Friday, October 9, 2009

Why have EMRs not been more widely adopted?

One of the nagging questions about EHR adoption is the following: why is there such a low EHR adoption rate at the same time that there is a broad consensus that EHRs are critical to future healthcare? Sebelius, the HHS secretary, recently noted that only about 10% of hospitals and physician practices are fully digitized. But I suspect that this number is more than generous. A recent Booz Allen report cited smaller numbers.

The difference is largely immaterial. The point is that EHRs are not nearly as widely used as one would expect given the broad consensus on the need for them.

Of course, different healthcare institutions have different kinds of challenges, but I think the challenges really amount to the three “Cs” of EHR adoption:

  1. Cost;
  2. Culture; and
  3. Content.

Cost:

When you compare the relative cost of the computers, IT support, software, etc that are necessary to implement and use EHRs, there is no question that it is more expensive than grabbing a legal pad and a pen to jot down some notes and throw it in a manila folder. Now I am not so naïve as to think that this is really how MDs track and store medical information, but for many small practices, this isn’t far off.

Even with the efficiencies that can be gained through the use of an EHR, the start-up costs are so substantial that it may never pay for itself. But for many small practices, there are no efficiencies to really gain. If you are a single MD with a single receptionist/medical record keeper and you implement an EHR that gains you even a 25% efficiency on administrative time, what do you do? You can’t exactly reduce your staff by 25%, so is there ever going to be an ROI for you?

And for lager facilities, there may be administrative cost savings, but implementing an EHR is a major undertaking that takes many years for results that are even further down the road.

Culture:

By way of analogy, thirty to forty years ago, it would have been unheard of for a CEO to type his/her own letter. Now, they write their own emails all the time. Why? It is far more efficient and they get a much quicker response. I believe we are at the dawn of a similar cultural shift with regards to the recording of medical information. MDs have been used to recording things one way and EHRs are a major shift away from that. But the problem is twofold: 1) we must make EHRs that are intuitive to use for recording information, and 2) we have to make that information available in actionable ways to the MD. In other words, the cost/benefit of email is readily apparent to CEOs, so they use it. We have to build EHRs functionality with the end in mind. The information and analysis that physicians derive from an EHR must be worth the extra effort of the information they contribute to the EHR.

Content:

EHRs are, in many instances, less intuitive and more cumbersome than paper records that can easily be modified to fit the MDs needs. And the information it produces is not actionable by its intended user. Let me illustrate.

I have a good friend who is a nephrologist in Michigan. He regularly gets referrals from other MDs in the area—some of whom dictate letters, others of whom send over an EHR. Guess which one is more useful? The letter. Typically, the EHRs come over as six or seven pages of small font forms, with most of the boxes empty, and no narrative description of why the patient is being sent. He or his staff then has to hunt through the document to find recent lab scores, and other more useful information to determine exactly why he is being asked to see this patient.

Compare this to the following dictated letter: “Mrs. Johnson is 65 and has seen decreased kidney function over the last six months.”

Part of the challenge of EHRs and the sharing of healthcare information in general is NOT capturing all the information and making it available to all appropriate clinicians, but creating the right lens for the right clinician so that the most useful information is mostly present most of the time. And, of course, there must be opportunities for exploring the information in more detail, if necessary. Now that’s a tall order.

Thus the real reason we as a country lag behind most others in EHR adoption is that EHR products haven't sufficiently proven their value to MDs such that the cost savings, convenience, quality improvements demonstrably outweigh the start-up costs and inconvenience of using an EHR. Until EHRs solve this "content" problem, we will continue to stumble over "culture".

Next up: How will the HITECH provision of ARRA address EHR adoption?

Monday, October 5, 2009

Has Healthcare Reform Already Passed?

While the Senate continues to debate the contours of health insurance reform (i.e. public option or no public option), I have recently come to the conclusion that healthcare reform has largely already been passed and approved. ARRA—or the American Recovery and Reinvestment Act—included significant funding for a variety of projects that, once implemented, will dramatically change the healthcare landscape. The section of ARRA is known as the HITECH Act, or “Health Information Technology for Economic and Clinical Health Act”

This act does a number of things from an administrative perspective, but I want to highlight for a moment some of the ways in which different provisions of the act are intended to work together. In particular I want to focus on electronic medical records.

