The healthcare industry has always been quick to jump on buzzword bandwagon. As an analytics and business intelligence company; we are confronted with clients and prospects that have been romanced by new ideas and concepts. Many of these are well worth additional effort on our part. Others turn out to be convoluted ideals that are well intended, but not always completely thought out.
Of late it seems that the new Don Juan of healthcare is a combination of predictive analytics and risk scoring. Both of these functionalities have their merits; however, the expectation of the benefit to be derived is inflated.
There is a certain “carnival fortune teller” mystique to predictive analytics. If we can see the future, then we can cure all of the problems to come. Unfortunately, the future is based upon the present, which in turn is based on the past. If you do not understand what is happening to your population now, if you have not learned from their past; how can you hope to alter the future?
Risk scoring is simply a combination of patient demographics, lifestyle choices, and comorbidities. Compiling all of this information leads to a score that indicates your sickest individuals, presumably both now and in the future. By throwing a number against a person, we supposedly can address who requires the greatest resources. The problem is that a number is not an individual. We need to remember that healthcare consumption is also impacted by social, economic, and emotional overtones!
Step outside of healthcare for a moment and ask any IT or operations manager what one characteristic is shared by their most successful and effective software. The answer they will share with you is that it is purpose-driven and actionable. In short, analytics results are simply information. By being able to take action on that information, you reach knowledge. By measuring the results of that knowledge application, you attain wisdom!
The major problem with the two topics we have addressed so far is three-fold. First, risk scores and predictive analytics have to be built off of a solid data foundation. If you cannot get to all of the data for an individual, you are getting only a partial picture. Second, even if you can get all the data; more than likely a large portion of it reflects dubious data quality. Third, presuming that you have all the data and it is of the highest quality; you still only have a static picture. In short, the actionable component is missing. Nothing changes if all you do is look at graphs and gauges, the entire time wringing your hands!
The taste and appeal of the sandwich always lies in the secret sauce; so here is some hush-hush insight into your population health. There is a magic way to reduce risk scores and to make your predictive analytics rosier. By taking action on care gaps you can directly affect outcomes. Do not measure risk; rather, reduce it! Do not predict outcomes; instead, create them! Use todays analytics to surface care gaps, to communicate the gaps to care teams, to engage the provider and the patient, and to deliver the right care at the right time by the right provider in the right setting.
We agree whole-heartedly that risk scoring and predictive analytics have their place in the healthcare ecosystem and that they deliver value. However, it is time for healthcare to stop reacting and instead start responding. In its simplest form, plug the care gaps and you will alter the risk score and the future. By using analytics to determine what those gaps are; you accomplish a simple, yet honored goal called the Triple Aim. You will improve quality, lower costs, and provide a better experience for patient and provider alike!
Here’s an alarming fact: the meaningful use dropout rate is already 17%.
A recently published assessment of the government’s April EHR attestation data revealed that 17% of the providers who earned an $18,000 EHR incentive in 2011 did not earn the $12,000 second incentive in 2012. Although the analysis was performed by the venerable Wells Fargo, my immediate response was, “That’s impossible! They must have miscalculated the data.”
So I crunched the numbers for myself, and to my astonishment, the conclusion is absolutely correct. A staggering 17% of the providers who succeeded at demonstrating meaningful use for 90 days were unable to sustain that performance for a full year—the second required reporting period—despite the fact that the program’s requirements remained exactly the same and the providers already had the necessary workflows in place to support those requirements. What makes this fact even more troubling is that the 2011 attesters were typically the early EHR adopters and therefore most experienced in the use of the technology.
A 17% loss rate in any business is wholly unacceptable, and this failure does not portend well for the future of the EHR Incentive Program. If $12,000 proved to be insufficient motivation for physicians with meaningful use experience to meet the relatively low requirements of Stage 1 on an ongoing basis, it would be foolish to expect physicians to muster the wherewithal to meet the increasingly demanding requirements of Stage 2. The incentive for a year’s performance at that point will be a mere $4,000.
