Assessing the Inter-Rater and Inter-Trial Reliability of the NeurOptics Pupillary Light Response-3000 Pupillometer

Background An automated pupillometer is a handheld device used to stimulate the pupillary light response (PLR) and track the entirety of the response from constriction to dilation. Pupillometers provide objective data that clinicians can use to identify and assess brain injury. The validity of these devices has been previously established; however, the inter-rater and inter-trial reliability are unknown. Purpose The purpose of this study was to assess the inter-rater and inter-trial reliability of the NeurOptics PLR-3000 pupillometer device in measuring pupil size changes, constriction velocities, and dilation velocities. The authors hypothesized that inter-rater and inter-trial reliability would have intraclass correlation coefficients (ICC) greater than or equal to 0.70 for all PLR parameters. Study Design: Observational, reliability study Methods: Forty-eight healthy adults (age 18-40 years) without a history of neurological injury, optical surgery, or cognitive impairment participated. Two independent raters used the NeurOptics PLR-3000 to measure PLR parameters in the left and right eyes of each subject. Data for the average and individual trials of each PLR parameter were used to determine inter-rater and inter-trial reliability, respectively. Inter-rater and inter-trial reliability was evaluated using descriptive statistics, ICC, the standard error of measurement, Bland-Altman plots, and the minimal detectable change. Results Seven out of eight NeurOptics 3000-PLR parameters demonstrated moderate-to-excellent inter-rater (ICC range 0.72-0.96) and good-to-excellent inter-trial reliability (ICC range 0.76-0.98). The 75% recovery time parameter exhibited moderate inter-rater (ICC range 0.64-0.67) and poor-to-moderate inter-trial (ICC range 0.41-0.65) reliability. Conclusion The NeurOptics 3000-PLR demonstrated acceptable reliability in measuring initial and end pupil size, constriction and dilation velocity, and latency to change between different users and trials. However, the device exhibited unacceptable reliability when measuring