Technician Training and Reproducibility of Drusen Measurement

Abstract

 

Purpose:

 

To determine whether training will improve reproducibility in measurements.

 

Methods:

 

A retinal surgeon (MD) and three trainees (TR) measured a single drusen of the same digital photograph daily over the course of sixteen days using the Topcon camera 50EX, OIS Winstation-XP 3200, V.10.2.44. Training was the time the trainee had exposure to the Topcon camera system. The MD trained each of the trainees. Each user was masked to any previous measurement. Reproducibility was evaluated with Shewhart process control analysis, Bland-Altman repeatability coefficients, and coefficient of variance, all techniques to analyze variability and reproducibility in measurements.

 

Results:

 

TRs ranged in training from four weeks to one day. TR 1 (4 weeks), TR 2 (3 days), TR 3 (2 days), MD (4 years). Lesion measurement: MD Range: 221-281um, TR 1 Range: 154-248 um, TR 2 Range: 221-328um, TR 3 Range: 228-315um.Shewhart analysis indicated that two measurements by the least trained user fell outside of two standard deviations, indicating the presence of variation in the measurement process.  

 

Conclusions:

 

The MD measurements had a smaller repeatability coefficient and coefficient of variance than the trainees suggesting a correlation between training and reproducibility. This small study suggests that ophthalmic measurements can be recorded and analyzed for improved training techniques.

 

Keywords: clinical (human) or epidemiologic studies: systems/equipment/techniques • age-related macular degeneration • drusen