Compounding this finding is the fact that 14% of physicians who attested to Stage 1 have already stated that they have no intention of attesting to Stage 2, according to another recent survey. And we can be sure that this number will rise as physicians begin to familiarize themselves with the labyrinthine requirements. If physicians are not motivated by the remaining incentives, it’s equally clear that the imposition of penalties for noncompliance will yield no better results. There is already a groundswell of objections to the penalties, including a bill introduced in the House seeking numerous exemptions, letters from AMA and AHA, etc.
So, is this the beginning of the end of meaningful use? What is keeping physicians from continuing to participate in the program? Are they bailing or failing? In either case, it is just too complicated—physicians are demonstrating that they are not willing to divert their attention from treating patients to consistently devoting the time necessary to keep track of the myriad measures on which they must successfully report. Instead of making meaningful use increasingly complex, we need to simplify it—focus on interoperability and leave the physicians and their clinical staffs to practice medicine. If we do not, the entire program will go down the drain. Let’s not throw the baby out with the bathwater!
Healthcare has always been infatuated with new buzz words and sexy concepts. Case in point; do a search on your favorite search engine using this string: “healthcare big data.” My own test of this search yield 136 million results on Google. In fact, just in the past 24 hours; over 1000 new results have made it onto Google. Apparently writing about big data has become big data in and of itself! Ah, the irony!
As a society, we Americans have a tendency to believe that more is better. That said; our supersize mentality has led to an obesity epidemic that shows no signs of relenting. Case in point; the contents of a children’s Happy Meal was the actual adult portion in the 1970’s!
We in the healthcare profession have fallen prey to this enigma. We assume that if we can get our hands on endless volumes of data; somewhere in there we will discover the magic bullets that enhance quality and reduce costs, all the while creating happy patients and providers! Big data mean big results, right? Wrong! In short, big data is getting in the way of the big picture. That big picture is providing the right care, to the right person, at the right time, by the right provider, in the right setting!
Perhaps it is worthwhile to assure that we all understand the meaning of big data. By pure definition: Big data usually includes data sets with sizes beyond the ability of commonly used software tools to capture, curate, manage, and process the data within a tolerable elapsed time. If you are a multinational fast-food chain trying to determine how many Tweets were written about your competitor’s new sandwich; then you are dealing in big data. If you are trying to map the entirety of the human genome, you need a big data approach.
Big data is three-dimensional; made up of the “Three V’s.” Volume is the amount of data. Velocity is the speed at which data moves in and out. Finally, Variety is the range of data types and sources. Let us now compare healthcare data to social network data, in light of the big data paradigm.
Volume – The average person Tweets about 1,600 times per year, yet only sees a physician 3.2 times in the same year. Suddenly, healthcare data volume seems insignificant.
Velocity – Every year, Facebook users make over 20 billion status updates, yet the number of healthcare transactions in the same year comes in at 1.2 billion. It appears that healthcare data is low velocity.
Variety – Claims records follow one of three formats (ANSI X12 837, 835, and 834) and provider electronic health record or HIS systems capture a limited number of fields, numbering perhaps in the thousands. By contrast, Amazon has well over one million SKUs. There goes the variety argument with respect to healthcare big data.
The fact that healthcare data does not pass the Three V’s test, comparatively speaking; one has to wonder what our collective fascination is with big data. Why are we entertaining the use of Hadoop, massively parallel processing, and storage devices that can handle petabytes? Again, one would be inclined to believe that the more data you grab; the more likely you will find a way to increase quality and / or lower costs. That is a case of faulty logic!
If a provider or payer can identify one diabetic who has not had an A1c test in over a year; the chance of them showing up at the ED in diabetic shock are lessened. If a diabetic who is not adhering to the medication regimen can be identified; the chance of them experiencing a lower limb amputation is lessened. If I can isolate the diabetic who has not had a recent vision exam; chances are I will lower the odds that they will have future retinopathy.
By identifying the gaps in care described above for diabetics; we can assure three specific results. First, we will enhance the provider and patient experience. Second, we will improve quality of care. Third, we will lower overall costs for the system. Sound familiar? It should! We just described the goals of the Triple Aim.
Do we need advanced query and analysis tools to meet this goal? Do we need to invest in massively parallel processing software and hardware? Do we need storage so large that it becomes unfathomable?
