Intra-examiner repeatability and agreement in accommodative response measurements.

PURPOSE Clinical measurement of the accommodative response (AR) identifies the focusing plane of a subject with respect to the accommodative target. To establish whether a significant change in AR has occurred, it is important to determine the repeatability of this measurement. This study had two aims: First, to determine the intraexaminer repeatability of AR measurements using four clinical methods: Nott retinoscopy, monocular estimate method (MEM) retinoscopy, binocular crossed cylinder test (BCC) and near autorefractometry. Second, to study the level of agreement between AR measurements obtained with the different methods. METHODS The AR of the right eye at one accommodative demand of 2.50 D (40 cm) was measured on two separate occasions in 61 visually normal subjects of mean age 19.7 years (range 18-32 years). The intraexaminer repeatability of the tests, and agreement between them, were estimated by the Bland-Altman method. We determined mean differences (MD) and the 95% limits of agreement [coefficient of repeatability (COR) and coefficient of agreement (COA)]. RESULTS Nott retinoscopy and BCC offered the best repeatability, showing the lowest MD and narrowest 95% interval of agreement (Nott: -0.10 +/- 0.66 D, BCC: -0.05 +/- 0.75 D). The 95% limits of agreement for the four techniques were similar (COA = +/- 0.92 to +/-1.00 D) yet clinically significant, according to the expected values of the AR. The two dynamic retinoscopy techniques (Nott and MEM) had a better agreement (COA = +/-0.64 D) although this COA must be interpreted in the context of the low MEM repeatability (COR = +/-0.98 D). CONCLUSIONS The best method of assessing AR was Nott retinoscopy. The BCC technique was also repeatable, and both are recommended as suitable methods for clinical use. Despite better agreement between MEM and Nott, agreement among the remaining methods was poor such that their interchangeable use in clinical practice is not recommended.