I was recently reading an article in the Wall Street Journal titled, “Everyone hates customer service; this is why.” It’s an amazing piece detailing how customer service is getting worse while businesses leverage technology to understand how poorly they can treat their customers before they lose sales. Yeah, I just wrote that. Some businesses have learned that they can mistreat their paying customers without deleterious effect, and have taken advantage of these findings to do just that (hello cable and phone companies.) Other more “respectable” groups are using technology to do more with less, and I want to focus on them.
How does this technology work? “Some companies now equip call centers with software that analyzes a caller’s tone of voice and pace of speech to determine how upset the person is. Angrier callers get routed to agents skilled at de-escalating conflict, who are in turn warned in advance,” the article stated. Another business “transcribes call-center conversations in real time, sensing when a call is taking a turn for the worse. That can help prompt a manager to listen in, view the transcript, and step in if necessary without having to ask customers to repeat themselves.” My favorite is a group that “crunches data gleaned from consumers’ demographic profiles, credit scores, and past interactions with a company to determine which customer-service agent is the best fit. An algorithm then matches the caller to the agent who has had the most success with that type of caller.”
This is all cool stuff, but how does it relate to the practice of medicine? Should we be recording patients as they call our clinics asking for advice about a rash or wondering about that refill that should’ve been at the drug store but isn’t? As I contemplated these questions, the article answered the questions for me. (You can probably figure out that the answer is no, but please read on anyway.) It turns out that all the easy stuff that folks used to call customer service lines to get help with in the past are automated nowadays. Who would’ve thunk it? “Businesses have digitized basic customer-service functions, such as checking bank statements and making online returns. That means customers who connect with customer-service agents are generally those with the most complex problems. That leads to more difficult interactions.”
That’s when the connection clicked for me. When I call customer service today, it’s inherently a complicated and difficult situation, otherwise I would have already dealt with it online or with the company’s mobile app. Sounds sort of like hospitalizing patients! If a patient ends up in the hospital today, he/she is inherently sicker than the average patient a few decades ago. Why? Because of course, we now have novel treatments that allow us to deal with serious illness in the outpatient setting. Oh, and perhaps there are financial considerations as well. We’re putting sicker patients in the hospital and are expected to discharge them in less time. What could go wrong?
To make matters worse, at least from a customer service perspective, most hospitalized patients are now taken care of by physicians they hardly know. Professional “hospitalists” have replaced the primary care doctor that patients see day in and day out, often for years. There are definite benefits to this system. Hospitalists specialize in taking care of very sick patients in the inpatient setting. They know the ins and outs of the hospital, and they’re often in-house around-the-clock. But a hospitalist can’t know their patient like someone who has taken care of him/her for a long time in an ambulatory setting. And a randomly-assigned hospitalist may not have the personality or “way of doing things” that matches a given patient (reference the company mentioned above that matches a customer with an agent who has a complementary personality.).
Can we use healthcare IT to help? While we can’t check our patients’ credit scores and give all of our clinical employees personality tests (ok, maybe we can do the latter), we can ask obvious questions about how patients like to deal with their caregivers and display those data at the appropriate time and place in the EHR. We might also be able to automate and infer some customer-pleasing facts from our interactions with our patients. For example, could we predict how much information the patient might want the care team to share with him/her based on how he/she uses the patient portal? Is there an active account? Are lab results and messages checked immediately? Heck, I dislike overused terms like “big data,” but can we predict how best to interact with our patients by correlating the way they pay their bills with how they request medication refills? I don’t know the answers to these questions, but it might be fun to find out!
Craig Joseph, MD, is the Chief Medical Officer for Avaap, an EHR and ERP consulting firm. He is a pediatrician and physician informaticist with experience in both clinical practice and the use of the EHR to improve patient care and physician efficiency. Dr. Joseph also currently serves as the Interim CMIO at El Camino Hospital in Mountain View, California. This piece was originally posted on Avaap’s blog page.