Problems in comparisons of data for the prevalence of myopia and the frequency distribution of ametropia

There is currently great interest in comparing data for the prevalence of myopia in different parts of the world, particularly in view of the suggestion that, in recent decades, marked increases have occurred in prevalence among children and young adults in some areas. This work investigates the factors that affect the comparison and interpretation of sets of myopia prevalence data for different age groups, locations and dates.

Using data from the literature, the problems caused by the effect of the reliability and validity of the method used to measure refraction, the threshold chosen to define an eye as myopic are discussed. The influence of slow drifts in refraction with age is considered and it is recommended that if mean refractions at different ages are to be compared, the interpretation of the results should take account of the normal age-dependent trends in refraction. The value of specifying the distribution of refractive errors, rather than simply their mean and standard deviation, is emphasised and possible parametric fits to describe these distributions are reviewed.

There remains a need for greater standardisation in sampling strategies, refractive measurement procedures and definition of myopia in prevalence studies. The use of an ex-Gaussian or other approximations to describe the refractive error distribution appears to give useful insights into the nature of the changes that may occur with age and other factors.