Objective: To evaluate the reliability of strabismus assessment using telemedicine (TM) technology.
Design: Two prospective interobserver agreement studies. One study compared the agreement between a standard and a TM examination, whereas the other assessed agreement between two independent standard examinations.
Participants: Strabismus patients over 4 years of age examined at a remote community clinic and patients assessed in a strabismologist's urban practice.
Methods: Forty-two patients were examined in person by a pediatric ophthalmologist at the remote community and independently by a pediatric ophthalmology fellow by means of TM (TM-standard study). The TM examination was performed with the help of a qualified ophthalmic assistant at the remote telecommunication center using a Power Cam 100 camera, a Picture Tel Concorde 4500 teleconferencing system, and a 224 kilobyte bandwidth. For comparison, independent in person examinations were performed on 43 patients by both examiners (standard-standard study). Agreement was measured using unweighted kappa (k) for categorical data, the intraclass correlation coefficient (ICC) for continuous data, and percent agreement. The odds of disagreement with TM (comparing the TM-standard versus standard-standard studies) was assessed with logistic regression analysis.
Main outcome measures: Three parameters were assessed: (1) category of strabismus, determined by observation without cover test; (2) angle of deviation at 0.33 and 6.0 m; and (3) ocular muscle action.
Results: Agreement on the category of strabismus was good (k > 0.61) other than for vertical deviations. However, there was good to excellent agreement between TM and standard examinations on the vertical (ICC = 0.78) and horizontal (ICC = 0.79) angles of deviation with 6-m fixation with the cover test. Muscle ratings agreed within one point for the lateral, superior, and inferior rectus muscle actions in more than 90% of the eyes examined. Although good agreement was observed in the TM-standard study, it was inferior to the agreement in the standard-standard study. Examination by TM increased the odds of disagreement compared with examination in person by twofold to threefold.
Conclusions: Strabismus examination can be performed with a good level of reliability with the use of medium bandwidth video teleconferencing equipment. However, reduced reliability has been noted in the detection of small vertical deviations by inspection and in evaluating oblique muscle actions.