News & Views: Education Makes It Worse?

Two recent papers by researchers at Dartmouth University may have disheartened you if you heard about them:

• Nyhan B, Reifler J, Richey S, Freed GL. Effective messages in vaccine promotion: a randomized trial. Pediatrics. 2014 Apr;133(4):e835-42.
• Nyhan B, J Reifler. Does correcting myths about the flu vaccine work? An experimental evaluation of the effects of corrective information. Vaccine. 2014 Dec 2. pii: S0264-410X(14)01542-4.

In both of these papers, which were covered by the media, the authors concluded that vaccine messages meant to educate might actually misinform and even decrease intent to vaccinate. Let’s take a closer look at these studies, so that we can better understand what they mean for those of us who educate about vaccines. Should we just stop trying? By no means!

Effective Messages in Vaccine Promotion: A Randomized Trial

Study format: The study was conducted in two waves of online panels in which respondents during the first wave were re-contacted to complete the second wave.

Participants: 1,759 parents of children ages 17 or younger

Phase one: Measurement of health and vaccine attitudes allowing respondents to be categorized by how favorably they viewed vaccines.

Phase two: Respondents were randomly assigned to one of five experimental groups:

1. Autism correction group – this group received information adapted from messages promoted by the Centers for Disease Control and Prevention (CDC), which corrected the myth that vaccines cause autism.
2. Disease risks group – this group received information, adapted from CDC materials (VIS sheets), which explained the risks of diseases prevented by the MMR vaccine.
3. Disease images group – this group was shown pictures of children with diseases prevented by the MMR vaccine.
4. Narrative danger group – this group was provided a descriptive narrative about an infant who almost died from measles.
5. Control group – received information about the costs and benefits of bird feeding.

Outcomes: Three main outcomes were measured; however, each was only determined by respondents’ answers to a single question (per outcome):

• Misperceptions about vaccines as the cause of autism
• Perceived side effects of vaccines
• Intent to vaccinate

Study findings:

• The group of people who received the “autism correction” information was significantly less likely to believe vaccines were a cause of autism after reviewing the information; however, they were also less likely to express an intention to vaccinate future children.
• The group of people who received the narrative describing the dangers of disease (“narrative danger group”) was more likely to believe that the MMR vaccine caused serious side effects.
• When respondents were sub-categorized based on how favorably they viewed vaccines, differences were not seen between the different experimental interventions for either “cause of autism” or “perceived side effects” outcomes. However, for the “intent to vaccinate” outcome, those subjects who had the least favorable attitudes towards vaccines AND received the “autism correction” information decreased their intent to vaccinate. No other vaccine attitude groups or information intervention types showed a similar decline.

Does correcting myths about the flu vaccine work? An experimental evaluation of the effects of corrective information

Study format: Data for this study were generated as part of an online survey primarily focused on politics and government. Two waves of surveys were conducted.

Participants: U.S. adults drawn from the YouGov/PolimetrixPollingPointPanel and the E-Rewards and Western Wats panels. Phase one had 1,000 respondents, and phase two represented phase one recipients who accepted an invitation to participate in phase two. Phase two had 822 respondents.

Phase one: Prior to introduction of the interventions, all respondents were asked about their concerns related to serious vaccine side effects and this information was used to segregate respondents into high- or low-concern groups for purposes of comparison. The high-risk group represented 24 percent of total respondents.

After the concern group information was obtained, respondents were randomly assigned to one of the following experimental groups:

1. Correction group – received CDC information addressing the myth that influenza vaccine causes influenza
2. Danger group – received CDC information about the dangers of influenza disease
3. Control group – received no information about influenza or the vaccine

Phase two: Designed to measure longevity of impact of interventions; however, the group suffered non-random attrition that prevented drawing valid conclusions related to this measure.

Outcomes: Three main outcomes were measured in both waves of the experiment; however, each was only determined by respondents’ answers to a single question (per outcome):

• Misperceptions about the influenza vaccine
• Beliefs about influenza vaccine safety
• Intent to get vaccinated

Study findings:

• Although the authors cite that randomization was successful, 74 percent of respondents had only “some college” education or a “high school degree or less.”
• Although one of the most publicized data points from the study was that 4 in 10 Americans believed the flu vaccine could give you the flu, the data also showed that more than 8 in 10 Americans believe the flu vaccine is safe.
• People who got information correcting the myth that the flu vaccine causes influenza (“correction intervention”) were less likely to believe the vaccine caused flu after reviewing the information. Likewise, these people were less likely to maintain incorrect beliefs about vaccine safety although statistical significance varied depending upon whether respondents had low- or high-levels of concern about vaccine risks. In contrast, people who got information about the dangers of influenza (“danger intervention” group) did not change their misperceptions or beliefs.
• Intent to get the flu vaccine did not change in the study population as a result of either intervention (“correction intervention” or “danger intervention” groups). However, for the one-quarter of respondents with high-risk concerns, the “correction intervention” showed significant decreases in intent to get vaccinated. Of note, based on data provided in Table 1, this group (high-risk group that received the “correction intervention”) represented 69 of the total 1,000 respondents. The high-risk “control group” represented 81 respondents, and the high-risk “danger intervention” group was comprised of 92 respondents (Note: these three groups were reported to comprise 226 total respondents, so group numbers were likely to be slightly lower than calculated using table 1. Differences are likely due to rounding.). In contrast, low-risk groups represented 240 “control” respondents, 246 “danger intervention” respondents and 272 “correction intervention” respondents, and no negative intervention effect was found.

So what does this mean for educating about vaccines?

While it is important to consider these findings, we need to realize the framework in which they were found and evaluate them in the context of the bigger picture.

• Realize the framework – Both of these studies use single questions to measure outcomes and while the questions of choice were appropriate for the measure, they do not establish the complexity of the emotional framework from which vaccine decisions are made. Additionally, intent to vaccinate was also measured by a single question and is somewhat arbitrary in that it is an answer to an online question and takes into account only that moment in time rather than in a setting where these types of information are typically shared, such as in a physician’s office, and accompanied by conversation with a healthcare professional. Indeed, the influenza survey was part of a politics and government survey.
• Look at the bigger picture – Anyone who communicates with parents or patients about vaccine decisions is likely to already realize and understand that different people will respond to different approaches and to different presentations, so that these studies found different responses to different materials is not necessarily surprising or contrary to what is known. And while these findings may be helpful to get us thinking about different forms of communication, they represent a scenario that differs from typical educational efforts. Specifically, whereas these studies were completed by giving all respondents a pre-determined piece of information via an online method, in practice, information is accompanied by conversation and likely relates to the specific questions or concerns indicated during the conversation. These in-person interactions also allow for an assessment of potential, or openness to receiving information.

From our experience, we can tell pretty early in a conversation which people will not be convinced, and it is likely that if you have these conversations in your office daily, you can too.