Have You Started Your Data Expedition Yet?

In 1803, Thomas Jefferson sent Meriwether Lewis and William Clark on their now famous expedition. The initial goal was to find a water-based route to the Pacific Ocean in addition to exploring the unmapped West. They imagined they’d find woolly mammoths, mountains of pure salt, lava-spewing volcanoes and never before seen creatures. What they found was quite different. Why did they even risk life and limb to do this in the first place?

It turns out that Thomas Jefferson was a visionary and a bit of an intrapreneur. When Jefferson took office in 1801, one of his top priorities was to gain control of the port of New Orleans. He saw this important water access point as an enabler of economic growth for farming. In one move, he more than doubled the size of the country for what turned out to be the real estate deal of all time … at the bargain price of less than three cents an acre. He paid $11.25 million in 1803 or roughly $234 billion in today’s dollars. That investment has paid for itself in incalculable terms in 1803. Of note, part or all of 15 states were created from this transaction. Ironically, Jefferson’s desire to control New Orleans what is the motive for the deal and the rest of the territory was pretty much a throw-in. This transaction is probably his greatest legacy and was arguably the most important step taken to build The USA into what it is today.

What would you do if you found yourself sitting on a massive untapped asset … and no one knew what it contained … or what to really do with it?

Jefferson decided to conduct an expedition of discovery by creating the Corps of Discovery in 1804 to explore his new acquisition.

Ironically, if you are a healthcare provider, have the same opportunity. Your big data is your untapped asset. What are you doing to explore, understand and leverage it? Leading providers are already conducting data expeditions of discovery. Most importantly, they are creating new streams of revenue from what they are learning.

Jefferson turned to trusted allies .. James Monroe to negotiate the Louisiana Purchase and Meriwether Lewis (who added William Clark) to lead the expedition of the acquired territory. Along the journey, other team members joined them (Sacagawea and Touissant Charbonneau).

You will need trusted allies and partner also for your expedition. You also need to think outside the box. In order to take advantage of new insights, it will require new thinking and new business models.

The Corps of Discovery faced nearly every obstacle and hardship imaginable on their trip. They braved dangerous waters and harsh weather and endured hunger, illness, injury, and fatigue. Along the way, Lewis kept a detailed journal and collected samples of plants and animals he encountered. It’s no wonder that it became known as the wild wild West.

You may not face hunger, illness and injury on your data expedition journey (insert IT joke here) but you will need the right kind of tools to prosper from it. Most importantly, navigating your way through big data will require advanced technologies … especially since more than 80% of it is unstructured in nature.

What the Corps of Discovery found was mind-boggling … some 300 species unknown to science, nearly 50 Indian tribes, and the Rockies (the mountains, not the baseball team).  They created the foundation and landscape for future governing states and grateful future generations and citizens.

You are almost certain to find many new ideas, insights that are really opportunities waiting for you to identify and exploit them.  Most amazingly, you have the same opportunity to impact the lives of generations of the yet unborn .. just in a different manner.

But unless you act now and start your own expedition … you are going to fall too far behind. Here’s why … new business models and new ways of delivering healthcare are already emerging.

Plus, you might swallowed up by big data bigfoot. A big data bigfoot … really!?!!(yes, I know this is a stretch)

By 2020, the footprint of medical data will double every 73 days according to the University of Iowa, Carver College of Medicine in 2014 (get it … bigfoot … as in footprint).

Seriously, This is going to change everything!

Data is going to become imaginably big. Get ready for a new term … Exogenous Data or data originating from outside (derived externally) … as in data from Apple Watch, Fitbit, medical devices, smartphones … you get the idea.

Exogenous Person The future of health is all about the individual and having a complete picture of the many factors that affect a person’s health. But we need better ways to tap into and analyze health information in real time to help doctors, researchers, insurers, case workers and other stakeholders determine the best approaches, and give the patient greater control over his or her own care. This is where you fit in. Bring your own data and take advantage of this now before others do. Start your data expedition here!

Like every other industry, healthcare is being disrupted and transformed by the exponential growth in data, such as medical records and clinical research, and these growing pools of information are difficult to share because they are fragmented. In addition, they do not readily incorporate critical information about individuals’ non-clinical conditions, which may have a strong bearing on health.

As a result, patients and their healthcare providers are forced to make decisions that are not based on all the evidence. And the problem is expected to get worse: between electronic medical records, digitized diagnostics and wearable medical devices, the average person will likely leave a trail of more than 1 million gigabytes of health-related data in his or her lifetime … the equivalent of about 300 million books. Those on data expeditions today will be at the forefront of capitalizing on new business opportunities that come out of all of this transformation.

