline the consistent methodologies and definitions used in their comprehensive set of data, and document increasing prevalence, earlier onset of myopia, and an increasing prevalence of high myopia. While Park and Congdon raise the theoretical possibility that the school sampling framework may have distorted the data, they provide no evidence that this is, in fact, the case. Thus, the data from Taiwan provide strong evidence of a cohort effect. In relation to the data from Singapore, their methodological criticisms have greater foundation, in that the early data on conscripts rely on measurement of visual acuity rather than actual refraction. However, some of the later data on conscripts have been obtained on samples on which refraction has been carried out, and thus clearly establish the recent high prevalences of myopia. 10 Park and Congdon suggest that the earlier data may be flawed because some of the low visual acuity may have been associated with hyperopic rather than myopic errors. However, if this was the case, then the prevalence of myopia would have been even lower 40 years ago, and thus the increase since then would have been even greater. Similarly, if their argument, that some of the low visual acuity may be explained by conscripts attempting to escape military service, were valid, this would again lower the real prevalence of myopia in the earlier studies, magnifying the increase that has actually taken place. Thus, neither of their hypotheses actually weakens the case for cohort effects. On a slightly different tack, Park and Congdon imply that there may be no epidemic of myopia in the sense of a cohort effect in which more recent birth cohorts have a much higher prevalence of myopia. They quote Rasmussen 11 as demonstrating that the prevalence of myopia was as high as 50% to 70% in China, at least since the 1920s. Unfortunately, this error perpetuates a misinterpretation of the work of …