From an administrative perspective, the HITECH Act is in the process of defining interoperability standards for healthcare data. These standards are important because they are defining the ways through which healthcare information can and will be shared across different healthcare entities. For example, right now, there are no common standards for sharing electronic healthcare information between a clinic and a hospital—they must be worked out on a case by case basis. Defining interoperability standards can change that.

But this effort also dovetails nicely with efforts on a few other fronts: 1) to establish and build upon the network of health information exchanges (HIEs), 2) to create regional “extension programs” to provide technical assistance related to Electronic Medical Records (EMRs) to individual practices, 3) to provide incentive payments to small practices who purchase EMRs.

This may not be health CARE reform, but it is health INFRASTRUCTURE reform—and this, more than the expansion of healthcare coverage being debated by the Senate now—has the potential to change healthcare practice.

Tuesday, September 8, 2009

Getting Old Fast

So there it is, folks. The trend that spells disaster for our healthcare system. Do you see it? It's that bright red line through the middle of the graph that runs counter to all the other lines. What is it? It is the line that represents the percentage of our national population taken up by people 65 and older. According to AHRQ, in 2002 about 13 percent of our population was 65 or older--yet they were responsible for 36% of our healthcare expenditures as a country. By 2030—merely 20 years from now—those over 65 will account for nearly 20% of the population according to the US Census.

But what is really scary is that people 65 and older spend, on average, more than three times the amount on healthcare as their “working adult” counterparts who are 18-65. According to the AHRQ data from 2002, people 65 and older cost just more than $11,000 per year, while “working adults” cost about $3,350. When in the 2030’s we reach nearly 375 million Americans, 19.1% of our population will be 65 or older, each of whom will cost more than three times the amount of their “working adult” counterparts. The net result? In 2030 more than 50% of our healthcare dollars will be spent people 65 or older. Of course, this assumes that healthcare costs for older and younger American’s increase at the same rate. In fact, the rate of increase in healthcare costs for the elderly has far outstripped the increases in healthcare for the non-elderly, so we should only expect this trend to get worse the longer we wait.

Sunday, August 9, 2009

Is Healthcare a Commodity? Part 1

In our current healthcare system, there is no question that we treat healthcare like a commodity. What concerns me, however, is whether this is, in fact, an effective paradigm from the perspective of healthcare consumers like you or me. In other words, is it efficient and does it produce the desired outcomes?

Ultimately, I am concerned with addressing one of the major criticisms of healthcare reform: that healthcare needs to be treated like MORE of a commodity in the traditional sense. I'm still not sure how this will address the lack of universal coverage, let alone some of the quality deficits in our system, but this seems to be beside the point of this particular line of attack.

(Incidentally, I also wanted to acknowledge a subtle change in language being used by the Obama administration. They have started calling this initiative "health INSURANCE reform, not health CARE reform. Factually, I actually think this is a much more accurate description of the current proposals floating through congress. It also implicitly acknowledges that the heavier lift of healthcare reform can come at a later time--once everyone is adequately covered.)

Concering healthcare as a commodity, I think there are three main issues that require further exploration:

  1. Is "healthcare as a commodity" an effective paradigm for consumers?
  2. Are there tweaks that can be made within the current framework to improve quality?
  3. Are there tweaks that can be made within the current framework to lower cost?

We generally make the same assumptions about healthcare that we make about other things (TVs, cars, furniture, ect). Namely, we assume that:

  1. Newer is better;
  2. More is better;
  3. Brand loyalty matters (and its offshoot- brand names names are better than generics); and
  4. There is a direct correlation between cost and quality.

In other words, all of the rules of thumb that we might use to decide which product to purchase are completely invalid in a healthcare setting. In healthcare, there is one gold standard: evidence. After that, there is clinical consensus/expert opinion. And after that, it is all marketing. Unfortunately, we are operating in an environment that lacks good comparative data on different kinds of interventions. For example, pharmaceuticals in order to be approved by the FDA need only prove they are more effective than a placebo. At no point do these drugs have to prove they are more effective and/or have a lower side-effect profile than an existing treatment.

But, a drug company that owns a patent on a new drug is highly incented to market the hell out of the drug to make a profit. Even if it is not a "better" drug, it will be a more profitable drug. And here is where the market starts to break down. Why? Well, a more expensive drug that is less effecacious is bad for the entire healthcare system even if it is where the money is. And it is wore for you and me.