Give me a RDBMS or SQL database, a single server, and a true analytics solution. I will find those gaps in care, deliver the results to case and care managers, and actually enhance my population outcomes. All of this will be accomplished without having to super-size my analytics meal!
“Large collections of electronic patient records have long provided abundant, but under-explored information on the real-world use of medicines. But when used properly these records can provide longitudinal observational data which is perfect for data mining,” Duan said. “Although such records are maintained for patient administration, they could provide a broad range of clinical information for data analysis. A growing interest has been drug safety.”
In this paper, the researchers proposed two novel algorithms—a likelihood ratio model and a Bayesian network model—for adverse drug effect discovery. Although the performance of these two algorithms is comparable to the state-of-the-art algorithm, Bayesian confidence propagation neural network, by combining three works, the researchers say one can get better, more diverse results.
I saw this a few weeks ago, and while I haven't had the time to delve deep into the details of this particular advance, it did at least give me more reason for hope with respect to the big picture of which it is a part.
It brought to mind the controversy over Vioxx starting a dozen or so years ago, documented in a 2004 article in the Cleveland Clinic Journal of Medicine. Vioxx, released in 1999, was a godsend to patients suffering from rheumatoid arthritic pain, but a longitudinal study published in 2000 unexpectedly showed a higher incidence of myocardial infarctions among Vioxx users compared with the former standard-of-care drug, naproxen. Merck, the patent holder, responded that the difference was due to a "protective effect" it attributed to naproxen rather than a causative adverse effect of Vioxx.
One of the sources of empirical evidence that eventually discredited Merck's defense of Vioxx's safety was a pioneering data mining epidemiological study conducted by Graham et al. using the live electronic medical records of 1.4 million Kaiser Permanente of California patients. Their findings were presented first in a poster in 2004 and then in the Lancet in 2005. Two or three other contemporaneous epidemiological studies of smaller non-overlapping populations showed similar results. A rigorous 18-month prospective study of the efficacy of Vioxx's generic form in relieving colon polyps showed an "unanticipated" significant increase in heart attacks among study participants.
Merck's withdrawal of Vioxx was an early victory for Big Data, though it did not win the battle alone. What the controversy did do was demonstrate the power of data mining in live electronic medical records. Graham and his colleagues were able to retrospectively construct what was effectively a clinical trial based on over 2 million patient-years of data. The fact that EMR records are not as rigorously accurate as clinical trial data capture was rendered moot by the huge volume of data analyzed.
Today, the value of Big Data in epidemiology is unquestioned, and the current focus is on developing better analytics and in parallel addressing concerns about patient privacy. The HITECH Act and Obamacare are increasing the rate of electronic biomedical data capture, and improving the utility of such data by requiring the adoption of standardized data structures and controlled vocabularies.
We are witnessing the dawning of an era, and hopefully the start of the transformation of our broken healthcare system into a learning organization.
I believe if we reduce the time between intention and action, it causes a major change in what you can do, period. When you actually get it down to two seconds, it’s a different way of thinking, and that’s powerful. And so I believe, and this is what a lot of people believe in academia right now, that these on-body devices are really the next revolution in computing.
I am convinced that wearable devices, in particular heads-up devices of which Google Glass is an example, will be playing a major role in medical practice in the not-too-distant future. The above quote from Thad Starner describes the leverage point such devices will exploit: the gap that now exists between deciding to make use of a device and being able to carry out the intended action.
Right now it takes me between 15 and 30 seconds to get my iPhone out and do something useful with it. Even in its current primitive form, Google Glass can do at least some of the most common tasks for which I get out my iPhone in under five seconds, such as taking a snapshot or doing a Web search.
Closing the gap between intention and action will open up potential computing modalities that do not currently exist, entirely novel use case scenarios that are difficult even to envision before a critical mass of early adopter experience is achieved.
The Technology Review interview from which I extracted the quote raises some of the potential issues wearable tech needs to address, but the value proposition driving adoption will soon be truly compelling.
I'm adding some drill-down links below.
Practices tended to use few formal mechanisms, such as formal care teams and designated care or case managers, but there was considerable evidence of use of informal team-based care and care coordination nonetheless. It appears that many of these practices achieved the spirit, if not the letter, of the law in terms of key dimensions of PCMH.