Advances in data availability, analytics and connectivity are giving doctors, researchers and other health professionals the tools they need to make better, faster and more cost-effective decisions and individuals the insights they need to understand more about their health and receive personalized care. There is a vast amount of meaningful data that can help tell the whole story about an individual’s health and needs. We plan to help stakeholders in the care ecosystem use that foundation to improve the quality, effectiveness and cost of their care.

Have you started your data expedition yet?  If not, what’s holding you back?  Take advantage of your untapped assets before the opportunity disappears.  After all, there are going to be any other Louisiana purchases … cheap land is pretty much gone.

Land is purchased at a premium these days and that model is to your business advantage now. Don’t be one of those laggards who will end up looking for wooly mammoths or mountains of salt when there is so much more out there. Put your data to work for you!

In simple terms — do three things:

  1. Start by taking inventory of the various types of data you have – pay close attention to the unstructured data.
  2. Take another step by forming a team to study how to benefit internally from re-use of that data (better reporting, process improvements, etc.).
  3. Also assemble a group of intrapreneurs to figure our out to leverage that data into new business areas (licensing of data, partnering on IP, etc.).

“I like the dreams of the future better than the history of the past.”   I have to agree with Thomas Jefferson on that.

I will be speaking at a couple of upcoming conferences on these topics and more as I share details from IBM’s data expeditions:

Watson Health builds on IBM’s unique strengths to create the ecosystem needed to transform the global healthcare system, as well as to provide the open, secure and scalable platform of data, insights and solutions needed to make it all possible. Information on IBM Watson Heath can be found here. My recent posts …

As always, please comment below … or feel free to reach out to me directly.

Amputations or Analytics … a Call to Action for Entrepreneurs and Intrapreneurs Alike!

Doctor George Shearer practiced medicine in central Pennsylvania from 1825 to 1878 (in the Dillsburg area). He was a pillar of the community and is believed to have been an active surgeon during the Civil War. He was 61 at the time of the Gettysburg battle.

According to the National Library of Medicine, the exact number is not known, but approximately 60,000 surgeries, about three quarters of all of the operations performed during the Civil War, were amputations. Although seemingly drastic, the operation was intended to prevent deadly complications such as gangrene. There were no anti-biotics during this era.

Back then, amputation was the recommended treatment for major injuries, such as damage from gunshots or cannonballs. These amputations were performed with a handsaw, like the one Doctor Shearer used (shown below). During the war, surgeons prided themselves in the speed at which they could operate, some claiming to be able to remove a leg in under one minute. Ouch! Literally!

Image

(Photo: Doctor George Shearer’s Actual Surgical Kit)

Keep in mind that local anesthetics were not invented until the 1880s and many procedures were performed without ether or chloroform … the only real anesthetics during the era.

In 1861, this was the best standard of care for those injuries. I think we can reasonably conclude that better treatment options (and outcomes) exist today.

Recently, The Mayo Clinic published an eye-opening report entitled, A Decade of Reversal: An Analysis of 146 Contradicted Medical Practices. The report focuses on a published medical practices and how effective they are. Things must have improved since 1861 … right?

The report examines published articles in prominent medical journals of new and established medical practices (such as a treatment guidelines or therapies), over a recent 10 year period (2001-2010). 2044 medical practice articles were reviewed. The findings are fascinating but one section of the report jumped off the page at me. Of the 363 articles that tested an existing standard of care, 40.2% reversed the original standard of care … and only 38.0% reaffirmed the original standard of care. The rest were inconclusive.

In other words, (in this case study) the current published medical standards of care are wrong MORE then often then they are correct. Wow!

I do feel obligated to point out that this is a very limited slice of the overall published standards of care … but still. It is just me … or is this mind-blowing!

I am not talking about gulping down some Jack Daniels so I don’t feel my leg being sawed off. This is researched and tested medical standards of care within the last 13 years. And yet … over 40% of the time, it’s WRONG. In fairness I should point out that they were right 38% of the time. No wonder the US Healthcare system checks in as the 37th best worldwide despite outspending everyone else by a huge margin (per capita).

It’s 150 years later, has the standard of care improved enough? We may not be sawing legs off at the same rate these days, but maybe it’s time for a new approach. Why are other industries so much farther ahead in leveraging their data with analytics to improve quality, reduce costs and improve outcomes? What could be more important then saving life and limb?

Years of data have been piling up in electronic medical records systems. Genomics is not new anymore. Isn’t it about time we brought analytics to this set of opportunities?

Some leading organizations already are … innovative solutions and companies are popping up to meet this opportunity. Entrepreneurs like Scott Megill, co-founder and CEO of Coriell Life Sciences, is a great example. Coriell Life Sciences is an offshoot of the Coriell Institute for Medical Research, a 60-year-old non-profit research organization. In 2007, the Institute launched an effort to bring genomic information to bear on health management. Coriell Life Sciences was established to commercialize the results of that research. Vast amounts of genetic information about individual patients has been available for a number of years, but it has been difficult to get at and expensive. “This company bridges the gap,” said Dr. Michael Christman, the Institute’s CEO.

