Physician Assisted Suicide: The Need for a Registry with Ongoing "Big Data" Surveillance

The means for escape......

While the Disease Management Care Blog was pondering a Final Exit methodology to escape the glut of repugnant political ads, it came across this "Not So Fast" editorial in the Halloween New York Times.  Author Ben Matlin gives pause to the notion that death with dignity is an open and shut issue. "Thin and porous" is how Mr. Matlin describes the border between coercion and choice.

In the meantime, policy is dominated by articles like this that reassure readers that physician-assisted suicide will remain reasonably appropriate, palliative care will be available to those who choose to struggle on, "vulnerable" patients will not be disproportionately impacted and that the slippery slope that leads to euthanasia will be prevented by solid safeguards. Ethics and organized medical opposition aside, the argument is that all that remains is the need for a "federal" mechanism with safeguards, transparency and non-physicians.

The DMCB has idea none where this will lead.  It suggests, however, given the significant estate tax considerations that surround the timing of a death, that this is an issue that calls for a registry with ongoing "Big Data" style surveillance.  If nationwide trends show a significant association between assisted suicide events and the various and shifting local, state or federal tax deadlines, we'll know that we've made a terrible mistake.

Image from Wikepedia

Associations. Correlations. Inferences. Signals. Yes, That's Big Data

America's corporate Directors
celebrate big data

The Population Health Blog's recent travels recently included a speaking gig at the just concluded National Association of Corporate Directors ("NACD") annual conference meeting.  It was part of a panel discussion focusing on health care innovation that was ably moderated by tech guru John Hotta.

The PHB's educational mission was to enable the persons who serve on Boards of Directors understand how "big data" is going to change health care.  After giving its standard definition (the use of large, disparate and unrelated data sets to find correlations and draw inferences that are actionable at the individual level), it turned to the following example:

"Imagine standing at the top of the Empire State Building and analyzing the noise from below to find out what's most likely happening down on Fifth Avenue."

In other words, its the use of computational analytics to separate the noise from the signals, and using those signals to ascertain a probability.

An informed guess.  Or, a probabilistic choice.

Folks in the audience seemed to get it, especially when the PHB noted that insurance (ICD-9 250), electronic record ("diabetes") pharmacy (insulin), public health (obesity prevalence data by zip code), survey ("have you ever been told you have diabetes?"), government (car registration; overweight persons prefer minivans), web-usage (recent interest in low calorie foods?) and purchasing (grocery purchases) data could be marshalled to assign a risk that diabetes is present, and if it's present, the risk of complications, and if there is a high risk, whether it's actionable.

The value proposition? 

By understanding the risk and being able to array it from high to low, precious health care resources can be scaled to the burden of illness in the population.  So, instead of "carpet bombing" all persons with a diagnosis of diabetes with one-size-fits-all reminders to see their doctor along with mass mailings of educational materials, personalized outreach can be targeted on those persons most likely to be hospitalized (and there are big data signals that can predict it) in the next year.

Bottom line: it can save money by rationalizing health care.

The PHB wanted to point out some other need-to-knows, which it did with variable success:

1. Quantum jumps in processing power and server capacity have put this within reach of desk-top personal computers.  As an added bonus, you don't need an army of mathematicians.

2. "Actionable" also means that the information is meaningfully available at the point of care, i.e. in the doctor's office where 80% of the decisions that drive health care spending occur.

3. Big data can also point to way toward more accurate diagnoses (imagine if all the risk factors for an Ebola infection had been rolled up into a single score in that Texas ER) as well as treatment (deciding on the "best" cancer treatment program after knowing the relative influences of genetics, lifestyle and past medical history).

New Insights on Big Data

Say hello to big data!

This week's The Economist discusses some of the implications of "big data."

Population Health Blog readers may recall the Facebook kerfuffle when it was revealed that the company had been "experimenting" on its users.  While this is another timely reminder that Facebook users are not customers but a product, what's far more interesting is how the company used its vast utilization data to monetize the changes of a hundredth of single percent in its users' behavior. 

