Closed claims analyses can have limitations when recommending improved practices for clinicians. Strive for actionable information rather than global data.

• It may be necessary to improve data points in closed claims before analysis.

• Seek specific results that may be counterintuitive.

• Avoid telling clinicians to just be more vigilant.

Eighty percent of physicians will face a medical malpractice lawsuit at some point, with those in high-risk specialties like obstetrics facing even worse prospects. Risk managers welcome any strategy that helps identify the physicians and situations most likely to result in liability.

Closed claims are an invaluable resource for identifying risks, highlighting the most common factors associated with adverse outcomes and liability. Like other organizations, TeamHealth, based in Knoxville, TN, with more than 20,000 clinicians nationwide, has used closed claims data analysis to help identify those risks and potential risk reduction opportunities.

But there is a limit to the information that can be gleaned from closed claims, says Kevin Klauer, DO, EJD, FACEP, chief medical officer for hospital-based services and chief risk officer with TeamHealth.

“Those databases are not specifically designed to identify risk trends,” Klauer says. “You can get high-level numbers like the claims in a certain area geographically, or with a certain diagnosis. But when it comes down to moving toward predictive modeling and getting actionable information for physicians, you need more.”

For instance, Klauer says it is not enough to remind physicians that stroke, sepsis, and heart attacks are often misdiagnosed or mismanaged simply because the closed claims data indicate those are frequent topics in civil litigation. Physicians are already trying to manage those issues properly, so simply telling them that they often result in lawsuits is not much help, he says.

“You have to give them actionable data, information that is useful clinically, rather than just saying ‘Here’s a problem, and here’s some reminders about best practices,’” Klauer says.

“You’re also implying that the clinician did something wrong, and we know in claims management that is not always the case,” he adds. “These are multifactorial claims, and frequently, there is nothing wrong with the care provided. But you still have a very unhappy patient who is unsatisfied with the outcome.”

Claims Database Improved

To collect more actionable data, TeamHealth added additional data fields retrospectively to the claims analysis. The goal was to derive data that were more useful for predictive modeling and risk management, Klauer says.

For example, the data analysis indicated that sepsis was the chief readmission complaint of ED patients. TeamHealth had already provided that information to its clinicians, along with reminders about best practices for sepsis prevention, and Klauer says that is necessary. But they wanted to go further.

“You can’t stop there. We also include in our database now, and we retro-populated it, the chief complaint when someone came in to the emergency department and was ultimately diagnosed with sepsis,” Klauer explains. “And what was the physician diagnosis given when they left that ultimately resulted in a diagnosis and a lawsuit for sepsis? That’s actionable information, and it may return some information that is counterintuitive.”

With sepsis, for instance, Klauer notes that it is reasonable to assume sepsis will be diagnosed in patients with undifferentiated fever. But interestingly, that was only the second most common complaint in patients who were discharged and returned later with sepsis. The number one complaint was abdominal pain, which can be especially difficult to pin down with a diagnosis.

UTIs Discovered as Sepsis Link

Once abdominal pain patients return with sepsis, the analysis indicated that the most common diagnosis was urinary tract infection.

“That is incredibly actionable information. You can tell your physicians that if they have examined a patient with abdominal pain and sent them home with a diagnosis of urinary tract infection, there is a subset of those patients who will return with sepsis,” Klauer says.

“Frequently, the urinalysis they did during the examination for abdominal pain will have had some nonspecific findings like a few white blood cells or maybe some nitrates. Those are not always diagnostic, but you can anchor on them and that might raise your suspicion of urinary tract infection.”

In the vast majority of patients with a urinary tract infection diagnosis after returning with sepsis, there were no urinary symptoms of that infection when they originally sought care for abdominal pain, Klauer says.

“So instead of telling clinicians that ‘sepsis is a concern and it’s a better diagnosis, please keep that in mind and do a better job,’ it is more useful to tell them that when you may miss sepsis is in this situation: You have a patient with undifferentiated abdominal pain who got a thorough evaluation, you thought they might have a urinary tract infection, and you sent them home,” Klauer says. “Give them actionable information rather than just global information. Mindfulness is not a risk management strategy.”


• Kevin Klauer, DO, EJD, FACEP, Chief Medical Officer for Hospital-based Services, Chief Risk Officer, TeamHealth, Knoxville, TN. Phone: (800) 342-2898.