Coriell’s approach is so innovative, they recently walked away with the coveted “IBM Entrepreneur of the Year” award.

Intrapreneurs at IBM have been busy commercializing the breakthrough innovation, IBM Watson – that originally debuted on Jeopardy! in 2011. Watson is based on a cognitive computing model.

Grabbing a few less headlines is IBM Patient Similarity Analytics, which uses traditional data driven predictive analysis combined with new similarity algorithms and new visualization techniques to identify personalized patient intervention opportunities (that were not previously possible).

These are a couple of obvious examples for me, but in reality we are just at the beginning of leveraging big data. New analytics and visualization tools must become the “handsaw” of today. We need these tools to be at the root of today’s modern standards of care.   If Dr. Shearer were alive today, you can bet his old surgical kit would be on the shelf, having been replaced by analytics that he could bring to the point of care.

For many Entrepreneurs and Intrapreneurs, the journey is just beginning, but there is a long way to go. A 2011 McKinsey report estimated that the healthcare industry can realize as much as $300 billion in annual value through analytics. Yowza!

What are you waiting for?

As always, leave my your thoughts below.

How Do Data Loopholes Slow Down the Treatment of Breast Cancer?

Considering it’s Breast Cancer Awareness Month, the timing of this post is hopefully helping a very important cause.  For reasons I won’t go into here, I’ve recently become more familiar with breast cancer then I would have otherwise.  When confronted with a new topic of interest, it’s my nature to dig in and learn everything I can about it.

The National Cancer Institute provides a wealth of information on breast cancer but being a “software guy” … the way a mammogram results combined with a clinical breast exam can detect early signs of cancer stood out to me as an important information issue.

I began to wonder where that information was captured and stored (after the test and examination) … and how it was ultimately used in follow-up care with the patient.  I didn’t expect to learn what I did.

The American College of Radiology (ACR) has established a uniform way for radiologists to describe mammogram findings.  The system is called BI-RADS and includes standardized structured codes or values.  Each BI-RADS code has a follow-up plan associated with it to help radiologists and other physicians manage a patient’s care.  These values are often used to trigger notifications of the findings or other follow-up steps.  This makes perfect sense to me except there is a (big data loophole) problem.

The BI-RAD findings (or values) are typically found on a text based report … or determined by the examining physician.  They are then captured or manually transcribed in the EMR as free text notes that are added to the medical record as text … unstructured data living in a structured data environment.  This is the loophole!  It’s technically there but not able to be used.

Sometimes this step can be missed completely and the results are not put into the EMR system at all (human error) … or, more likely, the BI-RAD value is not transcribed in the right place as a structured data field.  There are just two of the reasons reasons this loophole can be caused.

You may not be aware, but an Electronic Medical Records (EMR) system is generally optimized for structured data.  Most EMRs don’t leverage text based unstructured data (test results, physician notes, observations, findings, etc.) in ways that they could.  It’s a known weakness of many of today’s EMR systems.

To net this out … it’s entirely possible that cancer is detected using the BI-RADS value but the information does not find it’s way into the right place in the EMR system because it’s text based and the EMR cannot recognize it.  This EMR system limitation has no way of determining what the text based information is, or how to use it.

The impact of this is staggering.  Let’s think about this in terms of timely follow-up on cancer detection.  A system that is not able to use the BI-RAD value could mean patients are not being followed-up on properly (or at all) – even though they are diagnosed with breast cancer.  Yes, this  can actually happen if the value is buried in the text and not being used by the EMR.  The unstructured data loophole is a big deal!

Don’t take my word for it.  University of North Carolina Health Care (UNCH) has announced new findings from mining clinical data to improve the accuracy of its 2012 Physician Quality Reporting System (PQRS) measures, achieving double digit quality improvements in the areas of mammogram, colon cancer and pneumonia screening.  They are taking steps to close data loopholes.

The new findings indicate mammogram values are present in structured data 52% of the time … and present in unstructured data 48% of the time.  Almost half the time the unstructured data is not presented with the rest of the structured data.  Ouch, that’s a big data loophole.

The new findings also indicate CRC screening (colon cancer) values are present in structured data just 17% of the time … and present in unstructured data 83% of the time.  As a man of a certain age, this scares me in words that can’t be published.  Another big data loophole.

Thankfully leading organizations like UNCH are closing these data loopholes today with solutions that understand unstructured data and can “structure it” for use in EMR systems … pasted from an IBM press release dated today:

Timely Follow-up of Abnormal Cancer Screening Results:  Follow-up care for patients with abnormal tests is often delayed because the results are buried in electronic medical records.  Using IBM Content Analytics, UNCHC can extract abnormal results from cancer screening reports such as mammograms and colonoscopies and store the results as structured data.  The structured results are used to generate alerts immediately for physicians to proactively follow-up with patients that have abnormal cancer screening results.

