ALTERNATE COVER TEST (ACT) HOW RELIABLE IS IT?

Heather Johns

Abstract

PURPOSE. ACT is used to measure magnitude of a deviation but different endpoints may be used. Inter-& intra-examiner reliability for 2 methods was determined & methods compared.

METHOD. To determine reliability of 2 prism neutralization endpoints (1st neutral & midpoint of reversal), 2 experienced examiners measured near point phoria on 72 subjects. Measurements were repeated with examiners masked to each other & the first results. Paired t-tests/Wilcoxon signed rank tests were used to assess inter-& intra-examiner reliability & agreement between methods for each examiner within each session. Signed & absolute differences in phoria magnitude were calculated for different test sessions, examiners & methods. The influence of phoria magnitude was explored.

RESULTS. When 2 different examiners performed the same method within the same session (inter-examiner reliability) there were no significant differences between examiners. Average signed differences (ASDs) between the 2 examiners for each comparison was less than 0.5 pd with standard deviations (sds) less than 2 pd. The absolute differences ranged from 1.3 to 1.96 pd depending on examiner & method. Over 2 different test sessions with the same examiner using the same method (intra-examiner reliability), only examiner 1 using 1st neutral was significantly different between 2 sessions (p=0.038). ASDs between 2 test sessions for each comparison were less than 0.5 pd with sds less than 2 pd. Absolute differences ranged from 1.19 to 1.67 pd depending on test session & method. Comparisons of the two methods by the same examiner at a single test session were all significantly different (p < 0.005), but the differences (less than 1 pd between the 2 methods) were not clinically significant. ASDs were less than 1 pd with sd less than 2 pd. Absolute differences ranged from 1 to 1.58 pd depending on examiner & test session. Reliability was not influenced by large phoria magnitudes.

CONCLUSIONS. With experienced examiners either method provides high inter-& intra-examiner reliability (0.5 pd or less). Although the 2 methods differ statistically the differences are not clinically significant. Results are not associated with large phoria magnitudes.

Details

Year: 2001

Program Number: 3:00 pm

Author Affiliation: University of Houston

Co-Authors: Ruth Manny, Karen Fern, Ying-Sheng Hu

Co-Author Affiliation: University of Houston, University of Houston, University of Houston

Room: Room 108