Researchers have found a simple and effective way to reduce medical errors

We trust our doctors with our lives, but the sad and scary reality is that doctors can get things wrong. Each year, approximately 100,000 Americans die from medical errors, and recent studies show that 10 to 15 percent of clinical decisions regarding patient diagnosis and treatment are incorrect.

A team of researchers led by Damon Centola, professor and director of the Network Dynamics Group at the University of Pennsylvania’s Annenberg School for Communication, has found a simple and effective way to reduce errors in patient diagnosis and treatment -; Use structured networks to connect clinics with other clinics.

In a study published today in the journal Proceedings of the National Academy of Sciences (PNAS), the researchers shared results from a multi-year study involving nearly 3,000 doctors across the United States.

When presented with a case study and asked to make diagnostic and treatment recommendations for a patient, clinicians who anonymously reported their friends’ diagnostic decisions were on average twice as likely as physicians who made decisions alone.

Simply put, doctors make fewer mistakes when they have a support network.

The biggest risk with these information sharing networks is that while some doctors may improve, there may be an average outcome that leads better doctors to make worse decisions. But, that’s not what happens. Instead of regressing to the mean, there is constant improvement: the worst clinics get better, but the best don’t.

Damon Centola, Elihu Katz Professor of Communication, Sociology and Engineering

“We are increasingly realizing that clinical decision-making must be viewed as a team effort, including multiple clinicians and the patient. This study shows that having other clinicians present for consultation during decision-making can improve clinical care,” said study co-author Elaine Kung of the San Francisco General Hospital and Trauma Center in California.

Beyond the art of clinical collection

Over the course of several months, the researchers tested clinicians’ treatment and diagnostic decisions with an app they built and distributed through Apple’s App Store specifically for this purpose.

After signing up for the trial and downloading the app, doctors were asked to assess their clinical condition -; Patient cases based on real life -; Over three rounds. At the beginning of each round, the clinicians read the case study, then were given two minutes to answer two questions.

The first question asked the doctors to give the patient a diagnostic risk (for example, what is the probability that a patient with chest pain will have a heart attack in the next 30 days?) on a scale of 1 to 100.

Each clinic was randomly assigned to one of two groups: either a control group whose members answered all the questions individually, or an experimental group in which participants were connected to other anonymous physicians in a social network who could see their responses.

In rounds two and three, participants in the control group had the same experience as round one, answering questions independently. However, participants in the network condition could see the average risk estimates made by their peers in the social network in the previous round.

Each participant, whether they live in the social network or not, is given the opportunity to improve their answers from one round to the next.

Centola’s team used the same experimental design to study seven different clinical cases, each of which showed high levels of diagnostic or treatment error.

The researchers found that the overall accuracy of clinicians’ decisions doubled in networks compared to control groups. In addition, among the initially worst-performing clinics, the network increased control by 15% in the class of clinics that gave the correct advice.

“Using doctors’ networks, we can improve their performance,” Centola says. “Doctors talk to each other, and we’ve known that for a long time. The real breakthrough here is that we can set up information-sharing networks between doctors to dramatically increase clinical knowledge.”

Leveling the playing field

In medicine, face-to-face consulting networks are typically hierarchical, with senior physicians and junior doctors below. “Young doctors come into the medical community with different cultural and personal perspectives and are influenced by these top-down networks,” Centola says. “That’s how this persistent bias permeates the medical community.”

The researchers sought to recruit clinicians of varying ages, specialties, expertise, and geographic locations to the trial.

The researchers found that peer-to-peer networks, which are anonymous, eliminate barriers of rank and seniority, limiting many aspects of learning in medical networks. “Equal online networks increase the diversity of voices that influence clinical decisions. As a result, we have improved decision-making across the board for a variety of different professionals,” Centola said.

In the doctor’s office

“We don’t have to reinvent the wheel to apply these findings,” Centola says. “Some hospitals, especially those with low resources, rely on e-consultation technologies where a medical professional sends a message to an outside specialist for advice. It usually takes 24 to 72 hours to get a response. Why not send this request to a network of specialists instead of one person?”

Centola said each trial took less than 20 minutes. Moreover, it says that the network need not be huge. Actually, 40 members are suitable.

“Forty people in a network gives you a big jump in the collective intelligence of clinicians,” Centola says. “Beyond that, the incremental return — going from 40 to 4,000 — is very small.”

The researchers are currently working to implement network technology in doctor’s offices. The University of Pennsylvania Hospital has already funded a pilot implementation of this program, which is set to begin later this year.