Ten years ago, the Institute of Medicine published To Err is Human [PDF], a groundbreaking report that pushed the issue of medical errors into the public spotlight.
That we all make mistakes was certainly nothing new: Operational failures occur across all industries. But the impact of errors in the context of the health-care industry drew instant attention. Preventable medical errors resulting in injury cost the industry somewhere between $9 billion and $15 billion a year, the report stated. Even more shockingly, by some measures the number of patient deaths attributed to operational failures annually in the United States equaled the crash of one fully loaded 747 airplane every one-and-a-half days.
"At first, these campaigns are great—they bolster the frequency with which solutions are shared by a significant margin....[But] at what point do people shut off?" —Michael Toffel
Since then, much research has focused on the underuse of incident-reporting systems. After all, the thinking went, a system used to collect and report incidents will only help an organization learn from its mistakes and lead to better safety results—to the extent that employees report information that can be used for process improvement.
For incident-reporting systems to fulfill their promise, employees must use the system to "speak up" when they encounter a problem. Managers receive additional value when reporters speak up constructively by offering suggestions that facilitate process improvement.
A Harvard research team recently set out to better understand what managers can do to encourage employees to speak up about problems, and to investigate how managers can encourage employees to offer solutions.
The team's working paper, "Speaking Up Constructively: Managerial Practices that Elicit Solutions from Front-Line Employees" [PDF], considers data on nearly 7,500 incidents from a single hospital to determine whether two types of managerial actions increase the frequency with which frontline workers speak up by reporting incidents and do so constructively by including solutions in their incident reports.
The paper, authored by Julia Adler-Milstein, an HBS doctoral candidate in the Health Policy Management program; Sara J. Singer, assistant professor at the Harvard School of Public Health and Harvard Medical School; and HBS professor Michael W. Toffel, also considers how organizational information campaigns and department managers' engagement in process improvement interact to influence the extent to which frontline workers speak up constructively, which could enable organizations to improve their operating processes and ultimately (one hopes) improve patient safety.
Manufacturing and service organizations also benefit from worker input. But people, it goes without saying, are harder to work on than cars and hotel rooms.
"Hospitals are enormously complex," Toffel observes. "Imagine a factory where every part has to be custom-built and can require any number of 100 or 200 services and subprocesses. On top of that, the most knowledgeable people about those subprocesses-the doctors-come and go from the factory and are not employed by it."
That complexity makes errors inevitable. And despite the growing emphasis in health care on patient safety, the researchers note that our understanding of how managers can increase the value of reporting systems remains incomplete.
To shed light on how to encourage staff to share constructive feedback when using reporting systems, Adler-Milstein, Singer, and Toffel examined the influence of managerial engagement on problem solving and of an organization-wide information campaign.
First, the phenomenon of patient-safety information campaigns: Such campaigns increase the frequency of frontline workers' speaking up following an incident by 5 percent, the researchers learned. However, when it comes to sharing a solution to the problem, the campaigns had a much larger effect, nearly tripling the frequency with which frontline workers suggested a solution to the problem.
In addition, units in which managers "practiced what they preached" by actively engaging in problem solving saw substantial increases in the frequency with which staff reported solutions when they filed incident reports.
"When managers had been more proactive in responding to incident reports, there was a greater likelihood that staff would share their suggestions and actions taken to resolve the underlying problem, which is very valuable information for managers because they are unlikely to be able to get this information elsewhere," says Adler-Milstein.
That result prompted further investigation: Do staff members in units with high managerial engagement respond differently to a patient-safety information campaign, compared to staff in units with low managerial engagement?
"Interestingly, we only saw a meaningful increase in offering solutions by employees in units with low managerial engagement," says Adler-Milstein.
"In a sense, the units with high managerial engagement were already kicked into high gear," explains Toffel, noting that this result also suggests possible future research on the duration of campaigns.
"At first, these campaigns are great—they bolster the frequency with which solutions are shared by a significant margin. But there's a reason why we don't have a campaign all the time, whether it's in a hospital or for the United Way: fatigue.
"A campaign's optimal duration for maximum benefits is still unclear. At what point do people shut off?"
The team's database offers additional information to consider that was not examined in the working paper, Toffel says. Other questions to answer include: What types of responses to incidents are most effective? When should behavioral corrections be implemented? When should technological corrections be made?
"I'm excited to look at this data longitudinally," says Singer. "Ideally, one would hope that an incident gets reported and that a solution is implemented so that the incident doesn't recur. We can look at whether this happens over time. Knowing this will make a significant contribution to improving patient safety, because a lot of hospitals rely on these reporting systems and promote their use, if only to fulfill accreditation requirements.
"The real question remains, are they serving the intended purpose? It could be that very little happens with these reports in terms of the long-term learning that you would hope to see."
Says Adler-Milstein, "We could also determine if the same type of incident is occurring in a given unit over time, even when it is being reported. That would then make it possible to focus on how particular units resolve their problems."
Identifying pockets of excellence would enable more qualitative research to determine what exactly a unit is doing to achieve its success—and to identify how those practices could be codified and adopted elsewhere.
"Health care started out with largely independent practitioners and a limited body of knowledge," says Adler-Milstein. "Given the changes that have occurred recently, technological and otherwise, health care hasn't caught up quickly enough with the new practice methods that accompany this very different, modern-day model. I hope we will get there eventually, but right now there is a lag."
Drilling down to discover when frontline employees speak up most constructively, and how to translate this into problem solving, should help bridge that gap.