This is an example of what IBM calls Smarter Care … where advanced analytics and cognitive computing can enable more holistic approach to individuals’ care, and can lead to an evolution in care delivery, with the potential for more effective outcomes and lower costs.  If an ounce of prevention is worth a pound of cure, an ounce of perspective extracted from a ton of data is priceless in potential savings.  IBM Content Analytics is part of the IBM Patient Care and Insights solution suite.

I’ve written several previous blogs on related topics that you might find interesting:

I am also speaking at the PCPCC Annual Fall Conference next Monday October 14th at 10am and will be discussing Smarter Care, UNCH’s findings and more.  Hope to see you there.

As always, leave me your feedback, questions and suggestions.

Healthcare Data is the New Oil: Delivering Smarter Care with Advanced Analytics

It has been said that “data” is the new “oil” of the 21st century.  That is certainly true in healthcare where a unique opportunity exists to leverage data – as fuel for better health outcomes.  Everything that happens with our health is documented … initially this was on paper … and more recently, in the form of electronic medical records.

Despite billions of incentive dollars being dolled out by the federal government to purchase Electronic Medical Record (EMR) systems and use in meaningful ways, there continues to be significant dissatisfaction with these systems.

In a recent Black Book Rankings survey, 80% surveyed claim their EMR solution does not meet the practice’s individual needs.  This is consistent with my own observations, where many express frustration that “the information goes in … but rarely, if ever, comes out”.

If the information never comes out, or it’s too hard to access, are we really maximizing its value?

It all boils down to our ability to leverage years and years of longitudinal patient population data to surface currently hidden insights … and put those insights to work to improve care.

It’s incredibly powerful to combine years of clinical patient population data (longitudinal patient histories) with other types of data such as social and lifestyle factors to surface new trends, patterns, anomalies and deviations.  These complex medical relationships (or context) trapped in the data are the key to identifying new ways to achieve better health outcomes.  Some organizations are already empowering physicians with these new insights.

Context can be critical in a lot of situations—but in healthcare, especially, it can be the difference between preventing a hospital readmission or not. It’s not enough, for example, to know that a patient has diabetes and smokes a pack of cigarettes each week. These factors are only part of the whole picture. Does she live on her own, with family or in a care facility? Does she have a knee injury that prevents her from an active exercise program? Has she been treated for any other illnesses recently? Did she experience a recent life-changing event, such as moving homes, getting a new job or having a baby? Is she able to cook meals for herself, does she rely on someone else to cook, or does she frequent cafeterias, restaurants or take-out windows?

All of these things and more can—and should—influence a patient’s care plan, because these are the factors that help determine which treatments will be most successful for each individual. And as our population grows and ages, a greater focus on individual wellness and increasing economic pressures are forcing providers, insurers, individuals and government agencies to find new ways to optimize healthcare outcomes while controlling costs.
Today’s data-driven healthcare environment provides the raw materials (or “oil”) to fuel this kind of personalized care, and make it cost-effective as well. But it takes savvy analysis to turn that data into the kind of reports and recommendations providers, patients and communities need to make informed decisions.

The good news: IBM is uniquely positioned to help organizations and individuals achieve these goals. The IBM® Smarter Care initiative draws on a comprehensive portfolio of advanced IBM technologies and services to help generate new patient insights that can improve the quality of care; facilitate collaboration among organizations, patients, government agencies and other groups; and promote wellness through a range of public health and social programs.

IBM Patient Care and Insights is a key component of the Smarter Care initiative. By incorporating advanced analytics with care management capabilities, Patient Care and Insights can produce valuable insights and enable holistic, individualized care.

Advanced analytics: Leading the way to Smarter Care

Several leading healthcare organizations are already on the path to Smarter Care and demonstrating the real-world benefits of advanced analytics from IBM. For example, in St. Louis, Missouri, BJC HealthCare—one of the largest nonprofit healthcare systems in the United States—is using natural language processing (NLP) and content analytics capabilities from IBM to extract information from patient records that are valuable for clinical research. By tapping into unstructured data, such as text-based doctors notes, BJC HealthCare is surfacing important social factors, demographic information and behavioral patterns that would otherwise be hidden from researchers.

BJC HealthCare is also using IBM technologies to reduce hospital readmissions for chronic heart failure (CHF). The organization is analyzing clinical data such as ejection fraction metrics (which represent the volume of blood pumped out of the heart with each beat) to better predict which patients are most likely to be readmitted. These insights enable providers to implement tailored interventions that can avoid some readmissions.

