Validation of computer morphs for grading contact lens complications

Abstract

Abstract An alternative to the use of printed grading scales for assessing the severity of contact lens-induced ocular pathology is to use computer-generated grading morphs. This technique involves manipulation of a continuously variable movie sequence until this matches the severity of the condition being observed. To validate the use of computer morphs, nine experienced optometrists were each invited to grade - to the nearest 0.1 increment - an image of each of 16 contact lens complications, using printed scales (Efron Grading Scales for Contact Lens Complications) and computer morphs (Efron Grading Morphs). This entire procedure was repeated approximately 2 weeks later, yielding a total data base comprising of 576 individual grading estimates. Good accuracy was achieved using computer morphs, as evidenced by the similarity between the mean of the test and retest grading estimates for the printed scales (2.8 +/- 0.7) and the computer morphs (2.6 +/- 0.8) (F1,8 = 2.7, p = 0.14). There was no difference in median reliability between the printed scales (+/- 0.41) and the computer morphs (+/- 0.43) (Z = 0.1, p = 0.95). Computer morphs are thus considered to have been validated in view of their accuracy and reliability compared with printed scales. In addition, computer morphs offer considerable utility in terms of computer-based clinical record keeping, teaching and learning, and clinical research.