A new study has found, yet again, that it’s hard to change the minds of parents who are skeptical of vaccines.

Vaccines are the best way to protect children from infectious diseases such as the measles, but vaccination rates have fallen in some areas of the country over the past few years as parents have unfounded doubts about inefficacy and safety.

Alleviating these doubts—which are often based on misinformation—is a public health priority, as experts work to prevent outbreaks of dangerous and preventable diseases.

Previous research found that simply providing information are on the health benefits of vaccines does not change parents’ minds. In fact, a 2014 study found giving additional information to parents who had not vaccinated their children backfired. Though the parents who received the information were less likely to believe that the MMR vaccine (which treats measles, mumps, and rubella) caused autism, they were also less likely to say they would vaccinate their child than they had been at the start of the study.

For the most recent study, researchers tried to capitalize on the idea that doctors are the most trusted sources of information on this issue. They focused on increasing providers’ confidence in their ability to persuade parents.

Staff in 30 clinics in Washington state were given a 45-minute training session that encouraged providers to use respectful, open dialogue with parents while clearly recommending vaccinations. The researchers continued to be available to help providers for six months after the training. They then followed 347 mothers of newborn babies. Some of these mothers received care at these clinics; some did at 26 similar clinics where staff had not received the training.

The researchers explained that ideally, they would like to measure whether vaccination rates went up at clinics with the intervention. Because only a small percentage of parents in the state refuse to vaccinate, testing this measure would have required a much larger study with many more participants.

Instead, the researchers measured both the parents’ hesitancy to vaccinate and the providers’ confidence in their ability to discuss the topic.

Neither of these factors changed as a result of the intervention.

Vaccine hesitancy declined somewhat among parents at both the intervention clinics and the control clinics, but there were was no statistically significant difference between the groups. Moreover, other events in Washington during the six months of the study—a whooping cough outbreak and a state law designed to make it harder for parents to opt out of vaccine requirements—could have explained the decline.

“Obviously we were hopeful that it would improve vaccine hesitancy, so we would have preferred to see a different effect, but it really raised more questions about what other projects we could do moving forward,” Dr. Nora Henrikson, the study’s lead author, told NPR.

Experts agree that the study was worthwhile. Brendan Nyhan, a professor of government at Dartmouth who was the lead author on last year’s study, told NPR: “We’re headed in the right direction. We’re starting to ask better questions, and part of real science is that sometimes our experiments don’t work out the way we expect.”