Abstract
Purpose :
Unlike surgery in other areas of ophthalmology, strabismus surgery has evolved slowly. One reason is that efforts to improve quality usually concentrate on one or two tasks in the surgical procedure. Due to the large number of variables it is difficult to detect the effect of such improvement with statistical significance. We develop a framework model for quantitative comparison of surgical procedures.
Methods :
The variability of the result of strabismus surgery is described in terms of variances that originate from (1) inaccuracy during measurement of the angle of strabismus, from (2) differences between patients like in anatomy or innervation, from (3) differences between ophthalmologists and orthoptists in deciding which eye muscles to displace and how much, from (4) inaccuracy of the surgical procedure, from (5) postoperative wound healing and from (6) influence of binocular vision (Schutte S, Graefes Arch 2009;247:399). We analysed inaccuracy of common procedures for strabismus surgery by first dividing their surgical workflow into tasks. We then identified error sources for each task within each surgical procedure and quantified their variances.
Results :
It was assumed that the errors had a Gaussian distribution and were independent of each other. This is not incontrovertible for all errors, but does permit addition of all variances to arrive at the overall variance of the surgical procedure which, in the case of surgery for horizontal strabismus, is reflected by the variance of the postoperative angle of strabismus. An improvement of a method for a task in the surgical procedure could then be translated into a reduction of the variance of the postoperative angle. We calculated that, if the accuracy of surgery would be 1.0mm, 0.5mm or 0.1mm, the variance of strabismus surgery would comprise 30%, 20% or 4% of the overall variance that includes variance resulting from measurement of the angle and other causes. The model predicted the effect of improved precision of measurement in recession, from 0.3 to 0.1 mm (SD), on overall accuracy of the surgical procedure, that improved from 0.6 to 0.53 mm (SD).
Conclusions :
We provide a framework model that provides a comprehensive insight into the variances of the tasks that constitute common surgical procedures for strabismus. It allows a quantitative comparison between surgical procedures and study of alternative solutions.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.