Click here to

Session: Quality and Safety in Radiotherapy II [Return to Session]

Using FMEA to Evaluate the Changes in Patient Risk for a New Automated Breast Planning Workflow

K O'grady*, L KUO, G Tang, J Moran, L Hong, C Della Biancia, K Ohri, A Ng, E Lin, L Santanam, Memorial Sloan Kettering Cancer Center, New York, NY

Presentations

SU-F-BRA-8 (Sunday, 7/10/2022) 2:00 PM - 3:00 PM [Eastern Time (GMT-4)]

Ballroom A

Purpose: Using failure mode and effects analysis (FMEA) per TG100 to evaluate the change in risk prior to clinical implementation of an in-house automated breast planning workflow.

Methods: A committee of physicists and dosimetrists identified failure modes for manual and automated breast planning workflows. The committee scored each failure mode for occurrence (O), severity (S), and detectability (D). RPNs were calculated using the equation RPN=O*S*D. RPNs were assigned a hazard level of “low”, “medium”, or “high” with ranges of 1-36.9, 37.0-295.4, and 295.4 - 1000, respectively. Changes in RPN for each failure mode were calculated and compared between workflows. Finally, the average change in RPN and the average percentage difference were calculated.

Results: Incorporating FMEA into the software design process inspired additional QA checks in the automated workflow to mitigate risk. The FMEA committee identified a total of 20 failure modes (12 medium risks and 8 low risks) in the manual workflow and a total of 29 failure modes (27 low risks and 2 medium risks) in the automated workflows. There were 9 newly introduced failure modes in the automated workflow which were scored as low risk. Of the 20 failure modes mutually existing in both workflows, 18 decreased and 2 remained unchanged with automation. The three highest RPN scores in the manual workflow were reduced with automation from 166 to 14 (91.3%), 166 to 71 (57.6%), and 139 to 24 (82.7%). The average reduction of all failure modes was 47 points with an average percent difference of 55.1%.

Conclusion: By using FMEA to analyze pre-clinical software, we identified and then mitigated several risks in our breast planning workflow and significantly reduced risks when compared to our manual process.

Keywords

Treatment Planning, Software, Risk

Taxonomy

TH- External Beam- Photons: General (most aspects)

Contact Email

Share: