Abstract
Purpose :
The reported frequency of calibration error (CE) of Goldmann applanation tonometer (GAT), the gold standard of tonometry, has varied from 12.6% to 40%. We have earlier described several operational components in the maintenance of GAT. The purpose is to report our experience of running a comprehensive maintenance program and the development of error forecasting technique for GAT.
Methods :
In a year-long, prospective cohort study conducted at two tertiary eye care referral centers, we included 190 slit-lamp mounted GATs belonging to a single model (AT 900 C/M; Haag-Streit, Switzerland). The observers were masked to the CE readings. Qualified ophthalmologists did error checking and reporting while engineers maintained the tonometers. The team carried out CE check once a month, and repair of faulty tonometers, if any, within 24-hours. The main outcome measures were frequency of CE and factors contributing to the survival function of the tonometer.
Results :
The median age (range) of the tonometers was 10.7 (0.2, 25.1) years. The daily usage of the tonometers [median (1st, 3rd quartiles)] over a randomly chosen week was 16.6 (5.8, 22.6) times. The 95% limits of intra- and inter-observer agreement in the measurement of CE were within the reported limits. The frequency (95% confidence interval) of CE reduced from 22.6 (17.2, 29) % to 0.5 (0.1, 3) % at one-year (P< 0.01). The age of the tonometer [median (1st, 3rd quartiles)] did not differ between those needing [Study group, n= 63, 9.4 (5.1, 18.8) years] and not needing [Control group, n= 127, 10.7 (0.25, 19.8) years] repair (P = 0.24). The total number of repairs was 86. Forty-nine (25.7%) tonometers required one CE repair and their survival was 100%. Fourteen (7.3%) tonometers required >1 CE repair and their survival dropped to 40% at one-year.
Conclusions :
Our maintenance program for GAT was effective. User involvement and predictive maintenance were the highlights. The recurrence of unacceptable CE at a short interval might indicate wear and tear and could forecast the need for replacing the parts of the tonometer. ‘Number of Repairs’ could replace age of an instrument as the metric for suitability for use. The observations may have generalizability to other mechanical medical equipment.
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.