An exercise in Epidemiology: putting the world's access to cell phones to use.
I was thinking today about vaccines for various illnesses you can get while traveling and remember a fear I had about going to Nepal last year—Cholera. Out of sight and out of mind of most Americans, the disease affects roughly 3-5 million worldwide annually. That's all pretty sad when you consider there is a vaccination available for $.10—$4 a piece with 50—60% effectiveness1.
Is it perfect? No, but widespread vaccinations coupled with herd immunity have been historically proven to reduce disease vastly. My Wikipedia trails on slow Sundays always seem to take interesting directions, but today it led me to think about how we’re not using technology already in the hands of most people worldwide to better track outbreaks on a citizen level.
When I was in Nepal last summer, a country commonly courted in the top 15% poorest in the world, I was trekking through a village, and we were staying in a mud church with a dingy tin roof when the pastor comes into the building… holding a smartphone. I’m sure my reference to his religious role seems irrelevant but you have to remember Nepal is a predominately Hindu and Buddhist nation. Christians generally aren’t too wealthy, so seeing a smartphone in his hands piqued my interest. We have to also remember the Swedish telecom who covered Nepal in cell phone towers somehow to thank for this technological revolution, but however it happened, people living in some of the poorest conditions in the world have phones. If this pastor has a smartphone, and some Googling about world population with cell phones is accurate, roughly 60% of people on the planet are connected to our phone network. With this information, I imagine that most village/VDC (the only term I know for governance of collections of villages) leaders are connected.
Two ideas got tied in my head today, the prevalence of disease and the ubiquity of cell phones. But we don’t have a nice, real time location for that data outside the CDC/WHO who are both wonderful organizations but very much tied politically. They represent western influence with western leadership, which is often not the most effective way to work with majority-world nations. It would be really interesting if we could crowdsource disease outbreak reports from anybody with a phone, especially when we start talking about smaller sicknesses and places that aid agencies/governments have a difficult time accessing to start with.
So how could we build an international crowdsourcing operation? We’re not talking about making an app work in USA, Canada, Europe, and Hong Kong and calling it a day. We’re talking about asking some of the poorest people on the planet to remember who we are, find localized ways to establish trust in the concept, and finding the cheapest way per-country to accept submitted data of varying quality and entering it into a global database.
How can we be remembered?
I think this whole idea starts with the marketing. An interesting example is the Telia company, owner of a surprising amount of developing nation cellular networks.
I didn’t even know that their model was a thing until I was researching Kazakh phone carriers for a friend, but then realized the genius behind their system - they enter these markets and act entirely like a local player. All the Ncell advertising is entirely localized. Nepali language, Nepali models, Nepali social media presence. I’m sure the other markets are exactly the same. Ncell felt very sophisticated as a Nepali brand, but it never felt un-Nepali or overtly western.
The model is clear: your endpoint for data entry needs to be advertised like a local initiative ran by a local company. Give it a name that’s locally significant. If I were to do this in Nepal, I’d probably say “Be a Bahadur for health” (warrior). And I’d put Shah Rukh Khan’s face on it because it’s Nepal and I can’t imagine you need to ask for permission to use Indian celebrity endorsements. Don’t make it sound like texting or entering outbreaks is going to bring medicines or a helicopter, but make it look like somebody cares to receive the information about a persons community and that the information could be used to better serve their villages needs someday.
And for advertising? Find 50 truck owners and ask them to place a nicely designed, weatherproof poster or decal on their trucks for $100. It’s a relatively bribe-free way to spread the word about something in third-world countries, more permanent spots could require the infamous ladder of bribes that it takes to make anything happen in corrupt countries. And make sure the person asking to post the ad is a local. That should help on the pricing points.
How can we establish trust in our brand?
While we’re talking about marketing this idea of submitting outbreaks, we need to figure out how to establish trust in entirely different cultures. English words, white/western models, foreign feeling concepts won’t work here. We need to frame the texting endpoint or webpage to submit as something that appeals to the community we’re trying to reach. If they’re a very nationalist culture, maybe put a few flags on the poster. If they’re family-oriented, a picture of whatever a happy or respectable family would be there. If they’re populist, make sure the family in the photo is middle class, and so forth.
