Medical Error Reporting and Patient Safety: An Exploration of Our Underreporting Dilemma

Abstract

Studies suggest that the majority of hospital errors go unreported. Equally disturbing is that data surrounding near miss events that could have harmed patients has been found to be even sparser. At the core of any medical error reporting effort is a desire to obtain data that can be used to reduce the frequency of errors, reveal the cause of errors, and empower those involved in the healthcare delivery system with the insight required to design methods to prevent the flaws that allow mistakes to occur. Aligned with the adage that “we can’t fix what we don’t know is broke”, the question is raised why does underreporting exist? The likelihood of reporting medical errors is explored as a manifestation of culture. Factors studied include communication and feedback, teamwork, fear of retribution, and leadership support (top management and supervisor). Data is presented using a nationally recognized instrument—the Agency for Healthcare Research and Quality (AHRQ) Culture of Safety survey. Findings from the research are mixed with little positive relationship between the model and number of events reported although each factor is found to be positively associated with an employee’s perceived frequency by which near miss and no harm events are reported. While advances in patient safety have materialized, the act of employees’ actually reporting events still pales in comparison to the number of errors that have likely occurred, regardless of efforts to advance culture. To explore influencers beyond those found in the AHRQ Culture of Safety survey, an overlapping model is presented. This includes studying various underlying factors, such as understanding what constitutes a reportable event, ease of reporting, and knowledge of the processes supporting data submission, along with attempting to better assess the impact of the direct supervisor and incentives in influencing behavior. Findings suggest that these additional factors do contribute, albeit modestly, to the act of reporting errors. When adding tenure and patient interaction to the model, a higher percentage of the variance is explained. In terms of perceived frequency of reporting near misses and no harm events, this model yields similar results to the first, explaining approximately 28% of the variance. The two factors most positively associated with perceived frequency of reporting near miss and no harm events are communication and feedback and infrastructure —suggesting that some unexplored relationship may exist between the overlapping models.