Estimation of global visual impairment due to uncorrected refractive error

The paper by Resnikoff et al.¹ on the global magnitude of visual impairment caused by uncorrected refractive error published in the Bulletin of the World Health Organization in January 2008 does not refer to our previous publication² on the same topic in BMC Medicine that precedes their paper by about two years and is readily available in the public domain. The reason for this is unclear as our paper clearly shows up in a PubMed_® search for both “blindness” and “visual impairment”. In addition, the global estimates of blindness and visual impairment due to uncorrected refractive error made in our paper are know to the blindness community as they were presented at the World Ophthalmology Congress at Sao Paulo in February 2006. It is therefore surprising that Resnikoff et al. did not follow the usual norm of referring to previously published relevant literature. While there are differences in the inclusion criteria for studies in our paper and that of Resnikoff et al., with our paper using stricter criteria leading to fewer qualifying studies, and there were differences in the estimates of blindness and visual impairment due to uncorrected refractive error in the two papers, these should not be reasons for not referring to previously published work.

We estimated that globally there were 5 million persons who were blind due to uncorrected refractive error with distance vision worse than 3/60 in the better eye (plausible range 4–6 million), while Resnikoff et al. estimated this to be 8.2 million. Of particular note is that a large proportion of the estimate by Resnikoff et al. is due to the number in India, estimated as 3.15 million persons more than 50 years old who were blind due to uncorrected refractive error, which is 46% of their global estimate of 6.88 million for this age group. This is implausibly high and seems to have been influenced by data from a multistate survey done in India by Murthy et al., which reported that 5.34% persons older than 50 years had presenting visual acuity worse than 3/60 in the better eye which dropped to 3.37% with best correction, suggesting a very high improvement of 37% with refractive correction.³ The estimate by Resnikoff et al. for India seems to be a direct application of this 1.97% absolute improvement to the approximate 160 million persons aged more than 50 years old in India in 2004 (in order to arrive at 3.15 million persons blind due to uncorrected refractive error in this age group). However, the paper by Murthy et al. also mentions that a fifth of the blindness was due to uncorrected or poorly corrected refractive error, which was for blindness worse than 6/60. Presumably this proportion would be lower for blindness worse than 3/60. This is at odds with the 37% blindness worse than 3/60 due to refractive error that could be derived from the data presented for improvement with refractive correction, which was apparently used in the estimate by Resnikoff et al., indicating that methodological issues related to these data and calculations need to be looked into carefully.

Our estimate for the number of persons in India who are blind due to uncorrected refractive error was less than half that made by Resnikoff et al.^1,2 Undoubtedly our estimates could be refined with further availability of data but we believe that the global estimate of 8.2 million blind persons due to uncorrected refractive error by Resnikoff et al. is an overestimate, largely due to the inclusion of an implausibly high estimate for India. While on the one hand we should not overlook blindness due to uncorrected refractive error as it can be addressed relatively easily, on the other hand we should be careful not to swing the pendulum in the other direction by overestimating it. Related to this issue, we have also published a proposal for revision of the definitions of blindness and visual impairment in the International Statistical Classification of Diseases that would take into account the inclusion of refractive error as a cause of blindness and visual impairment.⁴ ■