Sunday, August 21, 2016

Pardon Me, Your Interface Is Showing

In a great post, "Doctor as Designer" Joyce Lee laments the "sad state of product and design in healthcare," and asks "when will device and drug companies create user-centered innovations that actually improve the lives of patients instead of their bottom line?"

I heartily agree with Dr. Lee's point, and think the question can be extended to the rest of the health care system.

Dr. Lee uses two examples to compare health care to consumer goods.  Heinz took a product design -- the glass ketchup bottle -- that had been around for over a hundred years, and greatly improved the user experience by changing to a squeezable "upside down" bottle.  This not only kept the ketchup from concentrating at the bottom but also avoided the need to hold the bottle at a special angle or to tap at a particular spot just to get the ketchup out.

She contrasts this with the Epi-pen.  It is not only hard to use correctly, but its manufacturer has used the recalls of competitors' medications to jack up its price by several hundred percent (from $100 to over $600).  Dr, Lee notes that some consumers are simply buying their own epinephrine and needles to create their DIY version, for about $5, "which means that we are paying $600 for a hunk of badly designed plastic!"    

Bad design for more money; it sure sounds like health care, doesn't it?

When I think about health care's lack of user-centered design, though, I think less about Epi-pens or medical devices and about more common patient interactions, like in doctor's offices or hospitals.

Kaiser Health News recently published two articles on the experience of elderly patients in the hospital.  The first noted that, ironically, elderly patients often are admitted sick but leave disabled.  It is important -- but uncommon -- for hospitals to focus on how to get the patient back home living as independently as possible.  Bed rest, catheters, IVs, interrupted sleep, and unappetizing food all can work against that goal.

The second KHN article stressed the need to keep hospitalized elderly patients moving.  A 2009 study found that such patients spend 83% of their stay in bed, being out of bed a median of only 43 minutes per day.  One nurse warns patients, "the bed is not your friend."

Hospitals are just protecting themselves from lawsuits.  A geriatrician explained that families won't sue if their parent gets weaker while in the hospital, but may sue if he/she falls, so preventing falls trumps preparing patients to go home independently.

As another geriatrician noted: "The older you are, the worse the hospital is for you."  Still another physician likened current approaches to a "smart bomb."  "We blow away the disease," he said, "but we leave a lot of collateral damage."

If that isn't a good description of our "health care" system, I don't know what is.

Design matters.  KHN cites examples of hospitals that have created special units that pay more attention to helping patients be more mobile -- through changes in room design, assuring that walkers are widely available, and focused care processes.  It can be done.

Certainly hospitals are much different than a generation ago, with semi-private rooms on their way out (who ever thought that was a good idea in the first place?) and amenities like WiFi more common.  Hospitals are said to be borrowing from the hotel industry to improve patient experience, but this may aimed more at marketing and revenue-enhancement opportunities than to improving patient care.

Still, I suspect that the next time a patient confuses a hospital for a hotel will be the first.

The health care system is recognizing that it needs to engage people differently.  Such engagement is seen as essential to getting them more involved in their health, especially in managing chronic conditions.  It is potentially big business, with the patient engagement market expected to grow from $7.4b in 2015 to $39b by 2024, according to Grand View Research.  

MobiHealthNews sees a big role for consumer health tech companies in this, particularly on the B2B side.  They cite numerous examples of alliances, acquisitions, and partnerships along those lines.  When it comes to improving patient experience, it asks, "What better place to turn than devices and apps that have already proven themselves engaging and delightful in the direct-to-consumer world?"

The problem may be that we're still not quite sure who the "customer" is.  According to Xerox, nearly 50% of consumers say they take "complete responsibility" for their heath, but only 6% of health professionals think that is true.  Nearly 40% of providers and payors think consumers don't even know how to take charge of their health.

It's hard to design for a health care system when we don't even agree who is "in charge" of our health.

If you think too much about the interface, it's bad design, creating friction.  The health care system is full of this kind of friction.  Think of selecting a health plan, understanding health coverage, finding a provider, getting an appointment, waiting to receive care, or understanding a diagnosis and treatment options.  And don't get me started on EHRs.

Martin Legowiecki, writing in TechCrunch, thinks UI should be "invisible" and that AI is the way to get to that.  The world, in his view, should be as easy as walking into your favorite bar and having the bartender have your favorite drink ready as soon as you sit down.  As he says, "that's a lot of interaction, without any 'interaction.'"

Or, as he puts it more pithily, "the ultimate UI is no UI."

In an Internet of Things world, we could use normal language to talk to our environment, with the omnipresent AI able to understand and apply conceptual awareness to accommodate our needs.  Picture a hospital bed that not only warns you when you've been immobile too long but also "helps" you get up, or a doctor's office that pulls together all the necessary information on you before you even arrive.

Design starts with making something functional, and good design tries to make easier to use, or at least more attractive.  Really good design doesn't make us think about how clever the designers are but, rather, allows us to forget that they did anything at all.

Health care could use some really good design.

1 comment:

  1. Hi Kim,
    I agree entirely with your proposal that the clinician community needs to be in control of the design of the EHR they have to use, however this requires a new level of technology we call Immediate Adatpability (

    However automatic language processing is vastly more difficult than people imagine so your statement "In an Internet of Things world, we could use normal language to talk to our environment” is somewhat of an overreach. Take for example the second last sentence of your blog with the clause "and good design tries to make easier to use, or at least more attractive”. It is grammatically malformed as the word “it” is missing between “make” and “easier”. Us humans can accommodate these types of mistakes quite readily. But how do you think an algorithm could detect that missing word, and even if it could put it in place how would it detect which word or phrase it is referring to, that is the coreferent, and subsequently what there had to be “least” of.
    As a computational linguist specialising in the analysis of clinical documents our clients ( will only be happy with a technology that is minimally 90% accurate which can be quite difficult to achieve let alone the desirable 100% which is virtually never attainable.
    Regrettably, the Nirvana of high accuracy language understanding is still over the horizon and probably far over it.

    Jon Patrick