One bit of good news about the Patient Centered Medical Home (PCMH) model: here is a study showing that in spite of considerable challenges to PCMH implementation, the transformations it embodies can be and are being implemented even in small primary care practices serving disadvantaged populations.
It's years ago, but I still remember the reception I received when during my first large-scale NHS IT implementation I suggested that doctors might like to record outcome information. I can still recall the smell of my singed fingertips. Until recently, the NHS has been obsessed with recording process data fitting an organization with its roots still in the mid-twentieth century.
I have this guy I hired about a year ago named Max. Max is awesome not only because of the work that he does on our marketing team, but also for things like his family (his mom is the super-woman Judy whom I love), his bow-ties, his ability to handle strange travel delays and his stories and wise sayings. In honor of Max, and to make up for, in some small way, that his whole family is in London and he had to stay home because he has a JOB, this is a blog post about one of those bits of wisdom. Somehow his single sentences hold clarity and meaning in a way that only Max can convey. Why is it number 501? Because it can’t be his first (it's way to insightful) and I know it is one of many.
Recently we were working with a client and we had trouble getting them to be in the present. There was all this talk and focus around what their system USED to do. No one wanted to think about what it was STILL doing or needed to do in the future. No one was able to see that the goals of the retiring system were different now that it has been replaced by Epic. It was quite frustrating.
Then Max says:
it's like looking in the rear view mirror to drive forward
And that is EXACTLY what it was. When you put it like that, everyone lights up because THEY GET IT. While looking in the rear view adds value, and lets you reflect on where you have been (which is important), you CANNOT look in the rear view to drive forward. You have to be looking at the front window in order to move on and get to where you eventually need to be. You HAVE to face the reality of the road in front of you. You have to remember that you have an ultimate destination and place you need to be. Everyone needs to put away the map of what the old system used to do and get out the Archive Strategy Map. It can hold some of the same sights, back roads, and detours of the original map, but it needs to be one that is driving your archive strategy forward.
Portuguese and Spanish researchers in the field of social robotics are working on the use of robots to interact with children who are hospitalized for the treatment of cancer, thereby providing emotional support.
The researchers are keen to take robots out of the laboratory and place them in a real environment. Until now, most of the research on social robotics has taken place in very controlled environments. As Professor Salichs from UC3M points out, 'The introduction of a group of autonomous social robots into surroundings with these characteristics is something new, and we hope that the project will help us to advance in the development of robots that are able to relate to people in complex situations and scenarios.'
Another cause for guarded optimism about health care robotics? My hope is that it will augment the efforts of often overworked staff and allow them to better prioritize the focus of their precious attention and energy. In addition to their potential social value, robots could act as in situ surveillance devices to watch for nascent or emergent health crises. My fear is that they will be used as justification for cutting costs through staff reductions, as self-checkout lanes have done in supermarkets.
If you are reading my blog, you know what my company does. Being the guys that are handling the “go dead” of an application, we are dealing with what is, agreeably, the most difficult and disappointing time for the vendor being retired.
I get it. The customers have passed you over. Putting you out to pasture. Picked someone younger, fresher, nicer, better suited to their business. Your revenues are impacted. Your market position is threatened. It makes a statement about where you are in your product lifecycle.
What I DON’T get is the attitude that comes along with it. Most HIT vendors today sell more than one product for more than one solution. Why would you treat a customer badly just because they are choosing to de-install ONE of your products at their work site? Raising their support fees? Refusing to help them? NOT GIVING THEM THEIR DATA? I just don’t get it. Its like you are 12 years old and breaking up with your first girlfriend.
Do you actually think this is the right way to treat these customers? That this WON’T come back to haunt you later? Do you know how much these customers talk to each other? Have you READ HISTalk?
Think about the bigger picture here. If you are a vendor that plans to stay in business, you had better treat each and every customer or potential customer like you would want to be treated. The good old Golden Rule applies here.
