J Brainard1, P R Hunter2, I R Hall3. 1. Norwich Medical School, Norwich, United Kingdom. Electronic address: j.brainard@uea.ac.uk. 2. Norwich Medical School, Norwich, United Kingdom. 3. Public Health England, United Kingdom; University of Manchester School of Mathematics, Manchester, United Kingdom.
Abstract
BACKGROUND: Concern about health misinformation is longstanding, especially on the Internet. METHODS: Using agent-based models, we considered the effects of such misinformation on a norovirus outbreak, and some methods for countering the possible impacts of "good" and "bad" health advice. The work explicitly models spread of physical disease and information (both online and offline) as two separate but interacting processes. The models have multiple stochastic elements; repeat model runs were made to identify parameter values that most consistently produced the desired target baseline scenario. Next, parameters were found that most consistently led to a scenario when outbreak severity was clearly made worse by circulating poor quality disease prevention advice. Strategies to counter "fake" health news were tested. RESULTS: Reducing bad advice to 30% of total information or making at least 30% of people fully resistant to believing in and sharing bad health advice were effective thresholds to counteract the negative impacts of bad advice during a norovirus outbreak. CONCLUSION: How feasible it is to achieve these targets within communication networks (online and offline) should be explored.
BACKGROUND: Concern about health misinformation is longstanding, especially on the Internet. METHODS: Using agent-based models, we considered the effects of such misinformation on a norovirus outbreak, and some methods for countering the possible impacts of "good" and "bad" health advice. The work explicitly models spread of physical disease and information (both online and offline) as two separate but interacting processes. The models have multiple stochastic elements; repeat model runs were made to identify parameter values that most consistently produced the desired target baseline scenario. Next, parameters were found that most consistently led to a scenario when outbreak severity was clearly made worse by circulating poor quality disease prevention advice. Strategies to counter "fake" health news were tested. RESULTS: Reducing bad advice to 30% of total information or making at least 30% of people fully resistant to believing in and sharing bad health advice were effective thresholds to counteract the negative impacts of bad advice during a norovirus outbreak. CONCLUSION: How feasible it is to achieve these targets within communication networks (online and offline) should be explored.