How can we cheaply accept data for each individual country?
We need to consider the technological savvy of people in cultures. Nowadays, a responsive form that works on most smart devices worldwide is a trivial thing to make. But for places like Nepal, where interacting with data, especially digital, may be more foreign, a number for people to text outbreak details to would be more effective. It’d be pretty easy to just point to an actual phone number, like an iPhone with an operator in Nepal somewhere, and have them take screenshots of every incoming text, and send a pre-generated personal sounding response. Those screenshots can then be entered into the global outbreak database by others, providing a quality filter for the data (data submitted by form should be cleaned up or rejected by the same operators)
In the early stages of growth, you could likely run an operation like this for under $10,000/year for two full time local staff members for a country, paying them well above market rates for the category of work. That may seem like a lot, but we’re talking about running 7-8 countries data input for roughly the same amount one professional health worker would cost. Phone bills wouldn’t have too much of a shell-shock either.
Bonus: how we can leverage the community of globetrotters, backpackers, and journalists who love wandering into obscure places?
There’s an incredible community of young (western) people roaming around the world, trying to get some experience away from the comforts they were raised in. There’s backpackers, missionaries, aid workers, and journalists who could be the first to notice a pattern of disease occurring in an area, and having access to their data would be an incredible win. We could even accept texts like these from satellite phones.
When we talk about having a crowdsourced data from a more technically savvy audience, we could even have them run through a required guide on how to report outbreaks (terms for symptoms, qualities to observe). We could gamify the process somehow or incentive them to be our “experts on the ground”. Globetrotters don’t have much of a problem finding internet nearly anywhere on the planet.
So now that we have all this data, what are we going to do with it?
For much of the above, I am speculating from my small amount of research into foreign aid, marketing, and web startups. Data processing is more my strength and this why this whole idea gets me more excited to start.
Data Processing / Entry
The data processing would be pretty easy once we have standardized outbreak information like amount of people affected, age groups, location, and reporting time. We need to convert locations to lat/lon, check it for duplicates, maybe check for any bias in the data, then we can just dump it all to Redshift or Google’s big data engine.
Real time visualization and pattern detection
I love maps and seeing a map of what is ailing people around the world would be an interesting way to track sickness. We could certainly compare it to where most disease elimination efforts are. But what gets interesting is when we can start tracking how sicknesses spread with big data. It would take months or years of continuous data input, but if we could start connecting rivers and roads with the spread of disease, and then pass this to governments of these countries who may be too resource trapped to find patterns, but could pursue leads with lots of proof, maybe we can help increase their own ability to control their future when it comes to diseases. With the wealth of new satellite imagery coming online, maybe we can link sicknesses with an image of a new factory up a river or downwind from smokestacks. And we could certainly provide an API to help others build tools to help us find these connections.
Imagine 10 years of crowdsourced data with weather data merged in, with links to recent news stories about the region or country, and notes about wars or famines2;. We could potentially find links to disease spread across continents we never could have seen on a micro-scale. I have code laying around to get predicted winds at nearly every altitude for nearly every point on the planet. Does the jet stream carry diseases?
As a disclaimer, I’m not in expert in disease, globalized initiatives, or cultural relevancy, but I hope that by risking being wrong about some things can inspire somebody who is right about things to connect these dots. I realize some of my generalizations here are dangerous and most of my cultural research is in Asian countries, so this will reflect some of the issues I would expect to happen in rolling out a crowdsourced initiative like this there. I also have no idea what goes into studying the spread of disease—I’m just a web developer who thinks that technology could go a lot father linking the world than it currently does. I have spent a lot of time studying southeast Asian cultures and see where a lot of American attempts to “change the world” fall entirely flat. I think if we really spent the time figuring out what conditions need to change for a crowdsourced data platform to work in individual communities, I believe we can leverage the locals input on disease outbreak to one day better prevent diseases from impacting their communities with improved prevention strategies, increased awareness and preparation time of incoming diseases, and new solutions we might not have been otherwise unable to invent without big data.