Archive for May, 2014

More on the invisible power of computing: how it affects the way we interact with one another

I’ve written before about the major influence of Moore’s Law on modern life in general, and health care in particular.  Moore’s Law states that the number of transistors on a chip doubles every 18 months, and therefore the cost of computing power halves in that same interval.  Result: simple information has become cheaper than zero, and ever more sophisticated information is becoming cheaper daily.  I am old enough to remember when four-function calculators came out, replacing the old adding machines that worked mechanically, and spit out tapes that resembled grocery store receipts.  The calculators were originally priced in the hundreds of dollars. 


But even sophisticated information has gotten cheap.  Think of all the functionality on your smartphone.  Although we each generally use a very small percentage of its potential, just think about what it does do for us.  On it is a GPS that used to be a couple of hundred dollars by itself.  I access TripAdvisor in strange cities and come up with great hotels and terrific restaurants, and directions to them as well as online registration and reservations.  And now that I’m three models behind, my model, brand new, costs about $200. 


While all of this is wonderful, there are some unintended consequences.  First, what we used to think was hard is now easy, especially if you have some money.  Ads cater to what we’re thinking, because of the powerful ability of vendors to infer what’s attractive to us by looking at our online data.  Our GPS devices reroute us around traffic jams, and some new car models anticipate accidents by tracking the data being generated by the car two cars ahead of us. 


I wouldn’t want to go back to the time when I had to go to AAA to pick up a TripTik, or customized route map, before a vacation trip.  (Young people: I’m not making this up.)  But the interface between humans and the computer-based environment isn’t static, for either party.  As I said earlier, Moore’s Law gives us more and more computing power more cheaply on an exponential curve, so things like standalone GPS were hot Christmas items five years ago, and have now been overtaken by apps on our phones.  I doubt you can give away your old GPS today with its static maps and slow processors.   Ditto your old 5 MP digital camera.


Perhaps the more germane accelerated evolution, though, is on the human side.  Humans are an amazing species, because of our adaptability.  This allows us to take virtual trips to different planets, summit Everest where there is almost no oxygen, and probe the sea floor three miles down to search for Malaysia Airlines’ MH370.  During this latest drama, the most common complaint I saw online: why is this taking so long?  Think about that: a decade ago this search might have been practically impossible.  Now we are complaining it’s taken too long.  We just assume that of course we can do that, and somebody must be bumbling the job, otherwise we’d have found it by now.


This applies not only to the rational parts of our brains, but also the emotional parts as well.  Scientists call this hedonic adaptation, and it essentially allows us to become happier in miserable circumstances.  Again, this is a really good thing.  But the same mechanism also causes joy to become routine, and reset our happiness to near its baseline after good things happen. 


It might be obvious by now where I’m headed.  As Moore’s Law makes information cheaper and cheaper, allowing us to search and find what we need in milliseconds, we as humans adapt to that hedonically, and we very shortly come to expect that which we found previously miraculous.  We expect answers to questions instantaneously, exhaustively researched and sharpened by crowd-sourced opinions.  We have evolved into a species that checks its smartphones while waiting the two minutes for our lunches to heat up in the microwave. 


This might be okay if we were just dealing with computers.  But we’re not.  Our worlds are populated by lots of other humans, and the lines blur pretty quickly between performance we expect from machines, and performance we expect from humans.  For example, I went into my local hardware store recently for some paint to touch up a wall in my house.  I couldn’t believe how long it took to shake up the paint to mix the colors.  I first stood there for ten or twenty seconds, and then I went to find something else interesting to look at while it finished.  And, I realized I was expecting my hardware guy, Billy, to somehow make the thing go faster!


The upshot of this is we are increasingly holding humans to inhuman standards.  No errors, instantaneous execution, and replace the human if it proves faulty in any way.  We see this in our politics, our corporate life, and our culture overall.  You just need one mistake to prove any politician isn’t worthy of his current or future office, or that a CEO has failed.  We hold Congressional hearings to prove that the problem isn’t that error is inherent in human decision-making, it’s we just had bad people in the wrong places, and if you just elect people from my party, all will be well.  Never mind that those people are also, well, people.  We are increasingly tempted to treat one another as processors that wear out, and need to be replaced quickly, before whatever circuit that blew can disappoint us again.   And because we have instant access to information, we have an opinion about everything, as if we are and have always been experts on deep ocean surveying, or other things we knew nothing about a couple of months ago. Our media have become a festival of second-guessing and I-told-you-so.  Heck, our online chats are almost nothing but that.  We are used to perfect computer-based execution as the rule, with delay, error, and lack of me-centricity the detestable exception. 


