Room: Exhibit Hall | Forum 3
Purpose: Investigate and identify error trends associated with radiotherapy electronic charts via proprietary chart checking process
Methods: Utilizing Visual Basic, an Excel-based GUI was created to facilitate chart checking procedures. An API script was then employed to extract basic patient and plan data from Eclipse into a CSV file, and subsequently imported into the database. The process contains items for 80+ checks of chart integrity ranging from plan quality, to field naming convention, to required documentation. Additional checks are populated for special procedures such as SRS and DIBH. For each check, a â€œNeeds Attentionâ€? checkbox exists in order to document checks that require intervention/modification to the plan by the chart checking physicist. Data from over 3000 initial charts spanning over 2 years was compiled, sorted, and analyzed to identify patterns for checks that were marked as â€œNeeds Attentionâ€?. For analysis, the checks were sorted into categories based on items that apply to every chart (ex. Prescription) and items that apply to specific charts (ex. bolus or 4DCT).
Results: It was found that, on a weekly basis, an average of 27% of the initial charts checked had errors related to the treatment plan with the most common error (10% of treatment planning errors) related to proper contouring. Additionally, an average of 16% of charts had prescription related errors with the most common errors related to dose/fractionation and energy matching treatment plan (~15% each). The most common error related to specific plan types involved plans with Implanted Electronic Devices where 45% of those plans required modification.
Conclusion: The ability to accurately track identified errors within the chart checking process will enable our clinic to focus on gaps in training and re-education regarding plan creation and finalization. Further research will entail isolating error patterns that may occur throughout the year.