Tuesday, February 28, 2017

Health Care Should Get "Smart"

Admit it: you're worried about your online privacy.  Admit it: your personal health information is one of the things you worry most about getting hacked.  Admit it: you don't understand why your health care providers seem to have a hard time sharing key information about you.  And admit it, you're not quite sure what health insurers really do, except for always saying no and for getting between you and your health care providers.

This is why blockchain is the new hope -- or hype -- for health care.  What intrigues me most about it, though, are its "smart contracts."

The GAO recently cited health as a key area of cybersecurity weakness, and TrendMicro profiled why cybercrime is a particular threat for health care.  The 2017 Xerox eHealth Survey found that 44% of Americans were worried about their personal health information being stolen, and one has to wonder if the other 56% are asleep or just don't care.

So it is no wonder that blockchain, with its touted higher level of security, is the new darling for health care pundits.

A previous post attempted to convey the blockchain basics and the hype (some of which interested readers were quick to try to deflate).  A recent article in NEJM Catalyst attempted to "decode" the hype, analyzing both the potential applications of blockchain in health care and the some of the challenges it would face.  The authors warned that:
the challenges and realities of health care and health care data may be insurmountable — even as some argue that blockchain could revolutionize how we share health care data.
The financial service industry, for one, is paying attention to blockchain.  A consortium of big banks and tech companies just announced a consortium to expand blockchain technology.  Improved security would be nice, but what may be catching their eyes most is the prospect for saving money.

Accenture estimated that blockchain could reduce investment bank's infrastructure costs by 30%, some $8b - $10b annually for the 10 largest such banks.  It could slash the cost for many back office functions, such as reporting, compliance, and clearance/settlement.

Smart contracts are one of the features blockchain offers to help achieve these savings.  Smart contracts are, essentially, automated programs that self-execute and self-enforce, based on satisfaction of the underlying terms. They can work between two parties, or for complex multi-party agreements, and do so without any middlemen -- no lawyers or other third parties.

Smart contracts have been compared (for better or worse) to a vending machine: put your money in, get your desired choice out.  Nick Szabo, who may or may not have invented Bitcoin, uses the vending machine analogy, and elaborates:
Think of blockchain as an army of robots checking up on each others' work.  Where traditionally you have accountants and lawyers, there are now a wide variety of things we can do with the vending machine-like mechanism to replace the job of traditional contracts plus added cryptographic mechanisms for integrity."
Electronic health records and exchange of patient data more generally are often cited as obvious potential uses for blockchain in health care, as the data is not stored in silos and can be shared by trusted partners.  Those may prove valid uses, but blockchain's killer app may be smart contracts -- such as to cut out health plans.

You get services from your health care provider, but your health plan decides if your contract covers the care, and how much, if any, it will pay towards the care.  Your provider may or may not be in your health plan's network, they may or may not have to contact your health plan for "permission" for some services, may or may not submit a claim on your behalf, and may or may not receive payment directly from them (as well as from your for your portion).

So, yes, they are in the middle.

Your health plan is really a contract between you and your health insurer.  If you pay the required amount, they are obligated to provide payment for a specified set of health care services -- under a specified set of conditions: application of deductibles/coinsurance/copayments, use of provider networks, limits on certain services, medically necessity, excluded services, etc.  Like many insurance contracts, to understand it you probably need a law degree.

Imagine a smart contract between you and your doctor.  She promises to, say, fix your broken leg and, if she does, you promise to pay her $X.  The two of you would agree how and when to decide if the leg is, in fact, fixed.  All that goes in the smart contract.

No ICD-10 or CPT codes, no unknown charges, no payment for care that doesn't work; just mutual agreement about what each party will do.   You fund a virtual currency account, she provides the services, the smart contract monitors when the conditions are met, and issues payment once they are.

None of that looks like the current health care system.

