Tuesday, August 16, 2011

Here's how to cure the world of disease 50% faster

Faster cures for more diseases. That sounds like a good thing, right? Clinical trials are often where a new drug or treatment meet success or failure as rigorous tests are done to ensure efficacy and measure side effects. Filling clinical trials with appropriate test subjects is a lengthy and expensive process but these trials are required before a meditation can be sold to consumers.

Hmm, sounds like a matching problem: how do we get the right trials in front of eligible subjects?

A few years back I was really pleased to discover that the US National Institute of Health maintains clinicaltrials.gov as a clearinghouse of federally funded and private trials. Today they have more than 112,000 trials listed in 175 countries. Wow, that's a lot of trials just waiting to be filled in order to judge whether or not it's a successful treatment protocol. Imagine if we could just fill these trials 50% faster - would that lead to more rapid success and failure discovery for pre-release drugs? And ultimately cures?

This is the question I asked myself while also wondering how Google Health could play a role in connecting eligible subjects with the appropriate test. Turned out it was really just some best practice work with the ClinicalTrials team to try and ensure Google was properly crawling and ranking the test information in our search index.

While that was a positive first step I think there's room to better connect people with this data. Where is our Hipmunk for clinical trials? A beautiful intuitive interface to this valuable data. They have a XML API so it's all waiting for you. You could be the developer who helps to speed the path to the next big cure - maybe you can help solve a disease that impacts, or has even taken the life of, someone you know and love.

Saturday, August 13, 2011

Smartphone aren't yet that smart (at least if you asked Charles Darwin)


Darwinism basically concludes that if smartphones were people, they’d be dead. Of course I love these brilliant little devices. Who wouldn’t be amazed to walk around holding their pocket a computer which just a few decades ago would have weighed several tons in order to match its processing power? But the reality is smart phones aren’t yet that smart because human intelligence isn’t just about your capabilities in a vacuum but rather your adaptability to environments. Right now the best my smartphone can figure out is time and place: where am I and when is it. but that’s not enough. Imagine if they could answer these other questions and make the answers dynamically available to developers.

“What am i doing?”
whether via accelerometer or just calculating change in position via GPS, my phone can start to make conclusions about what I am doing. For example, if it is moving quickly and also plugged in to charge, it should know that it is likely in a car. Why is this interesting? because I would love many apps to default to a ‘car mode’ which could for example, give me a simpler screen UI so I can hit the one or two most important buttons for your app while I am driving.

by just using speed, stability, local time and pattern recognition, my phone should be able to make educated guesses as to whether I am awake or asleep, still or walking or running, etc. tie this into my calendar and you really start to have some interesting possibilities for determining state.

“How healthy is the phone”
Can apps access battery status? I don’t think so and have never understood why my phone doesn’t make dynamic power management choices when the battery is low. For example you could decrease the refresh rate on apps running in the background.

“Have I been here before”
my phone certainly should know whether it is entering a geography that it has been before or not. Would this data be interesting to developers? I’m fairly certain all the mobile-local-social apps would like to know whether I was a first-time visitor to this neighborhood or a regular. It would also be very cool to have a map view of my city color-coded by the density of my previous visits - sort of a heat map for where I’ve been.

“Do I recognize the people around me”
this one is probably the most controversial but what if phones could sniff one another’s device IDs. From a biological perspective species definitely evolve with a heightened ability to detect strangers in the pack. What would the device driven equivalent be?

just some basic thoughts - what are other ideas for how this data could be used productively (obviously there are a ton of abuse vectors)? what other ways could devices better mimic human intelligence and awareness?