You may have heard about what happened in Flint, Michigan.
The city was poisoned by lead-laden water over the past two years when its water source was switched to the Flint River. At least six thousand children have been exposed to the heavy metal, and the impact of lead will continue to manifest — messing with brain development, changing behavior, weakening bodies and minds.
I heard of this crisis over the past few months when it made the headlines. After the numbers had hit the thousands. After cases were collated in neat Excel files and presented to city and state agencies. After Governor Rick Snyder finally acknowledged that the state had ignored complaints from the low-income city. And after three government officials had been charged for tampering with — and covering up — the toxic evidence.
But I didn’t hear about Flint in early 2014, when residents like LeAnne Walters first spoke out. She and her children were losing their hair, and her kids had strange bumps on their skin, as Julia Lurie reported in Mother Jones.
It took two years for the story to travel from those faucets to our ears, and land on the government’s lap with the weight of neglect. The problems in Flint needed to be exposed, tracked, and meticulously measured before serious action could be taken. And by that time, it was too late. Walters’ eyelashes had fallen out.
When it comes to health, both personal and public, we put data on a pedestal. The National Institutes of Health launched an initiative to make Big Data more accessible and user-friendly. And in 2014, United Nations Secretary-General Ban Ki-moon announced the Data Revolution for Sustainable Development initiative, a group of experts tasked with providing data and statistics for health, governance, and social interventions.
Scale is important. Numbers are what brought Flint’s crisis to a tipping point, and what forced action: 6,000 kids exposed to lead, 400 parts per billion of lead in the water. Same with the Zika virus; while the pictures of children born with microcephaly are heartbreaking, it’s the numbers and coordinates that have our attention.
But data collection has its limits. In Flint, we didn’t trust community voices enough to start collecting data when we heard the first whispers of a problem. In other places, political interests and human error may lead us to over- or under-represent a health risk.
And in the end, our over-reliance on hard data can make us sick.
II saw the limits of data firsthand in India, where my colleague Vivek Nemana and I spent over a year trying to find out who knew how many people died of malaria that year — or even that month, or week.
India has spent billions of dollars on its malaria program, supported by giants like the World Bank and the Global Fund. If you want to find out how well the program is working, you can ask the Indian government. In 2015 they reported that only 561 people had died, quite admirable in a population of 1.3 billion people and rising.
But ask that same question to independent researchers, and they will paint a wildly different picture. The World Health Organization estimated that the very same year, at least 15,000 people had died of malaria.
Our sources — a wide array of former government workers, community doctors and World Bank insiders — told us that India’s big data problem is multilayered.
Part of it is fear. Government health workers at all levels are scared to reveal that their part of the malaria program isn’t working as it should. Starting at the ground level, ordinary citizens don’t always know their own diagnoses, community health workers don’t document all the cases in their neighborhoods, and district officers file reports they know to be wrong. Meanwhile, they quietly watch the parasites take down families without any official channel through which they can communicate.
And that’s a vicious cycle: a weak program makes data collection even tougher. India is also trying to grow its influence in the global market, giving the government incentive to overstate its accomplishments.
A.C. Dhariwal, a malaria program officer with India’s health ministry, told me the government has commissioned a large-scale effort to collect better data in coming years. But the country’s starved health budget — weakened further when the World Bank stopped funding the malaria program — doesn’t reflect any renewed commitment to eradicating the disease, which it has promised to do by 2030.
This isn’t the only way data can hamper action. I spoke to epidemiologist Chelsea Solmo about this issue when she was studying malaria programs in Uganda. She found, after working with multiple international agencies, that the official number of people dying from malaria in Uganda is higher than reality.
She explained that government interests, and the desire for development, can influence data — the same, foolproof, hard numbers that we all crave to inform national strategy around the world. “Without reliable national statistics it is impossible to develop any interventions,” she said. “This really slows down the science.”
If the Indian government is hoping to cover up the failure of its malaria program, and Uganda is hoping for more foreign aid by inflating their numbers, can we use anything besides hard data to inform public health decisions?
To that end, some public health organizations actively use qualitative in addition to quantitative measures to guide their work, For instance, Jan Swasthya Sahyog, a community health organization in rural Chhattisgarh, India, works simultaneously on several levels. It engages community health workers to prevent disease, offers critical care at the central charitable hospital, and advises the government to better communicate what is happening on the ground to lawmakers.
The three times I’ve visited their center, I’ve seen what it means to be nimble in the bureaucratic world of public health and development. They’re able to “catch” illnesses around them because they’ve built trust with nearby communities through extensive programing and almost daily conversations. If a woman comes to them complaining of a spreading illness, or bad water, like Walters did in Flint, they don’t wait before acting.
And they have quick, cheap solutions — like tiny vials that test for bacterial water contamination — that can be sent out to a village in moments.
“At some level we [also] interact with the state to keep on telling them what they should do,” said Yogesh Jain, a community doctor with the organization. “They often listen but they always do a bit less.”
DDenying the importance of data would be a death sentence, and listening solely to community voices would mean treating disease without any diagnoses. Data has been enormously helpful in all of our recent public health crises and epidemics: from mapping out the trajectory of Ebola, to looking for patterns in cancer.
In the best scenarios, data and community voices can have a symbiotic relationship. A crowdsourced health project helped clarify that lithium didn’t slow the progression of ALS, a neurological disease that kills 100,000 people a year.
But I wonder how many years it will take, how many children will need to get sick, how many numbers are needed, for governments to be forced to act. Will parasites need to cross borders for the world to notice, like in the case of Zika?
In the case of Flint, it’s clear that the government and researchers should have collaborated as soon as they heard stirrings of families like LeAnn Walters’. The town has become a cautionary tale for the rest of the country, and cities like Newark, N.J. have been jolted into action over their own lead problems.
To prevent future crises, we need to treat community voices as a crucial part of public health interventions. And put resources behind connecting with these voices just as we do collecting data.
Because only someone who knows what their water looks and tastes like every day can tell you when it turns into poison.