The University of North Carolina (UNC) Health Care is using Patient Care and Insights for three new pilot projects. First, UNC is employing NLP and content analytics on free-text clinical notes to discover predictors of hospital readmission, identifying patients at risk and improving pre-admission prediction models.

UNC is also using IBM technology to empower patients. IBM NLP technology is helping to transform clinical data contained electronic medical records (EMRs) into a format that can be presented to patients through an easy-to-use portal. Streamlined access to information will help patients make more informed decisions and encourage deeper participation in their own care.

Finally, UNC is using NLP to help generate alerts and reminders for physicians. With NLP, the organization is extracting key unstructured data from EMRs, such as abnormal cancer test results, and then storing this data in a structured form within a data warehouse. The structured data can then be used to produce alerts for prompt follow-up care.

This is just the beginning. As organizations continue to launch new projects that capitalize on advanced analytics, case management and other technologies from IBM, we expect to see some very innovative approaches to delivering Smarter Care.

Learn more about IBM Smarter Care by visiting:

ibm.com/smarterplanet/us/en/smarter_care/overview/

For more about IBM Patient Care and Insights, visit:

ibm.com/software/ecm/patient-care/

As always, share your comments or questions below.

Moving Beyond One-Size-Fits-All Medicine to Data-Driven Insights with Similarity Analytics

Traditionally, Doctors have been oriented toward diagnosing and treating individual organ systems.  Clinical trials and medical research has typically focused on one disease at a time.  And today’s treatment guidelines are geared toward treating a “standard” patient with a single illness.

That’s nice… But the real world doesn’t work that way.

Most of us patients do not fit these narrow profiles … especially as we grow older and things get complicated.  We (patients) might display symptoms common to a variety of illnesses, or might already be suffering from multiple diseases.  Almost 25% of Medicaid patients have at least five comorbidities.[1]

This might explain why it’s estimated that physicians deviate from the recommended guidelines 40% of the time.  It might also explain why there is a real thirst in healthcare for evidence-based insights derived from patient population data.

In other industries, data-driven insights are often the only way organizations work with their customers.  Think of retailing and Amazon.com.  Amazon analyzes your past purchases, your past clicks and other data to anticipate what you might need and present you with a variety of options all based on data driven insights.  You might think that by now, every industry would analyze data from the past to predict the future.

That’s not true in healthcare where treating complex patients can be challenging and technology to handle this level of complexity really hasn’t existed.  Treatment guidelines are sometimes vague and may not exist at all when a patient has multiple diseases or is at risk for developing them.  In other words, one-size-fits-all approaches tend to be self limiting.

Treating patients with multiple conditions is also costly. In fact, 76% of all Medicare expenditures apply to patients with five or more chronic conditions.[2]  To reduce costs, doctors need ways to identify early intervention opportunities that address not only the primary disease but also any additional conditions that a patient might develop.

Consequently, Doctors are forced to adopt ad hoc strategies that include relying on their own personal experiences (and knowledge) among other approaches.  Straying from those guidelines (where available) might not deliver the best outcomes but it’s been the only option they have … until now

Similarity analytics offers a way to augment traditional treatment guidelines, enabling healthcare providers to use individual patient data (including both structured and unstructured data) as well as insights from a similar patient population to enhance clinical decision-making.  With similarity analytics, healthcare providers and payers can move beyond a one-size-fits-all approach to deliver data-driven, personalized care that helps improve outcomes, increase the quality of care and reduce costs.

IBM similarity analytics capabilities, developed by IBM Research, play an essential role in IBM Patient Care and Insights … a comprehensive healthcare solution that provides a range of advanced analytics capabilities to support patient-centered care processes.  Here is a link to a video (with yours truly) from the recent launch in Las Vegas (my part starts at 8:45 mins).

How do similarity analytics capabilities work?

Let’s take an elderly patient with diabetes (a chronic disease) who presents with ankle swelling, dyspnea (difficulty breathing) and rales (a rattling sound heard during examination with a stethoscope).   Diabetes by itself is bad enough … but the care process gets more complicated (and more costly) when other comorbid conditions are present.

With these reported symptoms and observed signs, the patient might be at risk for other chronic diseases such as congestive heart failure.  But exactly how much at risk and when?

In the past, Doctors have had no way of knowing this.  There are tens of thousands of possible dimensions that need to be understood, analyzed and compared to get an answer to this question.  Think of a spreadsheet where the patient is a single row … and in that spreadsheet and there are 30,000 columns of data that need to be analyzed in an instant … and someone’s life could be at stake based on the outcome of the analysis.  In other words, Doctors have been handicapped in their ability to deliver quality care because of the absence of this type of analysis.

