Rigorous QA strategies can help eliminate errors

Detecting, classifying, analyzing, and managing errors can help avoid the repetition of preventable mistakes and minimize harmful effects, according to a research team from Boston.

Most radiologists have little knowledge of either their error rates or the nature of their errors, said Dr. Anna Marie O'Connell of Beth Israel Deaconess Medical Center in an education exhibit at RSNA 2007. About 4% of interpretations by radiologists in daily practice contain errors, but current reporting systems are relatively limited in their feedback.

"We all make mistakes," she said. "When analyzing errors, it is far more helpful to look at them as arising as a result of human or systems contributors, rather than seeking to place blame on individuals."

The regulatory authorities now require that peer-review systems be implemented for error detection and analysis. The emerging focus is on being proactive rather than reactionary regarding the safety of patients and employees, but no system in radiology for classifying errors is widely accepted. Most classification schemes concentrate on human error, and little attention is given to systems errors, O'Connell said.

"Humans make errors for a complex variety of reasons," she said. "Knowledge-based errors occur when you simply do not know what you are doing and do not recognize that what you are doing is wrong. Rule-based errors occur when you think you know what you are doing but apply a bad rule or do not know that a rule exists. Skill-based errors occur when the actions are okay, but the plan is inadequate to meet the intended goal."

Most errors are both latent and active. They may involve human performance and practice guidelines such as patient management, procedures, and the communication of findings. Common causes of communication errors are incomplete or inaccurate information, questionable consent and disclosure processes, questionable documentation, failure to perform preprocedure timeout, and failure to pass on critical results.

Detection and analysis of near-miss occurrences provides a useful database for implementing change aimed at minimizing errors, O'Connell said. A clinical decision is altered before harm occurs, or a clinician does not act on an erroneous radiological diagnosis. A provisional radiological report is corrected after the attending physician detects a missed finding.

The exhibit cited ways of managing errors: introduce anonymous error reporting detection systems, establish a just and blame-free culture to promote reporting or errors, implement a comprehensive root-cause analysis process, evaluate both latent and human factors, and assess outcome as well as accountability. Corrective actions must be reasonable, achievable, and measurable, and anticipating barriers to introducing change is essential.

"Monitor these corrective actions, assign responsibility, and set timelines," she said. "Continue to actively manage, analyze, and improve the error-reporting systems in your department."