These weenie changes can add up to additional ad revenue.  Ditto for the techy Google's search, Amazon's placements and Facebook's ads which, according to Schumpeter, are getting a return on investment from "every pixel" of the monitor screen you're using to read this blog.

Unfortunately, says The Economist, this big data approach has been more difficult for the for the traditional "bricks and mortar" businesses, which have traditionally been fixated on traditional accounting measures.  But, as these businesses tether their infrastructure increasingly to information technology, they're getting there: UPS is monitoring 60,000 delivery vans, retailers are assessing how in-store placements generate the most revenue and businesses are measuring how the mix of different types of employees relates to productivity.

The PHB would rate most health care providers as being in the bricks and mortar category than in the technology space.  They have a way to go.

Two lessons:

1) Big data doesn't replace traditional business monitoring of revenue and expense, balance sheets, averages and standard deviations, it adds to it.  And yes, that ceaseless tinkering adds cost that - in the right hands - should have a return on investment.

2) The "tinkering" has its share of failures in addition to successes.  While The Economist can point to some wins, the landscape is probably littered with losses.

Speaking of the right hands and minimizing losses from big data, the PHB is proud to link this article appearing in the NACD Directorship Magazine.  While a password is necessary, the bottom line is that corporate boards can do ten things to help their companies successfully achieve a big data return on investment:

1. Provide analytic leeway;

2. Be clear on who is responsible;

3. Set realistic budgets;

4. Assess how big data is fitting into the pre-existing culture;

5. Be skeptical;

6. Link it to Enterprise Risk Management;

7. Task the Audit Committee with some oversight;

8. Assure privacy;

9. Reduce inappropriate incentives;

10. Get a big data expert on the Board.

Image from Wikipedia

More Big Insights on Big Data

Given the data, what are her chances of
getting breast cancer?

Unable to sate its big appetite for big data insights, the Population Health Blog glommed onto the New England Journal's just-published article on "Learning from Big Data."

As noted previously, "big data" is the use of statistical associations ("predictors") in a) large and b) disparate data sets  to gain insights at the individual level ("outcomes"). For example, a physician could know the likelihood - based on demographic, clinical and economic inputs - that a particular patient won't fill a prescription. As an other example, the PHB spouse could know the likelihood - based on prior active-passive behaviors, incentives and maternal upbringing - the likelihood, despite numerous reminders, that her husband will "forget" to take out the trash.

It's important to recall that big data is not about causality. Just because living in a certain zip code is an independent predictor of obesity (for example) doesn't mean living in [insert name of town] causes residents to be fat. Big data is "agnostic" about the cause, but that doesn't mean Big Data Architects (BDAs) can't use the information.

According to the author, the road from the promise to the reality of big data will be lined with:

1. generalizability, or being confident that the populations used in big data studies are similar to the populations where their lessons are being applied. Propensity matching or scoring is a good step in that direction;

2. automation, so that multiple questions can be answered simultaneously by many users;

3. "data refreshes," so that associations can be retested on repeated basis as new data come on line;

4. ease-of-use, so that even an orthopedist could use the software and understand the outputs.*

Politically, we'll also need to get

5. the owners of data warehouses - including the electronic health record vendors and insurers - to agree on either a) common data formats or b) methods that allow for the interpretation of data regardless of the format. An example of the latter the use of an order, entry or insurance claim for supplemental oxygen therapy as a marker of poor health status.

6) a resolution of our absolutist privacy "impasse.""De-identification" of patients' information makes it possible, but never guaranteed, to keep personal health information secure.

*okay, the New England Journal author didn't poke fun at the orthopedists by saying that, but the PHB couldn't resist. By the way, one way to do this would be to have the outputs be in pictures.

Image from Wikipedia

Big Data, Definitions and Population Health

What's the likelihood of diabetes?

Utter the term "big data" at any ACO, care management or managed care meeting, and one of two things will happen:

1) Your colleagues will admire your population health chops and your boss will be reminded that you deserve a raise, or

2) Your colleagues will tire of your faddism and your boss will wonder, once again, just what "big data" means

Either way, you may want to refer your colleagues and boss to this readable "on-line first" article appearing in JAMA.

