Meta-study results suggest most implicit bias training for health care providers has flawed methodology

A small team of psychologists and public health specialists from the University of Virginia, Virginia Commonwealth University, Old Dominion University and the University of Wisconsin-Madison has found via analysis of data from several studies that most implicit bias training efforts suffer from flawed methodology and translational gaps that compromise their integrity.

In their paper, published in the journal Science Advances, the group notes that little scientific evidence shows that such training programs lead to reductions in bias.

Implicit bias is defined as a type of learned stereotype that is automatic to a given individual, generally associative, unintentional and usually deeply ingrained.

Prior research has shown that an implicit bias can influence behavior, for example, giving less attention to Black pregnant women at health care facilities due to unconscious, stereotypical beliefs that Black women tend to complain more when experiencing "normal" problems. Such behavior biases have been shown to lead to a higher percentage of adverse outcomes for Black women during pregnancy and labor than for white women.

Over the past several decades, the health care industry has studied implicit bias and found it to be problematic and has therefore sought to address bias in health care settings through what has come to be known as "implicit bias training."

In this new study, the research team found evidence suggesting that many of these training programs use techniques that have no scientific basis, a finding that suggests many health care institutions or facilities are merely giving lip service to the problem rather than trying to solve it.

The researchers analyzed 77 studies conducted from January 2003 to September 2022 involving implicit bias training for health care workers. As part of that effort, they looked at how the bias training programs were designed and carried out, whether there were any gaps in knowledge translation, and if so, if they tended to decrease reliability and/or reduce the validity of the training.

The results showed that there was little scientific evidence among the studies to support such efforts. They also found little evidence suggesting that such training efforts have any meaningful impact on the people who were being trained—they found no measurable impact on behavior changes by people who had attended implicit training programs.

© 2024 Science X Network