In Reply We thank Yang and colleagues for their comments in response to our recent publication.¹ We elected to use routinely collected electronic health records to allow a large sample size and the inclusion of a wide range of covariates. These records have been shown to be nationally representative in terms of age, sex, ethnicity, and socioeconomic deprivation. The records were collated from approximately 25% of the UK population, with high validity for diagnostic coding, including fractures, to reduce the impact of selection bias.^2-4 Indeed, the limitations of this approach have already been stated in our article, including the inability to assess visual function or hospital treatments and a likelihood to report more serious falls requiring medical attention. Yet, other approaches, such as self-reporting of falls in surveys, are subject to recall bias.⁵ A deeper assessment into the strengths and limitations of our population-based approach has also been explored in a helpful commentary by Pundlik and Luo.⁶

The potential influence of confounders has been already highlighted in our Outcomes and Covariates section within the Methods.¹ We controlled for a large number of potential (measured) confounders, but as in all observational studies, there is a risk of unmeasured confounding, which we acknowledge. Although no method can address for unmeasured confounding unless the mechanism is known, we have used advanced methods to try to minimize the risk of measured confounding. Thus, we conducted sensitivity analyses 鈥渦sing propensity scores to account for covariate imbalance between the cohorts within further Cox proportional hazard models.鈥�¹ These 鈥渋nverse-weighted probability models included more than 50 confounders,鈥�¹ with adjustments listed in eTable 4 in Supplement 1. Both approaches showed similar results, giving weight to our findings.

Corresponding Author: Jung Yin Tsang, MRes, NIHR School for Primary Care Research, Centre for Primary Care and Health Services Research, Sixth Floor, Williamson Bldg, Oxford Road, University of Manchester, Manchester M13 9PL, United Kingdom (jungyin.tsang@manchester.ac.uk).

Published Online: April 25, 2024. doi:10.1001/jamaophthalmol.2024.1030

Conflict of Interest Disclosures: Dr Tsang reported receiving grants from the National Institute for Health and Care Research. Dr Ashcroft reported receiving grants from the National Institute for Health and Care Research, AbbVie, Almirall, Celgene, Eli Lilly, Novartis, UCB, and the Leo Foundation. No other disclosures were reported.