We read, with great interest, the article by Grunebaum et al “Effect of a comprehensive obstetric patient safety program on compensation payments and sentinel events.” As longstanding patient safety advocates, we applaud Weill Cornell’s tremendous investment of time, effort, and money into this program.

This article raises for us several questions:

What were the criteria used to classify an incident as a “sentinel event” in the study? Although the authors note The Joint Commission’s definition of a sentinel event, their own was much more narrowly defined (“at our institution, sentinel events included maternal deaths, and serious newborn injuries, including birth asphyxia and hypoxic ischemic encephalopathy”), accounting for only the rarest and most extreme outcomes. Because sentinel events are the “tip of the iceberg,” a decrease in the number of sentinel events, small to begin with, may not indicate an appreciable improvement in the overall safety or quality of patient care.

How might a broader definition of “sentinel event” to include nonfatal adverse outcomes and near misses have changed the study results and implications? The program’s impact might be more widely celebrated if results showed a decrease in often preventable adverse obstetric outcomes, such as unnecessary major surgical procedures (Weill Cornell has an extraordinarily high 40% cesarean section rate ), severe perineal tears, postpartum hemorrhage, neonatal ICU admissions, and postpartum posttraumatic stress disorder.

How did the criteria for identifying sentinel events change at the hospital over the study period? Despite systems to track errors, adverse events, and near misses, internal and external underreporting remains a significant problem. It is curious that 2008, the first year of the CMS “Never Event” policy limiting the ability of hospitals to bill for serious reportable events, was the same year that Weill Cornell’s sentinel event count dropped to zero.

What organizational system-level changes have occurred to prevent the recurrence of adverse outcomes? Most adverse events are the end result of a chain of chronic, repeatable, low-consequence deficiencies. Proactive, statistically rigorous approaches like failure modes and effects analysis and root cause analysis, would help determine the relevance, reliability, sustainability, and cost-effectiveness of the safety interventions.

Are there data on the number of and costs to defend nonlawsuit, noncatastrophic claims, and the frequency of error reporting and disclosure? A comprehensive analysis of total liability claims and costs associated with an open disclosure-with-offer program at the University of Michigan demonstrated compensation payment reduction without compromising ethics, transparency, or safety culture.