Tong P Y, Enke-Miyazaki E, Bassin R E, Tielsch J M, Stager D R, Beauchamp G R, Parks M M
Wilmer Eye Institute and the Department of Molecular Biology and Genetics, Johns Hopkins University, Baltimore, Maryland, USA.
Ophthalmology. 1998 May;105(5):856-63. doi: 10.1016/s0161-6420(98)95026-6.
This study aimed to determine the ability of healthcare professionals and lay volunteers to grade photoscreening photographs.
The study design was a cross-sectional study.
One hundred children 3 years of age or younger received a complete ophthalmologic examination and were photographed using the Medical Technology Innovations (MTI) photoscreener. Twenty-six children had normal examination results, and the remaining 74 children had conditions that are of interest for pediatric screening, including strabismus, refractive error, media opacities, and ptosis. Eighteen volunteers, including pediatric ophthalmologists, pediatricians, ophthalmic technicians, health department nurses, Prevention of Blindness Society personnel, and Lions Club volunteers, graded each of the 100 photoscreening photographs.
Sensitivity and specificity of vision screening and of photograph grading were measured.
Results from various graders yielded sensitivities ranging from 37% to 88% and specificities ranging from 40% to 88%. No single grader achieved sensitivity and specificity both greater than 70%. The grading of the manufacturer's representative had a sensitivity of 43% and a specificity of 85%. Sensitivity decreased to 31% for strabismus and 18% for refractive error when the correct type of strabismus or refractive error was required to be considered true-positives. Results were not positively correlated with the ophthalmologic knowledge of the participant.
The wide variability in sensitivities and specificities among graders indicates inconsistent photograph interpretation skills or deficient screening guidelines or both. For off-axis photoscreening as implemented by the MTI system to become a useful vision-screening method, additional photograph interpretation skill transfer may be beneficial, although not necessarily sufficient.
本研究旨在确定医疗保健专业人员和非专业志愿者对照片筛查照片进行分级的能力。
研究设计为横断面研究。
100名3岁及以下儿童接受了全面的眼科检查，并使用医疗技术创新公司（MTI）的照片筛查仪进行拍照。26名儿童检查结果正常，其余74名儿童患有儿科筛查关注的病症，包括斜视、屈光不正、介质混浊和上睑下垂。18名志愿者，包括小儿眼科医生、儿科医生、眼科技术员、卫生部门护士、预防失明协会人员和狮子会志愿者，对100张照片筛查照片中的每一张进行分级。
测量视力筛查和照片分级的敏感性和特异性。
不同分级者的结果显示敏感性范围为37%至88%，特异性范围为40%至88%。没有一个分级者的敏感性和特异性均大于70%。制造商代表的分级敏感性为43%，特异性为85%。当需要将正确类型的斜视或屈光不正视为真阳性时，斜视的敏感性降至31%，屈光不正的敏感性降至18%。结果与参与者的眼科知识没有正相关。
分级者之间敏感性和特异性的广泛差异表明照片解读技能不一致或筛查指南不足或两者皆有。对于MTI系统实施的离轴照片筛查要成为一种有用的视力筛查方法，可能需要额外的照片解读技能传授，尽管不一定足够。