I am not a Luddite; I don’t think there’s any way to go back to human-based execution with its inherent error rate as the norm.  But I do think we are losing many things, as we have hedonically adapted to the kind of performance powerful computing enables: forgiveness, empathy, and the presumption of good faith are among these.  Computers neither need nor want any of these things.  But in a world where fear and alienation seem to be the prevailing affliction, I think we humans all need some of these things, some of the time.  I know I do.  Not only do we need to receive them, I’m afraid we’re making our lives worse by forgetting to give them.  Good studies show that gift-giving, whether tangible objects or simply empathy and understanding, are at least as good for the giver as the receiver.   And that, I think, is what we paradoxically risk losing by expecting inhuman performance of humans, like you and me.  By condemning the inherent fallibility of our humanity in others, we risk disowning the same in ourselves.  


May 16, 2014 at 11:20 AM Leave a comment

Market failure and rural health care

For decades, the theory of American health care reform has been to introduce market forces and make them effective enough to control pricing and utilization, ideally at rates no greater than general inflation.  This has been true since before the failed Clinton era reform effort, and its core thesis of managed competition developed by Stanford economist Alain Einthoven. 

But what happens in locales that are too sparsely populated to support at least two competitors for a given service?  What if it can’t really support one competitor without outside subsidization?  Much of Colorado’s geography fits this description.  In the absence of market competition, what forces will match supply and demand?

The issue is more complicated than it seems.  From a purely technical standpoint, one might advocate for the decommissioning of facilities in such areas.  “Look, if the market can’t support them, then I certainly don’t want to use my tax dollars to do it,” some might say.   “If people choose to live there, then they have to pay their own way.  Many things are more expensive in rural areas, and health care is no exception.”

But health care services aren’t like others, in that people take “care” as a sign of “caring”.  If we stand by while their hospital fails, some would interpret that as a sign of neglect at best, and hostility at worst.  Recently in Boothbay Harbor, Maine, St. Andrew’s Hospital became St. Andrew’s HealthCenter, with no capacity to admit people overnight.   Despite the economic realities, the hospital seeing few emergent patients, and losing money because of expensive but underutilized staff, local residents were appalled that there would be no local hospital beds.

Someone at a local Maine foundation explained this to me.  She told me, “This isn’t just a matter of economics.  Many have contributed to their local hospitals over the years.  In some of our towns, it’s the largest remaining employer.  People identify with their local hospital, and its loss is a blow to civic pride.”

What this tells me is a lesson I have learned over and over again: the hardest problems aren’t technical, they’re cultural.  (This has to do with the American drive toward individual empowerment, but that’s a whole other post.)  Assuring adequate service to rural communities in the information age can be achieved, more cheaply and in many cases with an improvement in patient experience outcomes.   Sanjeev Arora in New Mexico is a pioneer in telemedicine through a program he started called Project Echo.  He’s demonstrated outcomes in many specialties in rural areas of New Mexico that are as good as those he achieves in his own specialty clinics in Albuquerque.  He does this by bringing nurse practitioners and other primary care providers on to his team remotely, through telemedicine.  I think he’s demonstrating that a static body of knowledge isn’t the main determinant of good care; it’s local relationships and the ability to stay current in an area through ongoing discussion with a team focused on the same kind of patient.   It tells me that the best repository of clinical knowledge isn’t single human brains last intensively programmed years ago, but a living human “cloud” intelligence supported by the actual computing cloud.  Both should continually update themselves.

Important questions loom in Colorado and other states with rural regions: how do we maintain adequate services for sparsely populated areas, at a price that is sustainable?  How do we need to start thinking about services differently in order to widen the possibilities to solve the set of equations that define this problem?  How do we honor the contributions of those who built these rural facilities that were once relevant, even vital, but before travel and information transfer were so easily available?   How do we keep up with residents’ expectations that rise as fast as those in urban areas, but whose infrastructure cannot? 

Complexity theory says that for complex problems with multiple independent actors, it is impossible to master plan solutions.  The independent actors respond unpredictably to attempts to change the system.  What can be done, though, is to plan interventions, measure results, adjust the intervention, measure, adjust, measure, etc.  Rural health care system redesign is likely one such complex problem.  Solutions to such a problem may involve innovative systems like telemedicine, mobile apps, and the local Walgreens being the main health care facility in town.  But as a cultural problem, surely anyone outside of communities suggesting any of these things for communities will fail.  Complex solutions cannot emerge without the leadership of the people affected by the problem.  The external pressures in communities are not under their control; their response to those pressures is.

May 5, 2014 at 4:35 PM Leave a comment


May 2014

Posts by Month

Posts by Category