Public Health
Can Reminder Postcards Help People Get Vaccinated?
What we learned from sending 200,000 Louisianans a postcard telling them to get vaccinated.
This post is co-authored with Lula Chen.
If you are an elderly person who lived in Louisiana during the 2017 flu season, you may have received this postcard in the mail:
This reminder to get up to date on vaccinations was sent by your friendly state public health department. And with their help, we measured the impact of the card. Embedded in the public awareness program was a social science experiment, in which we assessed the effect of the card on vaccination rates.
Using anonymized data from Louisiana Immunization Information System (Louisiana’s vaccine registry), we found that the postcard had a small but detectable effect on vaccine uptake. The cards caused at least 560 additional vaccinations — mostly flu vaccines, but also some shingles vaccines — over two months. Almost all of the extra vaccinations happened in the month of October, which was the first of four consecutive months in which the cards were distributed.
We learned four things from this initiative, with a mix of lessons for policy-makers and policy researchers alike:
The reminder worked, a little. We are encouraged by the positive impact of the card, especially because this postcard was not new — it had been sent out the year before — and we were studying adult vaccinations which tend to be low. But, we also note the relatively small impact. At a minimum of 560 additional vaccinations for over 200,000 postcards sent, it is debatable whether a similar undertaking would be worth it if funds are limited. Postcards are cheaper than many public health programs, but they are not free, so it is worth asking what the pay-off was. Unfortunately, we could not run a cost-benefit analysis in this instance (due to data limitations), but where cost-benefit analyses are feasible, they would be a good idea for officials who are thinking about starting similar initiatives.
The whole may be greater than the sum of its parts. As mentioned above, the card worked best when it was received in October. It worked less in November, and not at all in December. We did not anticipate this, but if we had to guess why it happened, we think the efficacy of the card was boosted by all the flu-season, vaccine-promoting activity that occurs in October and November, making vaccinations particularly salient in those months. In the context of this activity, our card further increased (already seasonally high) vaccination rates, while the exact same card failed to have a measurable impact just a month or two later.
Randomize creatively. In this case, all eligible citizens needed to get the postcard, which meant we could not create a traditional control group (who would receive not card at all) for purposes of comparison. But we could still run a randomized evaluation, by randomly assigning when people received the card. Sometimes, rigorous evaluation is possible within the requirements of an existing public policy program.
Explore the context before you nudge. This experiment had many features of “nudges” — small and frequently cheap interventions intended to overcome behavioral biases such as forgetfulness. We know that nudges can work, but there are probably better and worse scenarios for making a nudge successful. In this case, practical barriers (such as lack of access to transportation to the doctor’s office) may have been the bigger obstacles. If you can’t make it to the doctor’s office, whether or not you are forgetful about your vaccines may not matter. Before launching a nudge, consider whether a behavioral barrier is a key barrier — this will help you nudge in places where this approach is most effective.
Overall, the study reinforced that reminders are an important and effective tool in public health. Their effectiveness varies with context, and probably interacts with other active public health measures. We hope these lessons become useful as we look forward to a major upcoming vaccine uptake campaign — this time for COVID-19.