With IBM Patient Care and Insights (IPCI), a healthcare organization can collect and integrate a broad range of patient data from electronic medical records systems and other data sources (such as claims, socioeconomic and operational) … from past test results to clinical notes … into a single, longitudinal record.  Similarity analytics then enables the provider to draw on this comprehensive collection of data to compare the patient with other patients in a larger population.  With IBM Similarity Analytics (part of IPCI), the provider can analyze tens of thousands of possible comparison points to find similar patients … those patients with the most similar clinical traits at the same point in their disease progression as the patient in question.

Why is finding similar patients helpful?  First, providers can see what primary diagnoses and treatments have been applied to similar patients … some diagnoses and treatments might have otherwise eluded Doctors.  Second, providers (and payers) can identify hidden intervention opportunities … such as an illness that the patient is at risk of developing or the risk of the patient’s current condition deteriorating.  Surfacing hidden intervention opportunities is critical in addressing the costs and complexity of healthcare … especially when treating patients with multiple diseases.

Importantly, providers can also predict potential outcomes for an individual patient based on the outcomes of similar patients. Knowing what has happened to a patient’s peer group given certain treatments can help doctors hone in on the right intervention for this particular patient … before things take a turn for the worse.

There are many areas where similarity analytics are helpful.  Disease onset prediction, readmissions prevention, physician matching, resource utilization and management and drug treatment efficacy are just a few of the use cases.  My colleagues in IBM Research have been working on this technology for years.

By finding similar patients, pinpointing risks and helping to predict results, similarity analytics can ultimately help healthcare providers and payers improve the quality of care and deliver better outcomes, even for patients with multiple illnesses.  By working with other analytics capabilities to enable providers to apply the right interventions earlier, similarity analytics can also help pinpoint the specific risk factors for a given patient.  Those risk factors can become the basis for an individualized care plan.

In a future blog post, I’ll focus on the care management capabilities of IBM Patient Care and Insights so you can see how this solution helps put analytics insights into action.

Until then, learn more about IBM Patient Care and Insights by visiting:

http://www-01.ibm.com/software/ecm/patient-care/

Read specifically about IBM Research and Similarity Analytics by visiting:

http://ibmresearchnews.blogspot.com/2012/10/data-driven-healthcare-analytics-from.html

As always …  look forward to reading your comments and questions.


[1] Projection of Chronic Illness Prevalence and Cost Inflation from RAND Health, October 2000.

[2] KE Thorpe and DH Howard, “The rise in spending among Medicare beneficiaries: the role of chronic disease prevalence and changes in treatment intensity,” <link: http://content.healthaffairs.org/content/25/5/w378.full&gt; Health Affairs 25:5 (2006): 378–388.

Advanced Analytics … The Next Big Thing in Healthcare

If you are in the healthcare industry, you know you’ are facing a number of significant challenges. First and foremost, you are being asked to meet rising expectations for higher-quality care, better outcomes and lower costs. But at the same time, you face a critical shortage of resources and an aging population that will require a greater portion of those limited resources every day.

Chronic diseases present some of the toughest challenges. Approximately 45 percent of adults in the United States have at least one chronic illness.[1] Those chronic illnesses not only make life difficult for patients, they also stretch healthcare resources thin and cost the U.S. economy more than $1 trillion annually.[2]

Advanced analytics can give you an edge in balancing all of these demands, and in figuring out how to continue the balancing act as the industry evolves. With advanced analytics, you can leverage a broader range of patient information and surface early, targeted intervention opportunities that ultimately help you enhance the quality of care, improve outcomes and reduce costs.

Content Analytics

Content Analytics capabilities, such as those offered through IBM Content and Predictive Analytics for Healthcare, can help you analyze a wider range of patient information than you could before. In the past, analytics solutions were frequently limited to structured data—such as the data found in electronic medical record (EMR) and claims systems. But content analytics lets you incorporate unstructured sources as well, including doctors’ dictated notes, discharge orders, radiology reports, faxes and more.  Powerful natural language processing is at work to enable this.

To see how valuable that unstructured information can be in uncovering insights, read my previous blog post, “Playing the Healthcare Analytics Shell Game.”

Predictive Analytics

Predictive analysis capabilities can help you identify patients at risk for developing additional illnesses or requiring further interventions. You can use predictive modeling, trending and scoring to anticipate patient outcomes and evaluate the potential effects of new interventions. 

Similarity Analytics

Using patient similarity analytics capabilities, such as those developed by IBM Research, a provider could examine thousands of patient attributes at once. That includes not only clinical attributes but also demographic, social and financial ones. By assessing similarities of attributes in broad patient population, providers can better anticipate disease onset, compare treatment effectiveness and develop more targeted healthcare plans.