The decisions have been made. You were or were not asked to present or even attend the party. Suck it up. Do the right thing. Help your customer transition to the other vendor in a positive and supportive way. Make their data accessible and understandable. Help them through technical issues. Use the opportunity to find out what you can do differently next time to be the vendor that is chosen. MAYBE, just maybe, you will get a chance to present or sell something to this customer again. And MAYBE, just maybe, they will remember the positive and professional way you handled the prior system “end of life” and take that into consideration.
Because I can promise you this – the way some of these vendors are acting – they won’t even get in the parking lot, much less a ticket to the dance, even if they are the very best at what they do or sell. They are not only burning those bridges, they are nuking them.
Robot Aids in Therapy for Autistic Children
Wall Street Journal (05/01/13) Shirley S. Wang
University of Notre Dame researchers will present study findings at the annual conference of the International Society for Autism Research showing promise in the use of robots for teaching social skills to autistic children. The study, involving 19 autistic children, is believed to be the largest trial to date using robots in this way. The children interacted with a two-foot-tall robot therapist that was programmed to ask novel questions and engage children in conversation. The study participants showed greater conversational improvement with the robot than with a human therapist alone, and parents reported more significant improvement at home as well. Children interacted in six sessions with the robot as well as with a human therapist, who provided instruction on specific skills when interacting with the robot, such as making eye contact or taking turns talking. Simplified social interactions with a robot might be beneficial to children with autism, who tend to be very interested in technology but find complex social interactions challenging. The researchers hope the children will carry over the social skills to interactions with people as well, rather than just interacting with the robot.
Monday's ACM TechNews produced this very brief but tantalizing summary of a Wall Street Journal article.
This is one of those stories that leave me very ambivalent. In some ways, my automatic reaction to our collective desire to depend more on automation in direct patient care is fear. I am afraid we are going to abandon our elderly and otherwise hopelessly disabled kin to the unfeeling arms of robots, androids, whatever. This will spare us the feelings aroused by an out-of-control psychotic spouse, an incontinent and demented parent, or a profoundly developmentally disabled child, when we must intervene and our interventions are resisted, not appreciated, or insufficiently effective.
With this story, I see the situation is not so simple. Autistic children have difficulty relating to humans with whom they are intimately involved, and their difficulties are often reflected in others' responses to them. Machines are insensitive by nature, and can be programmed to reward positive behavior and ignore the negatives. This may be a situation, as the investigators assert, where robotic intervention is not only appropriate as an alternative but can even improve the patient's situation holistically.
I don't have a WSJ subscription so I can't follow the link ACM provides to the full story, and I don't have time at the moment to poke around on the Web for alternate sources of information about this research project. I would like to learn more, and will try to pursue this when I have more time.
1. I am loyal to a fault. You want me on your side. Even if you are wrong.
2. Don’t Wrong Me. That loyalty thing being said, don’t do me wrong. You don’t want to deal with that.
3. I take it to heart. All of it. I have a tremendous sense of right and wrong, and if I have Wronged You, I am devastated. I will do anything to make it up to you. If I am Right, you better be ready to convince me why you might be too.
4. That being said, don’t take advantage. See #2.
5. You can trust me. But be able to handle the truth, because I am going to give it to you.
6. I love my dog. She is a Cavalier king Charles Spaniel and she is fat and I love her.
7. I love my kids. More than my dog, most of the time.
8. I love life and adventure and change and the next thing. I am always looking to learn more, do more, be better. And there is nothing wrong with that.
9. I worry. All the time. You can stop worrying, I have you covered.
10. I love Prosecco.
A significant portion of the physician market has still not adopted an EHR, despite the lure of government incentives and the fear of the penalties looming on the horizon. The stock prices of most publicly traded ambulatory EHR companies are down sharply, as sales are lower and earnings projections have not been met throughout the industry. How can this be, when the EHR incentive program has successfully increased EHR adoption and was expected to be such a boon to EHR vendors?
