The Workflow Prescription: Patients Need Zapier, Workflow, and IFTTT-like Control Over Self-Care Workflow Automation At Home

This post was prompted by an Healthcare Leadership tweetchat about patient self-care at home and aging-in-place technologies. By the way, if this blog post interests you, I’m a guest on ’s “blab” immediately after the HCLDR tweetchat (click the “subscribe” link in the following tweet).

I often get into debates with proponents of patients owning their data. It usually goes like this:

  • Patients must own their data!
  • Me: Patients won’t own their data in a practical sense until they control the workflows creating (and consuming) that data.

Some times these debates really are this short. Other times, we both repeat ourselves a couple times. This particular HCLDR tweetchat is an opportunity for me to explain what I mean.

Let’s use the self-care, aging-in-place example. Consider this detailed plan of care and patient instructions (cited in the HCLDR tweetchat blog post)

There’s no patient data! It’s essentially an informal, through detailed, description of a set of personal care-related workflows. How might we use workflow technology to take this instructions of a piece of paper and make them come alive, so to speak?

These are essentially a set of reminders. If you don’t go up to your roof, you need a reminder. If the reminder doesn’t work, then you need a more insistent reminders. If that doesn’t work, then the reminder needs to escalate, to someone, perhaps a relative, friend, or perhaps the original primary care physician.

Intelligent reminders, ones that know if you’ve not done what they are reminding you to do and keep escalating, perhaps back to someone you really want to please (or at least keep off your back) are classic workflow technology (particularly when more than one person is involved). There is a model of process or workflow, in this case transition from the workflow state of someone not being on the roof to the workflow state of being on the roof (twice).

How would a workflow engine “know” whether someone is on the roof or not? That’s where the wearable and Internet of Things come in. They will increasingly feed into a workflow engine consulting the workflow diagram (picture workflow diagram arrow from a node labeled “not on roof” pointing to a node labeled “on roof”). Wearables and IoT will be both eyes and ears of workflow tech surrounding us, and arms, legs, and mouth as well. In other words, we’ll get reminders in lots of different ways, from traditional smartphone notifications to vibrating wearables to a little voice in our ear. Reminders will even follow us around, appearing in lots of other places, such as the LCD screen in the elevator (when we are alone: no, really, the tech already exists!).

The above scenario, of interacting wearables and Internet of Things objects, providing data to workflow engines that in turn push alerts and reminders back out to wearables and the Internet of Things was essentially my February keynote at the Healthcare Systems Process Improvement Conference in Orlando. The only different is that the same tech is applied in the home instead of the hospital.

Unlike the quantified self and health IT geeks I hang out with online, patients don’t want their data. They want to know what to do, when, and why, and in the nicest but also most effective way possible. That’s where cloud workflow automation services such as Zapier, Workflow, and IFTTT come in.

For overviews of these services (plus Flow XO), skim the following articles:

Are these services easy to use? Not necessarily. Not yet. But here is the important thing. You don’t have to be a programmer. And they will get easier and easier to use.

Zapier, Workflow, and IFTTT aren’t healthcare workflow automation services in the cloud (yet). But these are exactly what we need. We need workflows between the systems the prescriber of the workflow interacts with and the systems the patient interacts with: smartphones, wearables, and ambient computing in the walls and hallways (and roofs).

Lets return to these detailed plan of care instructions. The rest of the bullets are also patient-centered workflows. I could tell the same story I told regarding visiting the roof. But I don’t want to be tedious. But let me drill down on a couple more of the workflows.

Number 5, “continue current medical regimen” is the most clinical, and is really referring to an entire other workflow, a clinical workflow. Which is OK, since one step in one workflow can refer to accomplishing an entire workflow with its own series of steps and logic and so on. Bullet number 7 isn’t literally intended. But I can easily imagine smartphone apps or smartwatches or ambient computing displays that follow you around from screen to screen, from display surface to display surface, to audible alerts and utterances: Well, we’re at UGA… did you say hello? 🙂

Yes, when a patient goes from one doctor to another, they need a copy of their medical records. But what patients really need are lots of little workflow helpers and helpmates along the way. Where they can be relatives and friends, great. But it there are no relatives and friends, then we need to provide friendly workflow technology as a fall back. And, even if a patient is lucky enough to have the love and support of family and friends, those incredible people need friendly workflow tech to make it easier for *them* to help: “Hey, I’ll be happy to participate in those workflows you physician prescribed for you. If you don’t go up the room twice a day, I want to know! So, after you ignore that especially loud and annoying reminder to go up the roof, but before your personal workflow system rats you out to your physician, send *ME* a notification, and *I’LL” come over to get you up to the roof.”

Workflow technology started out in factories and document management. It has since proliferated across industries, including services industries. It underlies some of the smarter services we interact with on our smartphones, from Google Now to Siri. Increasingly, workflows are learned and optimized using machine learning. We need to get this kinds of intelligent patient-centered workflows all the way from our traditional health IT systems into the home, to support, nag, cajole, call for help, and, if necessary, go tattle.

Let’s assume we have the magical workflow technology, and it’s everywhere, and it can sense stuff about us, and then make stuff happen to influence our behavior…. Where do the representations of workflow come from? Do I expect patients to author their own workflows? No. At least not most. However, if a patient, or someone used to be a patient, or someone who is taking care of patient, is sufficiently motivated to create or edit a workflow (and then share that workflow with the world, just as Zapier, Workflow, and IFTTT users currently do), then the health and healthcare IT infrastructure around us could meet us at least halfway.

Right now? Health IT doesn’t even meet us one-half of one-percent of the way. (Oh, OK. Maybe five or ten percent of the way. I do increasingly bump into health IT startups, and larger concerns, heading toward the deeper end of the healthcare workflow technology pool!)


Leave a Reply