Many public health interventions are designed to address already existing public health epidemics, for example by encouraging physical activity to address the obesity epidemic or creating a vaccine after a large outbreak. Public health challenges can be difficult to predict, which explains why interventions are often designed after a public health challenge is considered an epidemic (a passive approach). However, more public health departments are using what is called predictive analytics to identify future public health challenges and design interventions that prevent the issue from becoming an epidemic.
Predictive analytics can be used to guide scarce public health resources and determine where there would be the greatest impact per dollar spent. This week the Senate discussed funding for two public health epidemics – prevention for opioid addiction and combating the spread of the Zika virus. Both issues have received some attention as issues that merit the use of predictive analytics to inform resource allocation. About half of $1.1 billion funding to combat opioid abuse proposed by President Obama would go to expanding treatment facilities and the other half to prevention efforts, including tracking overdoses more closely. Making data publicly available and using predictive analytics has not been mentioned in the proposal, but many have written about the impact big data can have on preventing addiction and related morbidity. Similarly with the Zika virus, health experts have called for open data source sharing in order to accurately predict the spread of the Zika virus and use resources accordingly. Currently, the amount of publicly available data that is needed to predict the spread of the virus is limited.
Chicago has served as a prime example of the use of big data. The county health department uses predictive analytics to help prevent the spread of food-borne illnesses in the 16,000 food establishments it oversees. The city’s Innovation and Technology Department created an algorithm that predicts which of the establishments are most likely to have critical violations that could harm consumers’ health using various data sources. The algorithm looks at a variety of variables, including nearby garbage sanitation complaints, neighborhood population density and past inspection results. They also monitor Chicagoans’ tweets about food poisoning symptoms and send the user information about how to report their condition so the health department can investigate. The algorithm found violations 7.5 days earlier, on average, than the previous inspection protocol. The city made their algorithm code public so that other cities could duplicate their work if they have the technical knowledge and resources available. Chicago also uses predictive analytics to prevent childhood lead poisoning.
If funding is approved to prevent opioid abuse and the spread of the Zika virus, the use of predictive analytics could potentially help to make the most of every allocated dollar.