What sounds more like an investigation into a fantasy realm than the subterranean urban affair that it is, MIT’s Underworlds project hopes to unearth lessons about health from our very own sewers. But rather than send people down into the dark underworld, the team of researchers has developed robots to do the dirty work for us (one job that is not likely to have people complaining about robots taking our jobs). The robots are playfully named Mario (first generation) and Luigi (second generation), after the world’s most famous plumbers.
“We all flush valuable health data down the toilet. Sewers represent a unique opportunity where health data from everybody in a community is pooled together,” explains Eric Alm, Associate Professor of Biological Engineering at MIT and Co-Principal Investigator of Underworlds. “If we learn how to sample, test and interpret data from sewers, we are creating new powerful tools to monitor the health of populations.”
By analyzing urban sewage, the team believes that they can create a better understanding of epidemiology (the science of diseases) and equip public health officials to anticipate outbreaks and possibly even help them understand the causes of chronic diseases.
Headed by Director of the MIT Senseable City Lab and the Principal investigator of Underworlds Carlo Ratti and Eric Alm, the Underworlds Project was launched in 2015. The team is currently using the prototype Luigi to sample data in Cambridge, Boston and Kuwait.
“Current sampling technology allows us to collect multiple discrete samples, as composites or grabs. Our sampling technology is pushing this further by onboarding some of the preprocessing onto the sampler. Rather than taking liters of raw sewage to the wet lab, we’re able to begin the filtration process of fecal and urinary matter in situ,” explains Ratti. “Eventually we hope to onboard pathogen-specific detection to build out a real-time, or close to real-time, biochemical detection platform,” he adds.
The team aims to help scientists predict where and how outbreaks happen, as well as to map the health of urbanites. If the technology allows scientists to detect outbreaks, it would put policy-makers in a position to anticipate them at early stages. More importantly, the team foresees the potential of helping policy-makers and scientists understand the prevalence of certain types of diseases in different neighborhoods.
“We can also measure biomarkers for noncommunicable diseases such as obesity and diabetes, leading to more effective public health policies: for example, how large of a sugar tax is needed to impact health,” says Ratti. Ratti notes that the information gathered could have different applications for different stakeholders, with public health authorities looking at the prevalence of infectious disease outbreaks or overall community well being versus public works officials, who could look at illegal waste from industries, for example.
The duo believe that the project can be run in different cities around the world to provide data and insights about urban health in very different contexts. But even with their limited sample of three cities, they have already been able to make observations about the samples.
“First, our approach to sampling sewage near residences, rather than the traditional approach of sampling at water treatment plants, is beginning to pay off. We find that sewage looks much closer to stool and urine samples collected directly from individuals. We aimed to ensure that toilet water was no more than 10 minutes journey from its origin and our sample point. As a result, we can detect chemicals (e.g. pharmaceuticals) as modified by the human liver, giving us 100% certainty that those chemicals were ingested by humans or humans were otherwise exposed to them,” explains Alm.
“Because these modifications are often labile, this would not be possible using samples collected at a treatment plant. Finally, we’re using the DNA sequencing to estimate the number of different people contributing to each sample, so that we can develop aggregate statistics about public health,” he adds.
More critically, he explains that the team hopes to correlate findings to the demographics of different neighborhoods. “Where are we seeing more antibiotics? Why? Is that where we have a lot of young or elderly populations? This layered, real-time data can inform policy makers, health practitioners, designers, and researchers alike. As such, the implications of this platform extend beyond just disease monitoring to the development of a new type of human population census,” he adds.
But what sounds eerily like the Big Brother of sewage is anything but, with Alm reassuring us that the reporting is not likely to have any negative implications on the populations being monitored – although the samples could help identify trends in different neighborhoods. “For example, it’s not very interesting to learn that someone is taking heroin or smoking tobacco, particularly when you don’t have any identifiable information on that person. But it is helpful from a public health standpoint if you can identify the percentage of individuals using these drugs,” he says.
“A noncommunicable disease such as obesity is another good example. A recent study showed that looking at microbes from sewage correlated with obesity rates across different cities. However, measuring obesity rates through traditional methods remains the simplest way…except when we want to measure the effect of public health interventions. If the city decides to combat obesity by raising a tax on sugary drinks, it will be difficult, and it may take a long time to see if the policy changes are having an effect by directly measuring obesity rates. But measuring the change in obesity biomarkers in sewage, whether these are microbial as in the recently published study, or human biomarkers, could occur almost in real-time. That kind of feedback might allow policymakers to understand what policies were effective, and for example how large of a tax is needed to impact health,” he adds.
With 10 prototypes already developed, the team is continuously working to fine tune the robots, which are able to filter the fecal and urine samples that they gather on site. They are also building a communication infrastructure to allow sampling instruments to upload basic mechanical and environmental data to a central server remotely. In November of last year, Underworlds was awarded a grant from Kuwait Foundation, which will sustain the project for the coming three years.