Surface new intervention opportunities

The insights you gain from these analytics capabilities are the keys to  discovering opportunities for new, individualized and highly targeted patient interventions—interventions that can reduce expensive hospital readmissions for chronic patients, avoid the onset of other illnesses, prevent postoperative infections, slow the deterioration of conditions and more. That all adds up to better care and better outcomes at a lower cost.

In future posts, I’ll present a more in-depth discussion of patient similarity analytics and examine how advanced analytics can be integrated with care management.  In the meantime, I’d be eager to read your comments and questions.  In the mean time, check out some of the analytics research currently underway at IBM Research,


[1] S.Y. Wu, A. Green, “Projection of chronic illness prevalence and cost inflation,” RAND Health, 2000.

[2] Milken Institute, “An Unhealthy America: The Economic Burden of Chronic Disease Charting a New Course to Save Lives and Increase Productivity and Economic Growth,” October 2007, http://www.milkeninstitute.org/healthreform/pdf/AnUnhealthyAmericaExecSumm.pdf.

Playing The Healthcare Analytics Shell Game

When I think of how most healthcare organizations are analyzing their clinical data today … I get a mental picture of the old depression era shell game – one that takes place in the shadows and back alleys. For many who were down and out, those games were their only means of survival.

The shell game (also known as Thimblerig) is a game of chance. It requires three walnut shells (or thimbles, plastic cups, whatever) and a small round ball, about the size of a pea, or even an actual pea. It is played on almost any flat surface. This conjures images of depression era men huddled together … each hoping to win some money to buy food … or support their vices. Can you imagine playing a shell game just to win some money so you could afford to eat? A bit dramatic I know – but not too far off the mark.

The person perpetrating the game (called the thimblerigger, operator, or shell man) started the game by putting the pea under one of the shells. The shells were quickly shuffled or slid around to confuse and mislead the players as to which of the shells the pea is actually under … and the betting ensued. We now know, that the games were usually rigged. Many people were conned and never had a chance to win at all. The pea was often palmed or hidden, and not under any of the shells … in other words, there were no winners.

Many healthcare analytics systems and projects are exactly like that today – lots of players and no pea. The main component needed to win (or gain the key insight) is missing.  The “pea” … in this case, is unstructured data. And while it’s not a con game, finding the pea is the key to success … and can literally be the difference between life and death. Making medical decisions about a patient’s health is pretty important stuff. I want my care givers using all of the available and relevant information (medical evidence) as part of my care.

In healthcare today, most analytics initiatives and research efforts are done by using structured data only (which only represents 20% of the available data). I am not kidding.

This is like betting on a shell game without playing with the pea – it’s not possible to win and you are just wasting your money. In healthcare, critical clinical information (or the pea) is trapped in the unstructured data, free text, images, recordings and other forms of content. Nurse’s notes, lab results and discharge summaries are just a few examples of unstructured information that should be analyzed but in most cases … are not.

The reason used to be (for not doing this) … it’s too hard, too complicated, too costly, not good enough or some combination of the above. This was a show stopper for many.

Well guess what … those days are over. The technology needed to do this is available today and the reasons for inaction no longer apply.

In fact – this is now a healthcare imperative! Consider that over 80% of information is unstructured. Why would you even want to do analysis on only 1/5th of your available information?

I’ve written about the value of analyzing unstructured data in the past with Healthcare and ECM – What’s Up Doc? (part 1) and Healthcare and ECM – What’s Up Doc? (part 2).

Let’s look at the results from an actual project involving the analysis of both structured and unstructured data to see what is now possible (when you play “with the pea”).

Seton Family Healthcare is analyzing both structured and unstructured clinical (and operational) data today. Not surprisingly, they are ranked as the top health care system in Texas and among the top 100 integrated health care systems in the country. They are currently featured in a Forbes article describing how they are transforming healthcare delivery with the use of IBM Content and Predictive Analytics for Healthcare. This is a new “smarter analytics” solution that leverages unstructured data with the same natural language processing technology found in IBM Watson.

Seton’s efforts are focused on preventing hospital readmissions of Congestive Heart Failure (CHF) patients through analysis and visualization of newly created evidence based information. Why CHF?  (see the video overview)

Heart disease has long been the leading cause of death in the United States. The most recent data from the CDC shows that heart disease accounted for over 27% of overall mortality in the U.S. The overall costs of treating heart disease are also on the rise – estimated to have been $183 billion in 2009. This is expected to increase to $186 billion in 2023. In 2006 alone, Medicare spent $24 billion on heart disease. Yikes!

Combine those staggering numbers with the fact that CHF patients are the leading cause of readmissions in the United States. One in five patients suffer from preventable readmissions, according to the New England Journal of Medicine. Preventable readmissions also represent a whopping $17.4 billion in expenditures from the current $102.6 billion Medicare budget. Wow! How can they afford to pay for everything else?

