Validating FMEA output against incident learning data: A study in stereotactic body radiation therapy

Area of Science:

• Radiation oncology
• Medical physics
• Quality improvement in healthcare

Background:

• Failure Mode and Effects Analysis (FMEA) is increasingly used for risk assessment in radiation therapy.
• Validation of FMEA output against actual occurring errors in radiation therapy has been limited.
• Stereotactic Body Radiation Therapy (SBRT) planning involves complex processes with potential for errors.

Purpose of the Study:

• To conduct an FMEA of the SBRT treatment planning process.
• To validate FMEA findings against data from an incident learning system.
• To compare the comprehensiveness of FMEA and incident learning for identifying process failures.

Main Methods:

• A multidisciplinary team performed FMEA on the SBRT treatment planning process.
• Failure modes were identified, rated for severity, occurrence, and detectability, and a Risk Priority Number (RPN) was calculated.
• FMEA results were compared with historical near-miss reports from a departmental incident learning system.

Main Results:

• FMEA identified 63 potential failure modes, with RPNs ranging from 60 to 336 for the highest risk items.
• The incident learning system contained 33 near-miss events related to SBRT planning.
• Combining both methods identified 76 total process failures; FMEA missed 17% of actual failures (lower risk), while the incident system missed high-severity failures not captured by FMEA.

Conclusions:

• FMEA is valuable but has limitations, failing to identify all actual failure modes, particularly lower-risk ones.
• Incident learning systems alone do not capture all potentially high-severity process errors.
• Integrating FMEA with incident learning systems offers a more complete understanding of risks in radiation therapy processes.