Online platform crowdsources predictions about disease outbreak

An online platform launched earlier this year crowdsources intelligence to predict the outcomes of disease outbreak.

The Johns Hopkins Center for Health Security implemented its publicly available disease prediction policy dashboard 10 months ago, and officials hope forecasting efforts might lead to better public health policy.

“This is an attempt to try to supplement an existing model and existing surveillance with another data point to help policymakers make decisions,” said Tara Kirk Sell, one of the project’s co-principal investigators. “It’s not going to replace surveillance or replace modeling, but it is another piece of information that can help point toward resource allocation.”

The team at Johns Hopkins asks the crowd questions relating to outbreaks spanning the globe. Questions have ranged from, ‘Will the United States lose its WHO measles elimination status?’ to ‘How many total cases of tularemia for 2019 will be reported in Sweden?’ More than 1,000 people are registered on the platform and as of last month about 500 of those have made a forecast.

“We’ve been discovering also through this process that there are some terrible gaps in surveillance around the world. In an interconnected world, we think that we should have information on everything all the time,” Sell said.

What’s surprising is that expertise is not a predictor of being a good forecaster, Sell says, and that the results show that predictions are more accurate when they are made at a group level.

“We’re testing usability. So far some of our forecasting has been really good and other times it could use some work,” Sell said. “I think we’re trying to figure out under what scenarios these forecasts would provide useful information to policymakers.”

Awareness for the project is cultivated through Johns Hopkins’ networks, the school’s health security newsletter, and word of mouth. Collective Intelligence for Disease Prediction is currently in its second six-month phase, which will run until the end of the year. Organizers say it’s not too late to join the forecasting community.

The project is supported by funding from the Open Philanthropy Project.