They can’t … beginning in 2012, those hospitals with high readmission rates will be penalized. Given the above numbers, it shouldn’t be a shock that the new Medicare penalties will start with CHF readmissions. I imagine every hospital is paying attention to this right now.

Back to Seton … the work at Seton really underscores the value of analyzing your unstructured data. Here is a snapshot of some of the findings:

The Data We Thought Would Be Useful … Wasn’t

In some cases, the unstructured data is more valuable and more trustworthy then the structured data:

  • Left Ventricle Ejection Fraction (LVEF) values are found in both places but originate in text based lab results/reports. This is a test measurement of how much blood your left ventricle is pumping. Values of less than 50% can be an indicator of CHF. These values were found in just 2% of the structured data from patient encounters and 74% of the unstructured data from the same encounters.
  • Smoking Status indicators are also found in both places. I’ve written about this exact issue before in Healthcare and ECM – What’s Up Doc? (part 2). Indicators that a patient was smoking were found in 35% of the structured data from encounters and 81% of the unstructured data from the same encounters. But here’s the kicker … the structured data values were only 65% accurate and the unstructured data values were 95% accurate.

You tell me which is more valuable and trustworthy.

In other cases, the key insights could only be found from the unstructured data … as was no structured data at all or enough to be meaningful. This is equally as powerful.

  • Living Arrangement indicators were found in <1% of the structured data from the patient encounters. It was the unstructured data that revealed these insights (in 81% of the patient encounters). These unstructured values were also 100% accurate.
  • Drug and Alcohol Abuse indicators … same thing … 16% and 81% respectively.
  • Assisted Living indicators … same thing … 0% and 13% respectively. Even though only 13% of the encounters had a value, it was significant enough to rank in the top 18 of all predictors for CHF readmissions.

What this means … is that without including the unstructured data in the analysis, the ability to make accurate predictions about readmissions is highly compromised. In other words, it significantly undermines (or even prevents) the identification of the patients who are most at risk of readmission … and the most in need of care. HINT – Don’t play the game without the pea.

New Unexpected Indicators Emerged … CHF is a Highly Predictive Model

We started with 113 candidate predictors from structured and unstructured data sources. This list was expanded when new insights were surfaced like those mentioned above (and others). With the “right” information being analyzed the accuracy is compelling … the predictive accuracy was 49% at the 20th percentile and 97% at the 80th percentile. This means predictions about CHF readmissions should be pretty darn accurate.

18 Top CHF Readmission Predictors and Some Key Insights

The goal was not to find the top 18 predictors of readmissions … but to find the ones where taking a coordinated care approach makes sense and can change an outcome. Even though these predictors are specific to Seton’s patient population, they can serve as a baseline for others to start from.

  • Many of the highest indicators of CHF are not high predictors of 30-day readmissions. One might think LVEF values and Smoking Status are also high indicators of the probability of readmission … they are not. This could  only be determined through the analysis of both structured and unstructured data.
  • Some of the 18 predictors cannot impact the ability to reduce 30-day admissions. At least six fall into this category and examples include … Heart Disease History, Heart Attack History and Paid by Medicaid Indicator.
  • Many of the 18 predictors can impact the ability to reduce 30-day admissions and represent an opportunity to improve care through coordinated patient care. At least six fall into this category and examples include … Self Alcohol / Drug Use Indicator, Assisted Living Indicator, Lack of Emotion Support Indicator and Low Sodium Level Indicator. Social factors weigh heavily in determining those at risk of readmission and represent the best opportunity for coordinated/transitional care or ongoing case management.
  • The number one indicator came out of left field … Jugular Venous Distention Indicator. This was not one of the original 113 candidate indicators and only surfaced through the analysis of both structured and unstructured data (or finding the pea). For the non-cardiologists out there … this is when the jugular vein protrudes due to the associated pressure. It can be caused by a fluids imbalance or being “dried out”. This is a condition that would be observed by a clinician and would now be a key consideration of when to discharge a patient. It could also factor into any follow-up transitional care/case management programs.

But Wait … There’s More

Seton also examined other scenarios including resource utilization and identifying key waste areas (or unnecessary costs). We also studied Patient X – a random patient with 6 readmission encounters over an eight-month period. I’ll save Patient X for my next posting.

Smarter Analytics and Smarter Healthcare

It’s easy to see why Seton is ranked as the top health care system in Texas and among the top 100 integrated health care systems in the country. They are a shining example of an organization on the forefront of the healthcare transformation. The way they have put their content in motion with analytics to improve patient care, reduce unnecessary costs and avoid the Medicare penalties is something all healthcare organizations should strive for.

Perhaps most impressively, they’ve figured out how to play the healthcare analytics shell game and find the pea every time.  In doing so … everyone wins!

As always, leave me your comments and thoughts.