The broken leg is, admittedly, a relatively simple example, and even it would require some work to define the contract.  How badly is the leg broken?  How do you know if the price is reasonable?  How do you evaluate if the leg is fixed to your satisfaction?  And all this is supposed to happen while you are in pain in the ER or doctor's office?

However, none of that is insurmountable, nor alien to advocates of reference pricing or value-based care.  Defining expectations about care, price, and outcomes prior to receipt of care would be a real change, but one that would go a long way to changing the way our health care system works.

It is easiest to see smart contracts working for conditions with clear expected outcomes, costs that are relatively affordable, and between two parties.  However, they don't have to be so limited, and could be designed for use with multiple parties, be they multiple providers or multiple funding sources (e.g., using a Kickstarter campaign or even a catastrophic health plan).

Blockchain may prove to not fit the structure we've evolved in health care.  Smart contracts may not be made smart enough to understand what happens in our health care system.  But maybe, just maybe, blockchain could be the next big thing in health care, and smart contracts may be its killer app.

Tuesday, February 21, 2017

The Good, the Bad, and the Ugly in Health Care

I hate being a patient.

I have to admit that, although I write about health care, I am typically what can be described as a care-avoider.  My exposure to the health care system has mostly been through my professional life or through the experiences of friends and family.  The last few days, though, I unexpectedly had an up-close-and-personal experience as a hospital inpatient.

I want to share some thoughts from that experience.

Now, granted, any perceptions I gained are those of one person, in one hospital, in one medium-sized mid-western city.  Nonetheless, I offer what I consider the Good, the Bad, and the Ugly of the experience.

The Good:  The People

The various people involved in my care, from the most highly trained physician to the person who delivered meals, were great.

I loved my nurses; they fit all those great stereotypes people have about the profession.  Attentive, caring, cheerful, knowledgeable, hard-working -- the list goes on and on (full disclosure: I'm married to a nurse, so none of this came as a surprise).

I liked my doctors a lot.  Each of them spent literally hours with me -- answering my (many!) questions, discussing what they thought was going on with me, describing the various tests or procedures, developing care plans to fit me.  They were super-smart and a pleasure to talk with.

The aides, the lab techs, the imaging tech, the transportation specialists -- all of them doing jobs that I wouldn't be able to do -- were each friendly and helpful, taking pride in what they did and how it helped my care.

Whatever you might say about our health care system, you cannot say that it is not filled with people who don't care about the patients in it.

The Bad: The Processes

On the other hand, on the lists of criticisms about our health care system, many of its rules and processes truly do deserve a place.  They're like part of an arcane game no one really understands.
I'll offer three examples:

  • Check-in: I was literally on a table in a procedure lab -- still wondering how the hell I'd ended up there and not quite sure what was about to happen to me -- when I was asked to electronically sign several forms (Privacy Policy, Consent to Treat, Consent to Bill) that I could neither see nor was able to question.  No court of law could call that informed consent, but that's what the process required before I could actually receive care.
  • NPO:  At one point it was thought that, on the following day, I might have a procedure, so I had to be NPO (no food or water allowed) for at least 4 hours -- but starting at midnight.  I pointed out that it was highly unlikely that they'd be doing anything at 4 am, and even mid-morning was unlikely since nothing was yet scheduled, but that was not persuasive.  As it turned out, I'd gone something like 16 hours NPO when they finally listened to my concerns: by putting me on a saline solution IV.  I think they understood the physical problem but not the human one.  (It ended up I didn't have the procedure anyway.)
  • Discharge:  On my final day, the doctor told me around 1 pm that I was being discharged.  Around 3 pm his nurse practitioner told me she'd personally written the discharge orders.  Around 5:30 pm my nurse gave me all my discharge papers, but told me I had to wait for Transport to escort me out in a wheelchair (even though I was perfectly capable of walking).  Finally, around 6:30 pm my wife simply commandeered a wheelchair and we made a break for it.    
The rules and processes are all undoubtedly in place for good reasons, but we need to un-handcuff all those great people when rules and processes get in the way of better patient care.  

