To develop an objective, structured observational tool to enable identification and measurement of hazards in the built environment when applied to audiovisual recordings of simulations by trained raters. Simulation-based facility design testing is increasingly used to optimize safety of healthcare environments, often relying on participant debriefing or direct observation by human factors experts. Hazard categories were defined through participant debriefing and detailed review of pediatric intensive care unit in situ simulation videos. Categories were refined and operational definitions developed through iterative coding and review. Hazard detection was optimized through the use of structured coding protocols and optimized camera angles. Six hazard categories were defined: (1) slip/trip/fall/injury risk, impaired access to (2) patient or (3) equipment, (4) obstructed path, (5) poor visibility, and (6) infection risk. Analysis of paired and individual coding demonstrated strong overall reliability (0.89 and 0.85, Gwet's AC1). Reliability coefficients for each hazard category were >0.8 for all except obstructed path (0.76) for paired raters. Among individual raters, reliability coefficients were >0.8, except for slip/trip/fall/injury risk (0.68) and impaired access to equipment (0.77). Hazard Assessment and Remediation Tool (HART) provides a framework to identify and quantify hazards in the built environment. The tool is highly reliable when applied to direct video review of simulations by either paired raters or trained single clinical raters. Subsequent work will (1) assess the tool's ability to discriminate between rooms with different physical attributes, (2) develop strategies to apply HART to improve facility design, and (3) assess transferability to non-ICU acute care environments.