As payers, insurance companies help to stand in they way of these marketing efforts by pharmaceutical companies, in part, by using physician committees (Pharmacy and Therapeutics (P&T) Committees) to establish tiered formularies for managed care, where they may decide to place this drug in the most expensive category. Yet, even if this happens, you and I will still see the ads and request the drug from our MD.

In the US, managed care is the dominant form of healthcare coverage, which means your PCP gets a monthly payment to provide coverage for you, regardless of whether you actually come in to see her. In selecting a PCP how many of you actually found comparative quality data on your PCP? Now, maybe you went on Angie's List. Maybe you looked at where they went to medical school. Maybe you talked to other people about who they go to. In other words, you used proxies to evaluate quality. But here's what you DIDN'T do. You didn't look at any actual quality measures for your MD. You didn't, for example, look at the percentage of your MD's patients with chronic conditions that are compliant with their medications, or required tests and lab values.

And that probably means you were evaluating the service you received on your PCP's bedside manner, or the friendliness of their office staff. The bottom line is you probably evaluate your experience as you would other consumer services: how well you are treated. Let's face it, whether or not the care is any good is pretty ephemeral. Now, I'm not saying that these are unimportant considerations, but they don't really tell you whether the care you received was any good.

Bringing this back home to the issue of our pharmaceutical and cost, you come in to your PCP and request the expensive, less effecacious pharmaceutical not recommended by your cost-conscious insurance company. Your MD has two choices: she can say, "well that's reallynot the best treatment" and recommend an alternative. This assumes, of course, that she is familiar enough with the literature on this specific drug and the alternatives in this drug class. However, in doing so, she also runs the risk of providing bad customer service (the cusomter is always right, right?) and so annoying you that you choose another PCP. What is more, this probably doesn't even a address category of care quality that will be recorded or reported anywhere, let alone reviewed by someone. So from the MD's perspective, she's got a ton of downside to recommending an alternative, and no upside. If , on the other hand, she just gives you want you want so long as it does no harm, you walk away thinging that you've got a responsive PCP and you may even recommend her to your friends.

Now I amd NOT suggesting that players in this system are simply automatons that follow the dictates of the market. But this is an illustration of ways in which market forces confound each other in our system. If I was in her shoes, I know which decision I'd make.

Clearly, then, our market rule of thunb--"newer is better"--is not effective for evaluating healthcare quality.

As for the "more is better" rule of thumb, I would encourace everyone to read Atul Gawande's piece in the New Yorker from June. Not only is more NOT better, it often has deadly consequences.

Without going through and listing (and debunking) each of our rules of thumb, to me the overall conclusion is that the healthcare system is very much caveat emptor. Now, towards the end of Gawande's article he goes through the exercise of debunking the marketplace argument in the form of hospital use. (i.e. someone can't decide to wait for a better technology when they are going in for diagnostic imaging for metastatic cancer, or to question the cost-effectiveness of certain tests when they are receiving urgent/emergent care).

But I, for one, think better information comparing physician quality could help individuals to make better decisions about "purchasing" quality services from a provider. I would love to have the mandate of the Center for Comparative Effectiveness Research expanded to include MDs as well as treatment protocols.

Unfortunately, the problem with this is twofold:

  1. Quality reporting only covers a very narrow set of conditions--how do you ensure that your PCP can actually provide quality for the conditions you have and not just those that get reported on? And
  2. How do you control for differences in a patient pannel for a particular physician?

Let's assume for the sake of argument that these can two issues can be addressed. Although this would be a major step in the right direction and one I would fully support, it may not matter that much from a cost perspective.

If comparative quality information in physicians was available for PCPs, we would be "purchasing" quality at the least expensive level--outpatient ambulatory care. As Gawande notes in his article, the place where people are most likely to cost a lot is in an inpatient setting where they have the smallest capacity to be active participants in their own care. Even if we consumers HAD better information on comparative cost and quality, how exactly would we (be able to) use it?

To summarize:

1) Is "healthcare as a commodity" an effective paradigm for consumers?

No. Consumers generally lack the information to make informed decisions based on quality and cost, in part because the rules of thumb we use on our normal consumer lives are misapplied to healthcare. What is more, the countervailing forces in the healthcare marketplace in general confound attempts to make "quality purchasing" effective.

2) Are there tweaks that can be made within the current framework to improve coverage and/or quality?