The Ugly: The Technology

Oh, health care technology.  It is equally capable of delighting as it is of frustrating.  It is truly remarkable that the doctor could go up my arm to perform a procedure in my chest, just as the detail an MRI provides is simply astonishing.  

On the other hand, those gowns...

Let's start with the perennial whipping boy, EHRs.  All of the staff used them, seemed to accept them, and even (grudgingly) acknowledged their value.  But no one liked them.  Even the youngest users, to whom technology is a given in their personal lives, were frustrated by the interface.  And, on many occasions, EHRs did not mean that people did not still often have to drag in other electronic equipment or even paper in order for them to do their job.

EHRs could be better, should be better -- and better get better.

MRIs are a wonderful technology, but as I was laying in that claustrophobic tube getting imaged, I kept thinking: what the heck are all those clanging noises?  We can make stealth submarines, but we can't make an MRI that is quiet, so that anxious patients don't have more to worry about?

I was on various forms of monitoring devices, the smallest of which was the size of a 1980's cell phone and still required countless wires attached to numerous leads.  I kept looking at the set-up and wondering, hmm, have these people heard of Bluetooth?  Do they know about wearables?

My favorite example of ugly technology, though, came when I had to fill out a form (which looked like it dated from the 1970's), so that it could be faxed to the appropriate department.  That's right, faxed.  To a department in same institution, in the same building.  I couldn't fill it out online?  A paper form couldn't be scanned and sent securely to the other department?  

I'd love to be the boss of the guy who has to request a new fax machine, just so I could look at him with my best "you've got to -be-kidding-me?" expression. 

No health care system is perfect.  Every system has its own version of the Good, the Bad, and the Ugly.  No one wants to have health problems, and no one wants to need to be in health care settings.  When we do, and when we have to be, though -- well, our system can do better.  Let's give all those great people working in health care a better chance to help us.

If any of the above strikes home for you, perhaps you'll Like/Recommend, tweet, or otherwise share with your circle.

Monday, February 13, 2017

Ask Better Questions

I've been thinking about questions.

A few things I read helped spur this.  The first was a blog post entitled "Asking the Wrong Questions" by Benedict Evans, of VC firm Andreessen Horowitz,  Mr. Evans looked at a couple of long range technology forecasts (from as long ago as 1964 and as recently as 1990), and pointed out how they both managed to miss several key developments.  He attributed this to "this tendency to ask the wrong questions, or questions based on the wrong framework."

And we're still at it.  Mr. Evans, whose background is mobile technologies, said that people are now doing a lot of speculating about what comes "after mobile," such as AR and VR.  There are lots of good questions being asked, he noted, "But every time I think about these, I try to think what questions I'm not asking."

That, my friends, sounds like some pretty good advice, especially if you fancy yourself an innovator.

Then there was an interview with Warren Berger in Singularity Hub.  Mr. Berger labels himself a "questionologist" -- a line of work I wish I'd thought of! -- and wrote a 2014 book A More Beautiful Question.

You have to admire his ability to turn a phrase; I love the notion of a "beautiful" question.

Mr. Berger defined a beautiful question as "an ambitious yet actionable question that can shift the way we think about something and may serve as a catalyst for change."  As he further explained:

  • “Ambitious” because we have to ask bold questions to innovate
  • “Actionable” because big questions we can’t do anything about don’t lead to change.
  • Critically, the question has to cause a mental shift—it makes you step back and say, “Hmmm, that’s interesting. I hadn’t really thought about this question before, and I want to explore it further.”

He sees these kind of questions as important not just for technological innovation, but even basic questions like "what business am I in" (something, for example, the folks at Snap have recently been asking, with some surprising answers).  He further suggests organizations should turn mission statements into mission questions, to remind people to keep questioning.  And, of course, he urges that leaders foster a culture of inquiry, without demanding immediate answers to every question.

