Purchase this article with an account.
Simona Degli Esposti, Omar Abdul Rahman Mahroo, Praveen J Patel, Adnan Tufail; Exploration of the phenomenon of regression to the mean in visual acuity measurements in patients with age-related macular degeneration. Invest. Ophthalmol. Vis. Sci. 2017;58(8):4696.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
In clinical trials and real-world data, gains in visual acuity are frequently highest in those with lowest baseline acuity. Some of this effect may relate to the phenomenon of regression to the mean. We explored this by analysing visual acuity measurements made on two occasions in untreated eyes, and assessing change in visual acuity for those with the highest and lowest acuities.
Best-corrected ETDRS visual acuity recorded at baseline (“visit 1”) and week 1 (“visit 2”) from the untreated fellow eyes of participants in the ABC trial (Bevacizumab for neovascular age related macular degeneration; Tufail et al., BMJ, 2010) were analysed. Participants seeing fewer than 5 letters on either occasion were excluded. Those in the top and bottom 10% at visit 1 were identified, and the means compared between visits (visit 2 minus visit 1). The top and bottom 10% at visit 2 were also identified, and their means compared between visits (visit 1 minus visit 2).
Ninety-nine patients were included. Mean (SD) age was 78.5 (6.9) years. Mean (SD) acuities were 63.7 (24.7) and 65.0 (25.1) for visits 1 and 2 respectively. For the top 10% at visit 1, the mean visual acuity changed by 0 letters between visits 1 and 2. For the bottom 10% at visit 1, mean visual acuity changed by +3.5 letters (p=0.14) by visit 2. When comparing backwards, for the top 10% at visit 2, the mean visual acuity changed by -1.25 letters (p=0.015), i.e. the mean for visit 1 was lower. For the bottom 10% at visit 2, the mean visual acuity changed by +4.4 letters (p=0.16), i.e. the mean for visit 1 was greater.
In this exploratory study, effects consistent with regression to the mean were observed. When eyes with the lowest 10% of visual acuities were identified at either visit, they appeared to improve their vision at the other visit irrespective of whether this was forwards or backwards in time. The effect appeared smaller/absent for the top 10%, which may be due to the lower variability in acuity measurements in this group. The regression to mean effect is greater the more the variability in the measurement.
This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.
This PDF is available to Subscribers Only