Yes, but to a limited degree. Putting more comparative quality information into the hands of consumers will facilitate better decisions about whom to select as a PCP. Bu this may have limited utility in part because the state of quality reporting currently only addresses a limited number of conditions that may/may not be significant for all consumers and/or physician specialists.

3) Are there tweaks that can be made within the current framework to lower cost?

Not really. Most costs are incurred in inpatient settings where consumers have a limited ability to manage their own care (and the equippment it requires). Better preventive care in outpatient settings (a likely outcome of better quality reporting) in fact will increase costs because more people will get more of the right tests and will take more of the right medications. This may or may not lead to reduced inpatient stays, but it will improve and increase QALYs.

Friday, August 7, 2009

Coming Soon: Is Healthcare a Commodity?

Much of the ideological opposition to a national healthcare program hinges on whether healthcare can be considered a commodity--akin to electronics or legal services.
My next entry (after a brief August break) will address this important topic as well as some of the ways market forces are brought to bear on healthcare services.
In the meantime, feel free to share your thoughts here.

Wednesday, July 29, 2009

A Fragmented System with Multiple Frames of Reference

I heard bits and pieces of Maggie Mahar on NPR’s Fresh Air the other night. On the show, she made a point that I think is worth repeating and fleshing out a bit more—along with a Healthcare Intelligence twist. In talking with Terry Gross, Mahar said that the healthcare system was positively Hobbesian. Hmm. How so? Well, remember back in college when you read The Leviathian? Perhaps not. But surely, you remember Hobbes description of life in the absence of society—the intense competition between individuals, the lack of specialization, and the lives that were “solitary, poor, nasty, brutish, and short.”

Welcome to the US healthcare system! As Gross paraphrased it, MDs compete against MDs, who compete against insurance companies, who compete against other insurance companies, who compete against pharmaceutical companies, who compete against other pharmaceutical companies, who compete against regulators, etc. You get the idea. Eew. Mahar’s points were that this competition has made the system less stable, more expensive, and more fraught with poorly coordinated care. And, of course, she’s right.

Our healthcare system—from financing, to service delivery, to data tracking—is inveterately fragmented. And as a result, different healthcare entities (patients, primary care physicians, hospitals, pharmaceuticals, insurance companies, etc) have slightly different frames of reference with regards to healthcare data, healthcare costs, and healthcare quality. They differ in terms of what data they have access to, the combination and constitution of elements related to cost and quality that they pay attention to, and in the leavers they can use to manage costs and improve quality.

Hospitals, for example, are looking not only at quality issues, but how those quality issues intersect with their supply chain, the use of materials for various procedures, etc. Many hospitals have moved towards the use of Diagnosis Related Groups (DRGs). These provide reimbursement to hospitals on the basis of a certain diagnosis, regardless of the services provided and, ultimately, encourage hospitals to become more efficient. At least in theory. You can think of it as managed are for hospitals. But this ultimately places a greater emphasis on the use of materials and supplies, so long as it doesn’t have too much of a detrimental effect of quality.

(When my daughter was born, the nurses looked like they’d gotten into a fight in a sticker factory. For each piece of equipment they pulled out—saline drips, latex gloves, vacuum extractor, etc—there was a convenient barcode sticker. As each nurse took a new something, they’d peel off the barcode, and stick it to her scrubs. At the end, it was like being in the checkout line of a grocery store. They zapped all the barcodes and catalogued all the supplies and equipment they used. All I kept thinking was, “Boy, I’m sure glad I’m not paying for this.”)

If a hospital can reduce the number of certain supplies for a particular procedure, they can reduce their costs. They can also start to look at the cost of storing their supplies, as well as the quantity to purchase at a single time. All of this can affect a hospital’s bottom line. But all of this is likely to have some effect on quality. How much? Well, that’s where healthcare intelligence would come in—to help evaluate the change in terms of quality outcomes, and it would need to control for a variety of factors including the relative sickness of various patients, any seasonality issues, demographics, etc.

I can devote another entry to some of the different kinds of quality certifications that different healthcare agencies must obtain, but suffice to say that Hospital certification—specificallyJCAHO--is a very complicated process that requires hospitals to monitor and report on their performance on a number of different measures. Although it will rarely be a single measure that would prevent a hospital from achieving certification, it is not difficult to imagine that a hospital might devote significant resources to improving marks in a certain area. Or perhaps they might decide that they are prepared to accept a declination in a certain quality area where they are already strong in order to improve their bottom line.