Here's a handy PHB summary:

"Big data" can be defined as the linking of disparate large data sets to provide insight at the individual level.

It's been used by political campaigns (swing voters), business (expectant mothers) and the NSA (potential terrorists). Once they are identified, amenable voters can be individually lobbied, expectant mothers can be sent personalized coupons and evil-doers can be visited by Jack Bauer.

According to Weber and his co-authors, how should health care providers approach big data?

1) Inventory the available data sets.  Traditional examples include electronic health records, insurance claims and pharmacy data.  Big data architects should also be aware of non-traditional examples including social media, census records and credit card purchases (such as grocery store purchases, fitness club memberships or over-the-counter meds).

2) Anticipate "probabilistic matching," since two or more individuals may fulfill criteria.  This will involve trade-offs between accuracy and feasibility, since two individuals matching "John Smith" in a single zip code may appear to have the same risk. 

3) Worry about HIPAA. Unfortunately, while medical data sets are disparate, they're also walled off by privacy concerns and special regulations that govern genetic and mental health data. It's not insurmountable. The health care industry should also participate in the public square to and help shape evolving societal and legislative standards over privacy.

Fortunately, the population health industry (here's a modest example) is already engaged. They understand that big data can be used to estimate individual risk which can, in turn, guide outreach to individual patients.

Image from Wikipedia

Big Data Analytics Salesmanship: A Free Public Service for the Data Management Vendors

Data analytics for sale!

While the Disease Management Care Blog continues to delight in the clever electronic health record humor of the Extormity web site, little did it know that the same over-promising jargon typical of the EHR hucksters would be adopted by the "Big Data" health care analytics vendors.

Sensing that the hard-sell carpet bombing of the provider community from these outfits is only going to increase, the Disease Management Care Blog is pleased to offer the emerging community of data analytics vendors some cut n' paste bombast suitable for their press releases, web sites, trade show presentations and glossy collaterals:

(Insert name of your company here)'s business intelligence and work-flow solutions offer best-in-breed data analytics that both monetizes and obfuscates open-source programming.  As end-utilizers utilize any Windows 8-based interface to satisfy myriad and shifting passkey screens to eventually tap our physician-led and lawyer-vetted intelligence applications, a robust informatics ecosystem awaits, all in a series of hosted legally immunized informatics platforms located in India or Kuala Lumpur and Vladivostok.

Providers, administrators and auditors can then utilize (company name's) propriety specifications developed by dozens of PhDs to assess trending, risks, outcomes, costs, spending, incurred-but-not-reported, billed-but-not-paid, correlations, odds ratios, attributable risk, Z statistics and regressions for pre-packaged insurance claims-based definitions of diabetes mellitus, migraine, gout, dropsy, lumbago, and conkus of the bonkus.

(Insert name of company) can also access Centers for Medicare and Medicaid Services claims data and use benchmark advances to apply truly artificial intelligence to derive predictive assessments that can also predict your health system's business planning needs. Claims data from some commercial carriers are also suitable to import and derivation thanks to the availability of an upgrade plug-in Ver 2.3.1.1 licensing option.

"Our company's approach to analytics is truly unique," said CEO Vera Patented, adding "We seek a long-term relationship with our partners and aspire to leverage our proprietary technology to assure that's precisely happens."

"Our technology is compliant with all known, anticipated and over-interpreted federal statutes and regulations," added Chief Legal Officer Doan Suemoi.  "Our ironclad hold harmless contracts guarantees our protection from allegations related to HIPAA, HITECH, the ACA, the GNA, the FDA, the FAA and the NRA," she added.

Remarked customer and Chief Imagination Officer Dr. Mustafa Raturn, "As an early adopter of this technology, I was impressed by this product's ability to arbitrarily assign risk scores to two decimal places for populations using a methodology backed up by years of research. This functionality was paired with a cut-and-pace interface with bar graphs and pie charts configured to impress my hospital's Board of Trustees." 

About the company:

(Name of company) is a wholly owned subsidiary of Triple Aim Fail Ventures, a company that offers innovative, integrated and illusionary services designed to meet the full spectrum of analytic, management, consulting and predatory health care business needs.

