Detection of improper fixation in MTI photoscreening images

The cost and yield of photoscreening: impact of photoscreening on overall pediatric ophthalmic costs

BACKGROUND

Approximately 5% of preschool-age children suffer from amblyopia. Many of them have high or unequal hyperopia. Amblyogenic risk factors frequently can be detected by photoscreening.

METHODS

Free photoscreening was offered to Alaskan children ages 1 to 5 from urban and rural screening hubs. Screened images were mailed to the Alaska Blind Child Discovery coordinating center for physician photoscreen interpretation, specifically seeking latent or anisometropic hyperopia. Parents and screeners then were mailed results and information about amblyopia. Follow-up examination data were tallied, and a cost-consequence analysis was developed for various vision screening paradigms and eye care.

RESULTS

From 1996 through 2003, a total of 13,255 screenings were performed with a positive interpretation rate of 4.7%. Penetrance of screening was 22% in urban and 44% in rural communities. Positive predictive value was estimated to be more than 90%. Average cost to screen and inform an Alaskan preschooler was approximately 10.67 dollars, and cost to detect amblyogenic risk factors by photoscreening in an Alaskan was approximately 206 dollars. Compared to American Academy of Pediatrics (AAP) 1995 guidelines, implementing photoscreening added 9%, while mandating complete prekindergarten examination added 49% to overall eye care.

CONCLUSIONS

MTI photoscreening achieved high community penetrance and high positive predictive value for latent hyperopia and other amblyogenic factors. When follow-up costs are considered, adding photoscreening to current AAP guidelines may add 112 dollars per child over 10 years, but probably would assist in the reduction of amblyopia. Penetrance of urban photoscreening likely will remain low unless pediatric vision screening guidelines and reimbursement are revised.

Calibration of photoscreeners for single-subject, contact-induced hyperopic anisometropia

BACKGROUND

In 1999, the threshold anisometropic hyperopia for photoscreening toddlers and preschoolers was determined to be 1.50 D. We compared crescent size from induced anisometropic hyperopia using three commercially available photoscreeners and compared them with miniature digital video cameras, which have a flash-to-lens dimension similar to two of the three photoscreeners. Photoscreeners were compared with a remote autorefractor.

METHODS

Anisometropia was induced by placing several known minus contact lenses (-1.00, -1.50, and -2.00 D) in the nondominant eye of a visually healthy, orthophoric emmetrope older than 3 years. Photographs were taken in light and dim conditions with all three cameras with and without the contact lenses. Corneal diameter, pupil diameter, and crescent location were determined. A protocol for flash initiation, focus, and distance was determined for the digital cameras.

RESULTS

Under light and dim ambient light, no camera yielded a significant hyperopic crescent for emmetropia (no contact lens) or induced 1.0-D anisometropia. We propose a simple measure, "delta center crescent" distance from the center of the pupil to the edge of the crescent. Delta center crescent less than 1.5 mm identified most examples of threshold hyperopic anisometropia. The ideal distance for photoscreening using the digital cameras is approximately 1.6 m, whereas the commercially available photoscreeners focus at 1.0 m.

CONCLUSIONS

We have developed a protocol for photoscreening and its interpretation with a miniature digital video camera weighing 400 g and costing approximately dollar 1,000. Threshold anisometropic hyperopia can be determined from digital still or audio-labeled video flash images if the pupils are at least 4 mm.

The threshold for the detection of strabismus

PURPOSE

To identify the threshold at which horizontal or vertical strabismus becomes reliably detectable by observers and to determine the effects of interpupillary distance, age, gender, and observer experience.

METHODS

Six models of different gender, age, and interpupillary distance were digitally photographed in several predetermined gaze positions off-axis in the horizontal and vertical planes. Standardized distance, zoom factor, and lighting were used. The images were digitally altered to exactly superimpose one eye deviated and one eye aligned with the axis of the camera. This simulated horizontal and vertical strabismus ranged from 2.5 to 20 prism diopters (PD). The images were arranged in random order and presented to groups of lay and professional observers, and their responses were recorded.

RESULTS

The statistically significant threshold for detecting esotropia, exotropia, and hypertropia was 12.5 PD. Hypotropia had a higher threshold of 20 PD. Observer experience and model age each had a significant effect on the ability to detect strabismus.

CONCLUSION

Our study demonstrates a unique method for assessing the significance of different types and degrees of strabismus. Our findings may be used to help patients with strabismus as they consider others' perception of their ocular misalignment. Physicians can also use this information in making decisions regarding surgery.