One of Mr. Berger's favorite examples is how the Polaroid instant camera came about because founder/CEO Edwin Land's three-year-old daughter asked why they had to wait to see the picture he'd just taken.  As Land later recounted, “I thought, ‘Why not? Why not design a picture that can be developed right away?’”

One could argue that Polaroid's downfall came because it stopped asking "beautiful" questions.

In 2015, Tom Pohlmann and Neethi Mary Thomas, of decision science firm Mu Sigma, wrote in Harvard Business Review that we need to "relearn the art of asking questions.   They claim that "proper questioning" has become a "lost art."  They lament that, while most of small children's conversations consist of questions, the opposite is true for adults, and blame this on educational and workplace environments that reward answers, not questions.

They categorize questions as follows:

I suspect that too many questions in most organizations would be considered "clarifying," and not very many at all would be classified in any of the other three quadrants.  The authors agree with Mr. Berger that leaders need to encourage people to ask more questions, because: "In order to make the right decisions, people need to start asking the questions that really matter."

Let's turn this to health care.

Patient engagement is one of the hot topics in health care.  Improve patient engagement, the theory is, and all sorts of good things will happen.  Patient compliance with instructions would improve, we'd do a better job managing chronic conditions, and patients would have better ongoing attention to their health.  The questions being asked are often revolve around how can we use technology to improve patient engagement.

Certainly technology could improve patient engagement, but let's start from a different point.  I would be willing to bet that all or almost all providers have mechanisms to track payments, have statistics on late payments to which they pay close attention, and have procedures in place to reach out to patients whose payments are considered late.

On the other hand, I'd similarly bet that very few providers have mechanisms to track patient status post-visit/procedure/prescription, other than perhaps a simple follow-up phone call.  If the patient doesn't contact them to complain, all is considered good.  As a result, no tracking mechanism, no statistics, no procedures to escalate anything if a patient's status is not going as expected.  

So we have a form of patient engagement already, but it is built around money, not patient well-being.  Putting in a patient portal or an app doesn't change that underlying focus.  It's addressing the wrong question.

Or let's look at the big question that confounds everyone -- why does the U.S. spend so much more than every other country on health care (and yet only has mediocre health results)?  Maybe we're simply counting the wrong things.

Most of us have, by now, probably seen a version of this chart, detailing the drivers of our health care spending.

Basically, medical care gets the lion's share of the money, but is in itself not a major driver of health.

We're somewhat unique in this.  We invest in medical care, while most other developed countries spend more of their money on "social care" -- better housing, education, support income, etc.

The AHPA charts it as follows:

When you combine OCED's health spending figures and their social expenditures, the gap in "health" spending between the U.S. and other developed countries narrows dramatically (we're still higher, most likely due to our insanely high prices).

We're not asking the right questions -- or even looking at the right problems.

I'll close with two applicable quotes:

Tuesday, February 7, 2017

Dr. Leopard, Your Spots Are Showing

Just a few years ago, things were looking up for the American health care system.  We were going to start finding better ways to pay for care: call it pay-for-performance (P4P), value-based purchasing (VBP), or similar terms.  We were going to nudge -- or, rather, push -- providers into more clinically integrated systems (e.g., ACOs) to help improve outcomes and to control costs.  And, of course, with wider use of electronic health records (EHR), we'd be able to better coordinate care and make decisions based on actual data.  It all sounded very promising.

Now, though -- what's that old expression about the leopard not being able to change its spots?

Let's start with EHRs, As Dave Lareau of Medicomp Systems told Healthcare IT News, "the concrete has already been poured."  For better or for worse, we've got the widespread diffusion of EHRs that we were hoping for.  Unfortunately, it seems more for worse.

They're not considered user-friendly, interoperability is as much of a barrier as ever, and the sense is that they take away more from patient care as they bring to it.  Moreover, they're woefully unprepared for the flood of data that wearables and other mobile tracking devices are already starting to generate.