Mind you, this may be the right thing to do—cost and quality are inexorably linked, mostly in inverse proportions. And here again we get to our $1 Trillion question(s): how good is good enough? And how much are we prepared to pay to achieve it?

Friday, July 24, 2009

Obama and Unwarranted Variation

I watched with interest Obama’s press conference from Wednesday night, as I’m sure many people did. I think he handled himself quite well and showed an impressive command of the issues. I think the discussion definitely favors healthcare reform when it is kept high level, which is why it concerns me that the Senate is now delaying a vote on healthcare until after the August recess.

A few high level things that I noted throughout the conference: there were at least two points at which Obama used the term “unwarranted” in his answers. One was when he was talking about tonsillectomies; the other was when he was talking about bonuses for bank executives. Of course, “unwarranted” is the first half of a famous phrase in healthcare: Unwarranted Variation. This is the phenomenon observed in Medicare data that the delivery of healthcare in different regions around the country differs for reasons that have nothing whatsoever to do with illness, medical need, or evidence based medicine.

These observations were made by Dr. Jack Wennberg and colleagues at Dartmouth, and have been documented in an important project called The Dartmouth Atlas. The Dartmouth Atlas, in turn has been referenced by the likes of Peter Orszag, Obama’s Director of the Office of Management and Budget. (Incidentally, Orszag's Special Advisor for Health Policy is Ezekiel Emmanuel--Obama Chief of Staff Rahm Emmanuel's brother.) By some estimates, if unwarranted variation was completely eliminated, it would reduce Medicare costs by as much as 30 percent.

(As an aside, I think it is interesting that he is starting a subtle mantra of unwarranted variation as a bad thing, and the unseemly association of “unwarranted” with bank bonuses. If everyone had the same visceral reaction to unwarranted variation that they have to unwarranted bank exec bonuses, we'd all be much better off.)

In essence there are three categories of Unwarranted Variation:

  • Effective care: Care where the evidence is incontrovertible, yet not delivered. For example, if you have had a heart attack, you should be taking beta-blocker (BB) medication. Only about 45% of patients are adherent with a BB in their first year post-heart attack. This under-use of care is unwarranted variation.
  • Preference Sensitive Care: Care where the evidence supports different interventions, each with equal clinical validity. NOTE: on the link, there is a box in the lower left hand corner of the front page that talks about…wait for it….tonsillectomies. That’s right, the very example that Obama gave in his press conference concerning preference sensitive conditions is talked about as the classic example of preference sensitive care. Another more recent example (with great cost comparisons of treatments, is David Leonhardt’s discussion of Prostate Cancer in the NYTimes.
  • Supply Sensitive Care: Care that tends to be driven the availability of services, not by improvements in quality. It is, in part, related to Roemer’s Law, but has broader implications. Diagnostic imaging, is a great example. When a new MRI machine is available in a community, it gets used, but the quality of the care being delivered does not appreciably improve. But perhaps most scarily, on pg. 3 of the attached brief, there is actually evidence to suggest that more care isn’t just more expensive, it can be deadly. As the brief states, “The study (comparing outcomes for patients with one of three conditions) showed increased mortality rates in regions with greater care intensity.” Yowza.

I think unwarranted variation is a great example if the way in which the healthcare debate is extremely complicated. This suggests that the problems with our healthcare delivery system are not merely about providing more care or less care, but both—as well as care that is just different.

Let me explain: if we are talking about delivering effective care, people need it, so we are talking about MORE CARE. If we are talking about supply sensitive care, we are talking about delivering LESS CARE, because too much of it is a bad thing. And if we are talking about preference sensitive care, we need to deliver DIFFERENT CARE.

And one of the real questions for healthcare reform, is how to use public policy to address unwarranted variation. This is a fairly nuanced point that has trouble making it through the cacophony of vitriol on healthcare reform. However, steps are being made in this direction.

Perhaps most importantly, the ARRA legislation (the “stimulus bill” that was passed in February) included a provision to create the “Federal Coordinating Council for Comparative Effectiveness Research” and funded it with $1.1 Billion. Some of this will get at issues of unwarranted variation; others will address lacunae in research around the delivery of qualityhealthcare.

As I said earlier, I think the major thrust of this round of reform efforts is on coverage. Lets get the 45+ million Americans without coverage covered. Then lets really start tweaking the system so we 1) know what quality is; 2) measure and report on it; and 3) incent providers to provide it—financially or otherwise.