For-Profit Meets For-Publication For Big Data

The "for-profit" research side of health care and the academics have always had a strained relationship. The Population Health Blog witnessed it first-hand when it recruited volunteer participants for an protocol that was sponsored by a pharmaceutical company. It was a good experience, but the company made it abundantly clear who was in charge of the data.

As "big data" research grows, will large pieces of it be likewise run by self-serving and deep-pocketed healthcare corporations?

That's the question explored in this JAMA "online first" piece by Sachin Jain et al. Huge electronic health record and insurance claims data sets involving tens of thousands of patients can provide academically (publishable) as well as commercially (profitable) insights on treatment safety and effectiveness in the real world. The JAMA authors use Indiana School of Medicine's Regenstrief Institute's collaboration with pharma giant Merck as an example of how the relationship doesn't have to be anything but collaborative.

Their 5 year agreement centers on mining a statewide information exchange involving over 11 million patients. Scientists from both companies with similar interests - such as melanoma, heart disease in persons with diabetes, medication adherence, the progression of heart failure, treatment of osteoporosis, natural language processing and vaccinations - are encouraged to jointly present ideas to a steering committee that ultimately okays and funds projects.

What are some of the lessons learned?

1. Academics prefer funding that lasts 12 to 18 months, while pharma wants an answer ASAP. The fix was to create sustainable funding "cycles."

2. Protection of individually identifiable data is a priority; Merck has "arms length" access only to de-identified data, and that's just for starters.

3. Both institutions have to agree on the release of any research findings into the public domain.  Any disagreements are handled by the steering committee.

4. A separate operations committee keeps track of all the projects and their timelines.

5. Some research questions on the natural progression of chronic disease can only be answered over the course of years.  One big data project beats a gold-standard randomized clinical trial.

The PHB's take:

This may be a template for population health vendors to follow.

Because they're interested in the association of multiple risk factors with multiple outcomes, the vendors likewise have a lot to gain from mining big data. The good news is that many already have contracts with health care systems and other entities that are sitting on terrabytes of clinical and claims data. Smart vendors should be asking how to move past their for-profit reputation, leverage these relationships and take big data - with their academic colleagues - to the next level.

Image from Wikipedia

Big Data and the Coming New Value Proposition for Disease, Care and Wellness Management Providers

Disease Management Care Blog readers know that the its latest interest is "Big Data." While the researcher-DMCB has played in the sandbox of some insurance claims data sets, the idea of combining and combing through multiple terrabytes of clinical and public data remains a topic of endless fascination. It knows it's not alone.

So, it was only a matter of time until one of the major clinical journals published an article on the topic. JAMA has stepped forward, and not a moment too soon.

It's "must reading" for the disease and care management provider community.

Drs. Murdoch and Detsky point out that Big Data offers four value propositions:

1. Observational correlations may generate insights that cannot be found using standard research approaches. Scanning text for key words in electronic record systems involving hundreds of thousands of patients may find associations or trigger early warnings faster, quicker and cheaper than any formal scientific protocol or clinical trial.

2. Those insights, especially since they can be tailored to fit the circumstances of an otherwise unique patient, can be used to guide diagnosis or treatment. Physician judgement cannot be replaced, but if Big Data points out that there were other patients with a similar pattern of illness who responded best to one treatment versus another, patient outcomes could improve.

3. A Big Data approach to genomics can correlate genetic information with outcomes and further guide therapy. While the DMCB still wonders if "genomics," outside some narrow anecdotes, will always remain the science of the future, Big Data may turn out to be the key to finally unlocking its potential.

4. Since Big Data, by its very nature, can combine clinical information to other personal data (the foods you've bought or your driving history), Big Data will necessarily tilt toward the patient-consumer and away from the health care system. Not only does permission for access lie with the patient, but the insights will be less about sickness and more about wellness.

