July 2018
Volume 59, Issue 9
Open Access
ARVO Annual Meeting Abstract  |   July 2018
A ‘universal’ calculator to link data across myopia epidemiology studies
Author Affiliations & Notes
  • Noel A Brennan
    R&D, Johnson & Johnson Vision Care, Jacksonville, Florida, United States
  • Xu Cheng
    R&D, Johnson & Johnson Vision Care, Jacksonville, Florida, United States
  • Youssef Toubouti
    R&D, Johnson & Johnson Vision Care, Jacksonville, Florida, United States
  • Mark A Bullimore
    Marshall B Ketchum University, Fullerton, California, United States
  • Footnotes
    Commercial Relationships   Noel Brennan, Johnson & Johnson (E); Xu Cheng, Johnson & Johnson (E); Youssef Toubouti, Johnson & Johnson (E); Mark Bullimore, Johnson & Johnson (C)
  • Footnotes
    Support  None
Investigative Ophthalmology & Visual Science July 2018, Vol.59, 4151. doi:
  • Views
  • Share
  • Tools
    • Alerts
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Noel A Brennan, Xu Cheng, Youssef Toubouti, Mark A Bullimore; A ‘universal’ calculator to link data across myopia epidemiology studies. Invest. Ophthalmol. Vis. Sci. 2018;59(9):4151.

      Download citation file:

      © ARVO (1962-2015); The Authors (2016-present)

  • Supplements

Purpose : Refractive error in a population is not normally distributed, but is skewed and leptokurtotic. Epidemiological data for myopia are often presented in different ways, e.g., prevalence based on a certain dioptric value, mean, median, etc. Here, we develop a calculator to (i) enable conversion and comparison of prevalence data derived from different studies using varying criteria, and (ii) enable various descriptors of a distribution to be derived from a single value. The variance of the model across studies can also be used to assess consistency between populations, and whether a single ‘myopia’ factor drives the shape of the distribution.

Methods : A systematic review of the peer-reviewed literature identified over 100 papers with at least two descriptors of the refractive error distribution—typically mean and proportion with at least a given amount of myopia. Data were ultimately restricted to age 5 through to young adults. The log(odds of being more myopic than a specified criterion) was fit to mean spherical equivalent refractive error.

Results : 30 studies reported data on mean refractive error and the proportion of subjects with a given level of myopia for at least 530 subjects. Only one study was rejected because the data was inconsistent with the rest of the studies. The model was also found to be only applicable to children 5 years and above, teenagers and young adults and not for myopes above age 40. The following equation was developed:
ln(Odds) = –1.39μ – 0.11μ2 + 0.77α + 0.024αμ – 1.04
where μ is the mean refractive error and α is the criterion for myopia. R-squared was greater than 0.95. The graph plots the prevalence of myopia less than a certain dioptric value for different mean refractive errors in the group.

Conclusions : In general, refractive distributions can be described by knowledge of a single parameter; that is, the proportion of subjects above any level of myopia can be accurately predicted from the mean refractive error and vice versa. This information can be used to compare distributions described by different criteria and also in projecting future rates of high myopia in the population.

This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.



This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.