Scalable non-invasive pediatric cerebral visual impairment screening with the higher visual function question inventory (HVFQI)

Cerebral Visual Impairment (CVI), vision loss due to brain injury in early childhood, is the leading cause --- approximately 40% --- of bilateral visual impairment in children from industrialized countries and is the most rapidly growing cause among children in developing countries. Typical causes of CVI include abnormal brain development or brain damage, often consequential to birth-related complications such as hypoxic ischemic encephalopathy, meningitis, hydrocephalus, and head injury. The current gold standard for clinical diagnosis require a trained clinician to administer visual-motor integration tests in conjunction with a thorough review of a patient's clinical history. This approach does not scale to meet the needs of undiagnosed children with CVI. Recently, non-invasive screening methods such as administering a higher visual function question inventory (HVFQI), have been shown to accurately capture the observations of teachers and guardians of children. Analysis of those observations has been shown to have very high correlation with visual-motor integration tests (p values The HVFQI web app is an online clinical diagnostic tool and database system designed to gather participant responses to over 50 questions and follow-up questions. No personally identifying information is stored in the database. The web app provides three critical functions: (1) scalable and accurate administration of the HVFQI; (2) global coordination of screening efforts; and (3) a consolidated database for efficacy analysis studies and rapid iteration of the HFVQI to maximize impact, accuracy, and accessibility. The HVFQI consists of a question inventory, a scoring rubric, and a conditional intervention strategy list. The question inventory is carefully curated and designed to capture observations from teachers and guardians that may indicate specific pediatric higher visual function deficits (HVFDs). In many instances HVFDs are accompanied by normal visual acuity, making diagnosis difficult, and requiring multiple questions to elicit HVFDs. These questions are presented at random to avoid leading the participant. The responses are scored according to the HVFQI scoring rubric, which indicates which HVFDs may be present, and relevant intervention strategies. The rubric responds to varying degrees of affirmative responses in a 5-category Likert scale, which includes 3 non-applicable responses with different causes. The rubric also responds to categorical responses for multiple-choice and multiple-answer questions. The web app prepares a report of relevant intervention strategies for the participant, along with the questions and responses as context. The web app is also designed to coordinate international clinical efforts. The web application administrators, led by a superadministrator, can create, manage, and remove centers and staff corresponding to physical (or virtual) locations as demand grows. Each center has a local administrator that can authorize staff to administer the HVFQI, interpret and discuss the results and relevant strategies with participants. The data is stored in a central secure database, accessible only to authorized researchers, and the site administrator. All data are anonymous and non-personally identifying. Authorized researchers can analyze the results for their local centers to estimate efficacy and submit suggestions to update the HVFQI to meet the needs of local participants. Researchers authorized by the super-administrator have full access to the entire database and can perform global analysis for rapid improvement. The HVFQI web app is the result of a collaboration between the SeeLab of the Smith-Kettlewell Eye Research Institute, and the Kulkarni Group of the Computer Science Department at San Francisco State University.