Evaluating A Visual Annotation Tool For Weight Errors
Posted Date: Jun 20, 2019
- Investigator: Tzu-Yu Wu
- Specialties:
- Type of Study: Observational/Survey
Errors in body weight data can occur in data collection processes, leading to patient safety issues especially in pediatric medication and dosing. Since manually reviewing charts to identify and fix weight errors is not practical in busy clinical routines, there is a critical need for developing a reliable and valid computerized algorithm to automatically detect weight errors in real-time and in a retrospective manner. While the long-term goal of this project is to integrate the advanced algorithm with EHR system to alert weight errors and further support clinical decisions, a critical first step is to collect a large amount of expert-annotated weight data for algorithm training, which can be supported by a visual annotation tool developed by our research team. This visual annotation tool needs to be evaluated for its effectiveness and efficiency and validated for its design before being deployed to collect a large scale corpus of expert-annotated weight entry errors. This IRB protocol proposes a user-centered evaluation of the visual annotation tool and describes the actions taken to protect human subjects.
Criteria:
All Patients Who Visited Cincinnati Children's Hospital Medical Center (Cchmc) And Have At Least 4 Weights Taken And Stored In The Cchmc Electronic Medical Records. All Data Have Been De-Identified.
Keywords:
Pediatric Weights, Data Quality, Patient Safety
For More Information:
Danny Wu
513-558-6464
wutz@ucmail.uc.edu