All My Pharmacy Students Learn to Code

We all know that health care is changing rapidly. The mobile phone, augmented by health and wellness trackers, apps, and sensors is shifting care into the home. Pharmacists must be leading new practice model initiatives and, at a minimum, be familiar with connected technology. The chief clinical officer of Partners Healthcare, Dr. Gregg Meyer, likes to say, “Right now, there is a freshman at MIT tinkering in his dorm room who is going to put us all out of a job.”

With that in mind, a few months ago, I decided that I would take the concept “always be learning” to a new level by including a coding aspect into my hospital advanced pharmacy practice experiences institutional rotation. When it comes to technology, I am fascinated by machine learning and voice-powered digital devices, such as Amazon Alexa and Google Home. By one estimate, by 2020, 30% of interactions with technology will happen through voice.¹

The term the “democratization of medicine” was popularized by a prominent author and physician named Eric Topol. Major tech companies have “democratized technology” by creating many tools for any level of expertise. One does not need to be a computer programmer to create neat and interesting health care applications. By 2022, Forrester expects 66% of US households to have a smart-home device.²

I can foresee a time where care at home runs through a smart device. Thus, I chose creating an Alexa Skill as our goal.

My first student for this pilot was a talented clinician from MCPHS University named Mohannad Al-Nabolsi. I knew this first student for my novel rotation was going to be the most challenging. On his first day, I went through my normal orientation interview, but I asked probing questions about his openness to new ideas and familiarity with technology. It turns out that he was a former professional gamer and had dabbled in computer programming. Perfect! The first thing we did was Google how to program the Amazon Alexa. Tech companies have tons of informative videos on how to program, and sample code is available on repositories such as Github. That is all Moe needed to learn to program an Alexa Skill. He had his first one built in less than a week.

The next step was to find clinical partners for a pilot study and decide what problem we were trying to solve with our skill. There are a lot of problems to solve in health care, so it was not hard to find a partner. Enhanced Recover After Surgery (ERAS) is a care pathway designed to improve clinical outcomes and cost savings. One key aspect to this pathway is partnering with patients in their care journey. Most institutions provide patients with wordy pamphlets explaining all the various aspects of care. The patient is expected to comply with the pre-procedure recommendations. Maybe pharmacy students would remember all the details in such a pamphlet, but between my 2 kids and busy work life, I would never remember when to take my Neomycin before surgery. So, we put all the ERAS patient education material into an Alexa Skill.

Alexa is programmed using intents, utterances, and responses. For example, prior to colorectal surgery a patient must drink a carbohydrate drink to make sure that sugar levels are consistent during surgery. The intent for this skill was “drink.” The utterance is anything that a patient might say to get the response he or she is looking for. As an example, if a patient intends to find out when to take a carbohydrate drink, an utterance that they might use would be “Alexa, when should I have my carbohydrate drink?” In this case Alexa’s response is: “The drink should be started 3 hours before your surgery and finished no later than 2 hours before the surgery is scheduled to begin.” Needless to say, the surgeons were blown away by this skill. Right now, we are working with our compliance department on HIPAA-related issues, but we hope to start piloting our skill with patients very soon. Alexa is not HIPAA compliant so it is critical that personal health information does not run through an Alexa device. Alexa is always listening. Thus, privacy concerns have been raised when the device is used in the health care setting.

One aspect to internet of things devices, such as Alexa, is the insights gained from interactions with the devices. I hope to do interesting things, such as analyze if a patient asks Alexa a lot of questions regarding ERAS, is the outcome better or worse? If it turns out that patients who ask a lot of questions have worse outcomes, perhaps there could be an intervention early on for those who need additional services.

Since this first skill was developed, my students have created a number of other skills, including information about the hospital, education for congestive heart failure and commercial pharmacy formulary information to try to cut down on those pesky prior authorizations.

For other pharmacists, technicians or students looking to create their own digital tools, here are 2 pieces of advice: A demonstration is worth a thousand Powerpoints; and always start with the problem you are trying to solve.

My next student starts after the new year. I cannot wait. We plan to tackle the problem of expensive chemotherapy waste by building a machine-learning model. Wish us luck!

References

1. Levy HP. Gartner predicts a virtual world of exponential change. Gartner Inc. gartner.com/smarterwithgartner/gartner-predicts-a-virtual-world-of-exponential-change/. Published October 18, 2016. Accessed December 27, 2017.

2. Priya S. Forrester data smart home devices forecast, 2016 to 2021 (US). Forrester Research Inc. forrester.com/report/Forrester+Data+Smart+Home+Devices+Forecast+2016+To+2021+US/-/E-RES134441. Published May 11, 2016. Accessed December 27, 2017.