The authors do a good job of pointing out that there are plenty of challenges. Most doctors don't get it, privacy laws could be over-interpreted or enforced, it remains to be seen who will pay for it and Big Data is still in its infancy.  The DMCB also points out that while Medicare has just discovered that alternative research innovations are possible, Big Data promises to eclipse those approaches (like traditional time series analysis, propensity matching), again making CMS a day late and another dollar over budget.

The implications for the care management and population health community are considerable. The industry has amassed years of intellectual capital in the science of predictive modeling and Big Data is it's next step. Many care management vendors have multiple clinical partners and already have access to terrabytes of data involving millions of persons. Not only is the math and the informatics well within reach, they also "get" the tilt toward wellness and consumer empowerment. Last but not least, if anyone can monetize a value proposition like this and turn insights into revenue (or "shared savings"), these nimble vendors can.

A DMCB prediction: while academics will write about Big Data in scientific journals, the care management industry will be doing it.  In fact, they probably already are.

Two particularly good quotes to use to impress your CEO and stymie your competitors:

"Data has gone from refuse to riches."

and

Economic theory describes the quantitative conversion of 3 kinds of inputs (capital, labor, and raw materials) into outputs (goods and services)...The current revolution in data management makes it clear that a fourth kind of input, information, will become just as important as these other inputs in the future of many industries.

A Definition of "Big Data" for Health Care Providers and Five Useful Caveats

And you thought its only
function was to be an EHR?

Regular readers of the Wall Street Journal probably saw the Monday March 11 "big data" section that was filled with articles like this.

Written from a "business intelligence" perspective, there were precious few insights for the population and care management community. We're aware of the concept, but how, asks the Disease Management Care Blog, does it apply to our corner of the health care delivery system?

Unable to resist, the DMCB donned its snorkle and flippers and took a deep dive at the topic.

First off, when the DMCB performed a classic medical literature search using the key words "big data," it found that that the term has not entered the health care lexicon in a big way. Academics instead prefer to write about "registries," "data warehousing" and "predictive modeling." The DMCB also looked for a standard health care definition of "big data" and could find none in the published medical literature.

So, the DMCB offers up its own definition, culled from papers like this and this:

Health care "big data" is a branch of health care informatics that pools large and disparate data sets and applies a suite of mathematical approaches that derives associations, facilitates comparisons and generates insights that are otherwise not possible using standard mono-source analytics. It includes, but is not limited to, reporting, dashboards, ad-hoc queries, graphical displays, scorecards, predictive modeling, data mining and business intelligence. The data sets can be comprised of EHR data, insurance claims, pharmacy utilization, care management systems, consumer as well as government information, public health, surveys, point-of-contact information and web-usage.

The DMCB's simplistic off-the-cuff examples of big data queries include examining 1) the association between "hits" from a cluster of ISPs on an emergency room's web page and ER utilization, 2) complaints about a hospital's food service from family and the likelihood of being named in a malpractice suit, 3) looking for rare side effects among persons with a cluster of medical diagnoses who are using a just-released drug and 4) whether the number of household flat screens is a useful predictor of obesity.

Five DMCB caveats:

1. One data integrity trumps five Ph.Ds: The chief challenge is not the mathematics but combining and aligning the various databases.  Once the information is teed up, it's amazing how much can be done by a masters-level statistician and a desktop PC.

2. Associations, not causality. Whether a web page leads to ER visits or whether bad food fuels dissatisfaction is a different question.  It's possible that ER visits prompt web usage or that already dissatisfied patients find overcooked string beans icky. All the possibilities are still useful insights.

3. Not a panacea: It's "a" tool, not "the" tool.  Users will still need to also invest in faster, better and cheaper mundane data tasks (like admissions per thousand) while they simultaneously understand how big data's associations, comparisons and insights generate additional patient value.

4. Journey, not destination: There's a potent mix of art, science and wizardry in the evolving science of "big data."  There are no standard methodologies or best practices.  Get used to it.

5. Skepticism abounds: Data stakeholders who are used to standard analytics will refer, as the DMCB found out, to "big data" as "voodoo," and resist buy-in.  If a critical mass of an organization's leadership comes to believe it's useful, the rest will follow... eventually.

Image from Wikipedia