May 28, 2016

Incident Reporting: Is it really helping?

Incident Reporting (IR) in healthcare refers to any of a number of patient safety event reporting systems that rely on those who become aware of events to provide detailed information describing any accident or deviation from policies or orders involving a patient, employee, visitor, or student on the premises of a healthcare facility. It includes any event occurring in the process of patient care that is inconsistent with the desired patient outcomes or routine operations of a health care facility. The vast majority of hospitals today use electronic reporting systems, although some paper-based systems still exist.  In the opinion of many however, these systems which were implemented with the best of intention are now spinning out of control.

The idea of identifying, reporting and tracking of adverse events in order to improve the safety of flawed systems is essentially a good one. The Aviation Safety Reporting System is in large part responsible for the remarkable safety record in that industry, and healthcare systems are similar in design. However, improvement in healthcare safety has been slow to materialize. Although a recent report from the Department of Health and Human Services suggested that hospital-acquired conditions declined 17% from 2010 to 2013, preventable medical errors persist as the No. 3 killer in the U.S. – third only to heart disease and cancer – claiming the lives of some 400,000 people each year. Clearly, just putting reporting systems in place is not the whole answer.

Let’s examine how IR systems are being used (or, more accurately, not used). Many organizations use their reporting system to track their rate of improvement in patient safety. When the rate of serious safety events goes up, they congratulate themselves for having successfully created a “culture of reportability” and proclaim another victory for transparency. On the other hand, when the rate goes down, they declare that they have turned the corner and all of their patient safety efforts are finally paying off. This is flawed thinking, for several reasons.

First, a voluntary incident reporting system is not intended to be used as a true reflection of the rate of medical errors and even if it was, it’s not effective for that purpose. In one study, the adverse event detection methods commonly used to track patient safety in the United States today—voluntary reporting and the Agency for Healthcare Research and Quality’s Patient Safety Indicators—fared very poorly compared to other methods and missed 90%of the adverse events. The study demonstrated that the Institute for Healthcare Improvement’s Global Trigger Tool found at least ten times more confirmed, serious events than these other methods.

Second, the input is incomplete. Physicians rarely enter events in hospital IR systems and in fact most wouldn’t know how to do so even if they wanted to, which most don’t. Nurses and other front line care providers report the majority of safety events. The systems themselves have become complex in an attempt to provide enough data to allow accurate classification of the event, but as the complexity increases, so does the time required for data entry. System vendors often tout their product as requiring ‘no more than 5 minutes’ to enter an event, when in actual practice the time required is often 15 minutes or more. Suppose a floor nurse has just completed a 12-hour shift taking care of 6 (or more) complex patients. Since the nurse is encouraged/expected to enter not just serious safety events, but also near misses, precursor safety events, unsafe conditions and provider behavior issues, it is reasonable to assume that at least 3 or 4 such events are observed during the course of a single shift. In a world where management is pushing nurses to avoid overtime and meet sometimes unrealistic productivity targets, do we really think that the nurse is going to stay the extra hour or two required to enter these events into the system?

The data suggests that the answer is “No”. A 2012 report from the Department of Health and Human Services Office of the Inspector General found that Hospital staff did not report 86 percent of events to incident reporting systems, not just because of the time required for data entry but also because of staff misperceptions about what constitutes patient harm. Of the events experienced by Medicare beneficiaries discharged in October 2008, hospital incident reporting systems captured only an estimated 14%. In the absence of clear event reporting requirements, administrators classified 86%of unreported events as either events that staff did not perceive as reportable (62%of all events) or that staff commonly reported but did not report in this case (25%).

And finally, in order to attempt to make sense of those events that are reported, small armies of quality coordinators and risk managers are now spending most and in some cases all of their time trying to identify trends and produce reports. This takes them away from activities like safety rounds, safety training and process improvement that could actually reduce the amount of errors that they now spend their time categorizing.

I don’t have all the answers to the issues I’ve discussed, but I do know this: While Incident Reporting Systems have clearly not yet been perfected, we are much better off now than a couple of decades ago when medical errors were not even acknowledged, much less talked about openly. I also know that while we don’t really have a good way of knowing how many errors occur, even if the rate really is decreasing it’s still too high. One is too many and until we get to zero, we’re not finished.