I know why, and it is not—as commonly thought—because the initial EHR-adoption rush fostered by the incentives has ended. Rather, it is because of rampant physician dissatisfaction that has reached a more-than-palpable level. I have noticed a dramatic change in the tenor of conversations with physicians, most recently at professional society conferences, where physicians who have not yet purchased an EHR are frozen in their tracks. They are worried by the horror stories they hear from colleagues—even from those who have succeeded at meaningful use—because many of those physicians continue to experience major workflow disruptions and significant productivity losses from which they see no potential to rebound. Recent surveys point to the number of physicians looking to replace their EHRs, and based on my company’s experience in the replacement market, that number is growing. A recent article summarized the findings of a large study on EHR satisfaction and presented an insightful analysis of the potential reasons for these disappointing results.
This heightened level of frustration has resulted from frantic, insufficiently researched EHR purchase decisions by physicians and rushed, inadequate implementations conducted by resource-strapped vendors. Massive EHR failures are exactly what I predicted in an EMR Straight Talk post on the unintended consequences of the EHR incentive program in February 2010:
After an initial peak in implementations, long-term EHR adoption will slow—particularly among high-performance specialists—and the current failure rate will escalate. Many factors will contribute to this: (1) Some physicians will rush into EHR purchases without conducting proper due diligence. (2) Products that were overly complex and did not work in busy specialists’ practices in the past will surely not succeed now, particularly since these same products must now be used in an even more structured and demanding way. (3) Sorely needed implementation and training will be provided by inexperienced and rushed implementation teams, further reducing the likelihood of success with providers, many of whom are less technologically savvy than the early adopters. (4) Where there was never a convincing economic justification in the past, the addition of data-collection requirements will further lessen the economic feasibility of traditional, point-and-click EHRs. . . . The result? The high failure rate will leave physicians “holding the bag” after investing large sums of money, failing to earn the anticipated incentives, and owning a system that doesn’t meet their needs.
So, what can physicians do to avoid falling victim to EHR failure, and to instead reap the benefits of successful EHR adoption—government incentives and practice productivity? I have written extensively about the importance of physicians doing thorough and objective reference checking—that advice is as valid now as when I first wrote about it, and perhaps is even more critical today. For guidance on how to conduct a thorough and fair evaluation of an EHR, read EMR Selection: How to Uncover the Truth or 100% EHR Success – A Clinical Approach.
Well I have done it. My book Stop Saving the NHS and Start Reinventing it has been published in Kindle and paperback. It's aimed at NHS leaders and managers, but will probably interest anyone who is interested in the shape of 21st century healthcare.
The FAQ related to the Implementation of EN 62304 with respect to MDD 93/42/EEC was released by Team NB, the association of Notified Bodies.
You'll find in this FAQ many hot subjects I already mentioned in this blog:
This FAQ shows that the state-of-the-art is still evolving. But I think that it has reached a point of consistency and stability. Many questions in the FAQ hadn't clear answers one or two years back.
This week, six senators released a white paper, Reboot: Re-examining the Strategies Needed to Successfully Adopt Health IT, that argues that there is no evidence that the $32 billion in taxpayers’ money being spent on meaningful use is returning the results it was designed to deliver. Although it would be naïve to discount the political motivation of the authors—all six being Republicans—they raise some of the same criticisms and concerns that I have written about in the past. They also make some claims that I feel compelled to dispute.
The senators have it right on these issues:
I vociferously disagree, however, with the senators’ criticism regarding interoperability. Of course, we are not there yet—and clearly they are frustrated by that fact—but progress is underway toward that universally supported goal. Contrary to their claim that there are no meaningful use measures that require interoperability, there are in fact several in Stage 2, including the requirement that physicians electronically send a patient care summary for 10% of patients transitioned to the care of another physician or provider. This exchange is facilitated by the fact that all certified EHRs must communicate using the same formats.
Not only does interoperability relate to provider-to-provider communication, but it also allows for easy integration between products of different vendors, without requiring additional programming. I was recently speaking with another HIT vendor about a potential partnership arrangement, and we both talked the same language—XDR Direct for transport protocol and CCDA or HL7 in terms of content. This conversation would neither have been possible, nor would we be able to create a tight, simple interface between our products, were it not for the standards promulgated by the EHR incentive program. This kind of interoperability will ultimately be better for physicians and for their patients. The EHRA (EHR vendor association of HIMSS) hit the nail on the head: the appropriate role for government is to set the standards, but then the vendors should be free to innovate and let the market take over from there.