Mr. Lareau further noted that "their main purpose was for reimbursement -- to get it over to billing."  Jon Melling, of Pivot Point Consulting agreed: "As we move to value-based reimbursement, we have a variety of venues to select, including value-based care and fee-for-value, which are incompatible in the system."

Oh, yes, about those new payment mechanisms.

Harvard's Ashish Jha, MD, MPH, says that: "the evidence on P4P in general is largely mixed, and the evidence on Hospital Value-Based Purchasing (VBP), the national hospital P4P program, is discouraging."

According to Dr. Jha, VBP has had no positive effect on either mortality or patient experience, and this should come as no surprise.  He'd noted several years ago that successful P4P programs must have three design features:

  • incentives large enough to "motivate" investments in improving patient care;
  • focus on a small number of high-value measures to drive practice changes;
  • a simple design that people can know how they are doing.
VBP failed all of these, in his opinion.  

Dr. Jha acknowledges that some critics believe P4P and VBP "fundamentally cannot succeed," because our quality measurement abilities are "woefully inadequate" and resulting performance measures are so flawed that it is easier to game them than to use them to actually improve care.  He is more sanguine, believing that these programs can work if designed properly, but admits that "none of these changes will be easy."

Meanwhile, professors Stephen B. Soumearai and Ross Koppel, writing in Vox, flatly assert that physician P4P "doesn't work."  People believe in it, they say, because Econ 101 would predict that performance will improve if we pay for outcomes, and because several studies claimed to show a positive impact -- studies they believe have "fatal flaws."

They cite studies that either don't take into account improvements that were already happening prior to P4P, or ones where there simply were no differences in performance under P4P than from a control group (see figure below).  In their words, "when you single out the most rigorous systemic reviews, empirical support for pay for performance evaporates."

Effect of financial incentives to physicians, patients,
or both on lipid levels
. JAMA 2015.

It's worse that that.  With these programs, they point out that we're adding some $15b in regulatory burdens on physicians alone, and may also be discouraging physicians from treating sicker patients, due to concern over how they might impact their statistics.

Not exactly what we were hoping for.

Like Dr. Jha, Dr. Soumerai and Dr. Koppel aren't entirely discouraged, having faith that providers simply want concrete information -- based on better research about the reasons for poor performance -- that will help improve care.  They cite the ever-quotable Uwe Reinardt:
The idea that everyone’s professionalism and everyone’s good will has to be bought with tips is bizarre. 
I don't think I've heard P4P ever called "tips" before, but that's not far wrong.

Muhlestein & McClellan 2016
Then there are ACOs.  Their number has skyrocketed since the passage of ACA, with there being close to 1,000 nationwide.  Whether they've been effective in controlling costs or improving quality is less clear; at best the jury is still out, at worst the answer has been no.

What we have seen, though, is that provider consolidation has been on a spree in recent years, with no end in sight.  The argument for it is that such consolidation is necessary for the kind of clinical integration that ACOs and P4P require.  This is despite the fact that such consolidation has not delivered lower costs or better quality; if anything, costs have increased with it.  

As it turns out, though, the consolidation bears little relationship to ACO penetration or physician participation in them, according to research by Neprash et. alia.  The post-ACA consolidation simply continues previous trends, although it may now be "defensive consolidation in response to new payment models."

Which, it would seem, may not really work anyway.  

So, it would seem, our health care system can't quite seem to change its spots.  It's taken every reform we've thrown at it -- every new delivery approach, payment mechanism, regulatory oversight, new competitors -- and come out virtually unscathed.  Costs keep going up, unnecessary care continues to be delivered, and thousands of lives are damaged or lost that didn't need to be.

You can't blame the health system, or the people in it.  For the most part, everyone in it is just doing what they think is their job.  It's not going to change, not on its own.  Why would it?

Maybe it is us who have to change our spots.