Development and Validation of an Automated Measurement of Child Screen Media Use: FLASH

点击翻译按钮获取中文摘要

The obesity field is fraught with problems of measuring people's diet, physical activity and sedentary behaviors that are hindering the advancement of the field to find promising solutions for obesity prevention and treatment. An important example is screen media use, a prominent cause for sedentary time that has been linked to obesity and metabolic syndrome, as well as other unwanted physiologic, psychosocial, and academic outcomes in children. Given the important impact screen use has on children's health, our ability to measure the time children spend on screens is very limited. Current assessments typically rely either on (1) general questions estimating “typical use;” or self- or parent-report via screen use diaries, both which have reporting bias and recall bias; or (2) devises initially intended to restrict children's TV viewing (TV Allowance™), which may have intervention effects or miss-classify TV viewing time. While automatic and objective measures of physical activity have shown validity and reliability for quantifying children's physical activity, no system that is automatic, accurate and unobtrusive has been developed to assess children's screen use on different platforms. Advances in technology, such as person detection, accurate facial recognition based on images, and imaging, computer vision and signal processing algorithms now offer novel and promising solutions to objectively and automatically measure people's screen viewing behaviors. We will leverage these recent advances and integrate them to develop a first of its kind, in-home, unobtrusive, automatic, privacy preserving screen use monitoring system: Family Level Assessment of Screen use in the Home (FLASH) that uses an embedded computing platform connected to a video camera on larger, stationary screens (FLASH-TV); or functions as a background app using a front facing camera (FLASH-Mobile). Our trans-disciplinary group, consisting of behavioral researchers at Baylor College of Medicine (BCM) and electrical engineers at Rice University, propose to develop and asses the validity of the FLASH to accurately identify whether and for how long a child is using screen media devices. In this multiple-PI application, the development of FLASH is led by engineers at Rice. Once a final system has been developed, alpha and beta tested, a validation study will take place in observation labs by the BCM behavioral researchers with 6-10 year old children for FLASH-TV and FLASH-Mobile (n=43). Comparisons of FLASH output will be made to staff observations of children participating in a set of structured predefined activities. Next FLASH will be assessed for feasibility and accuracy for identifying children's screen use across platforms in a naturalistic home setting (n=46), compared to direct observation and screen use diaries. This project has the potential of creating a paradigm shift in scientist's ability to assess the role screen use has on health outcomes or intervention effects of screen media reduction programs, similar to that of accelerometers for physical activity research. FLASH has the potential of having a significant impact on public health and clinical research regarding screen media use.