We Are Not Guinea Pigs: The Effects of Negative News on Vaccine Compliance

"A growing body of research has highlighted the importance of public trust for the effectiveness of vaccination campaigns aimed at curbing the spread of disease during epidemics."

This paper studies the effects of negative news about vaccination on vaccine hesitancy and noncompliance. It does so by examining the aftermath of human trials of an experimental antibiotic in the Muslim state of Kano in Northern Nigeria, which had experienced a meningitis epidemic in 1996. Eleven of the 200 children died and multiple children were disabled following the trials. An exposé about the incident published in 2000 in The Washington Post sparked a series of protests in 2001. They were led by Muslim religious leaders who alleged that Pfizer and "the West" were trying to kill Muslims and called for boycotts of vaccinations among Muslims.

The subsequent reduction in routine child vaccination among Muslim mothers set back global polio eradication efforts by over a decade; negative effects lasted for up to 12 years after the initial disclosure. Specifically, following this incident, there was heightened distrust among Muslims toward vaccination campaigns led by "Western" nonprofits like the Global Polio Eradication Initiative (GPEI). In 2003, a boycott led by Muslim religious leaders of the GPEI-led polio mass vaccination campaigns in five Muslim northern states in Nigeria was driven by rumours that the vaccination effort was part of an effort by Westerners to spread HIV among Muslims and cause infertility in Muslim girls. In interviews published elsewhere explaining support for the 2003 boycott, respondents explicitly cited the Pfizer drug trials, with one respondent stating, "We cannot trust the white man or our federal government because many years ago they were in partnership when they brought medicine to poison our people". Although an out-of-court settlement with Pzifer was reached in 2009, "The Pfizer trials remain a point of tension among Muslims in Nigeria..."

The paper assesses the effects of the 2000 Washington Post disclosure on child vaccination outcomes using a difference-in-differences strategy comparing differences in child vaccination outcomes across Muslim versus non-Muslim mothers (first difference) and for children born before versus after the Washington Post revelation of the Pfizer drug trials in 2000 (second difference). The researchers use available data from the birth recode of the Demographic and Health Surveys for four rounds of surveys between 1990 and 2013. The dataset documents individual mothers' reported vaccination of children born between 1985 and 2013 and records the mother's religion.

The study assumes that the year of birth and mother's religious identity jointly determine a child's exposure to the shock induced by news of the Pfizer drug trials. Since the Pfizer drug trials occurred in 1996, it is plausible that news about the trials may have already begun diffusing among Muslims through religious networks, as has been documented in other settings. Given information diffusion among these networks, the researchers consider children born between 1996 and the 2000 Washington Post disclosure year as partially exposed to the news about the Pfizer trials. Children born after the 2000 disclosure - after the information about the trials was fully revealed - are considered fully exposed to the news. Children born before the trials in 1996 are considered non-exposed.

The empirical strategy requires a primary identifying assumption: that in the absence of news about the Pfizer trials, vaccination rates across birth cohorts would have been similar for children born to Muslim and non-Muslim mothers (see, e.g., Table A2 in the paper's appendix, which show that the Muslim-non-Muslim child vaccination gap was generally not systematically different prior to the 2000 disclosure).

The study finds that the 2000 Washington Post disclosure led to significant reductions in routine vaccination - namely, tuberculosis (BCG), diphtheria, pertussis, and tetanus (DPT), polio, and measles vaccinations - of children born to Muslim mothers after the disclosure. Muslim mothers reduced BCG, DPT, polio, and measles vaccination rates by 9.1 percentage points (pp), 8.9 pp, 5.1 pp, and 8.6 pp, respectively, equivalent to an 11-27% reduction in child vaccination of children born to Muslim mothers relative to the pre-disclosure sample mean. The disclosure did not affect other health-seeking behaviour of mothers; the reduction effect is specific to child vaccination.

The researchers show that the effects are largely driven by educated Muslim mothers and Muslim mothers residing in minority Muslim neighbourhoods with relatively stronger ties to religious networks (farther away from family networks and with relatively lower trust in neighbours). "The evidence of higher religiosity and trust in religious leaders among Muslim respondents supports the hypothesis that the reduction in vaccination outcomes for children born to Muslim mothers after the disclosure was linked to information spread among Muslim mothers with stronger religious networks."

In conclusion, the "results show that local trust networks can be key drivers of vaccine hesitancy, especially among minority populations within regions with potentially stronger own-group cleavage; hence, policymakers aiming to increase vaccination in the aftermath of epidemics must work to build and leverage trust within these local networks to enhance vaccine uptake. These findings highlight the importance of both careful, ethical, and transparent practices in vaccination efforts, and institutional and local community network trust in vaccine compliance. The negative externalities from reduced vaccination have global consequences for the resurgence of epidemics of infectious diseases that can persist for years after the initial event and disclosure."