This post was written more than four years ago. The world changes fast, and the information, conclusions, or attributions may or may not still be accurate. Check the sources and links, and email me if you have any questions.

IFTTT is one of my absolute favorite web services that makes life better and easier. It works like this: you setup channels that contain data — for example, your Gmail or foursquare account, the weather, the time, or an RSS feed. Based on the data in those channels, you setup condition-based recipes that contain triggers. For example:

  • Every time I’m mentioned on Twitter [Twitter channel], add the tweet to my Evernote account [Evernote channel]
  • If I get an e-mail from my boss [Gmail channel], send me a text message [SMS channel]
  • At sunset [Date/Time channel], turn my back deck light on [Philips Hue lightbulb channel]
  • Every time I post an Instagram photo [Instagram channel], save it to my Dropbox account [Dropbox channel]
  • Turn on my coffee maker [Belkin WeMo wi-fi outlet channel] after I have my morning weigh-in [Withings wi-fi scale channel]

You can quickly see that IFTTT automates web services and home automation to make life easier. But hold on a second — that last one is sort of an interesting match between behavioral design and health decision-making. Could health data feed into something like IFTTT?

For patients focused on fitness or patients battling multiple chronic conditions, there’s a long list of goals, activities, and interventions for them to follow. For a diabetic patient, for example, they need to check their blood glucose, increase physical activity, maintain or lose weight, maintain hemoglobin A1c, etc. There’s a lot of interventions and points of data that can easily fall through the cracks.

What if a patient’s health record — their vital records, labwork, diagnoses, encounters, procedures, and medications — was something resembling an IFTTT channel? With healthcare data becoming more easily portable via BlueButton+ Direct messages with a ‘push and pull’ methodology, it’s certainly a possibility. What if other medical, fitness, and mobile health devices were also feeding into IFTTT?

Suddenly, regardless of whether medical data comes from your regular clinic, a random quick-check clinic you visit in a grocery store while traveling, or a medical device in your home, these ‘IFTTT for Health’ or ‘IFTTT for Fitness’ use cases become possible:

  • If my triglycerides exceed 149 mg/dL [EHR data], send me and my doctor an e-mail
  • If I gain more than five pounds [EHR data or a FitBit or Withings scale], send me a text message
  • If I gain more than five pounds [EHR/scale data], set a new daily walking goal to 15,000 steps [FitBit]
  • If I eat at more than two fast food restaurants in a week [foursquare check-in data], send me a text message every work night reminding me to make a meal to bring to work the next day for a week
  • If I don’t walk 30,000 steps by end of day Wednesday [FitBit], schedule a Google Calendar event for Thursday at 2pm to go for a walk around the lake

While I think the simplicity is fantastic for the Quantified Self types with a passion for self-tracking and improvement, I think this can go much further — much more structured, and something that could actually be implemented in a primary care setting involving a physician, caregiver, or family member. Medical interventions are typically associated with a care plan, which can contain specific measures. A diabetic patient might receive a care plan with measures of acceptable ranges, limits, and goals for measured health data like HbA1c, blood pressure, or medication adherence.

Clearly this is straying from the simplicity of an IFTTT for Health, but if a provider can ‘prescribe’ a data-driven, live care plan based on actual measured data from consolidated electronic medical records, medical and fitness devices, it’s a patient population health tool that alerts providers to failed interventions as they happen. Could interventions intelligently change as interventions are successfully accomplished or as patients fail to adhere to a plan?

Taking it a step further, if the app knows your diagnoses and procedures from your electronic medical record data, it could suggest bundles of recipes for you based on what other patients recommend — or, since actual data is involved, what actually works for other patients.

I think this concept is interesting for a number of reasons, most importantly because BlueButton+ and the Direct Project means that medical data will (hopefully) be freely flowing between provider databases and consumer apps like never before — because not being able to manage, own, and interact with my medical data that belongs to me is not only unacceptable — it’s a requirement under Meaningful Use Stage 2:

“Beginning 2014, provide patients the ability to view online, download, and transmit their health information within 4 business days of the information being available…”
— 42 CFR § 495.6

Because this should not be the limit of how I interact and use my medical data:

epic-mychart-lab-values