Pieces Insights

Pieces Technologies at SXSW 2017

Pieces Technologies at SXSW 2017

This year, we’re coming to SXSW to meet with the influencers and leaders who can shape the next era of integrated healthcare.

Pieces Iris™, our platform for coordinated case management, connects social service organizations and healthcare providers – finally addressing the social and economic determinants of health in a patient’s care regimen. 

That means better care. And healthier communities:

  • 360-degree care management: Managing all the determinants of health outcomes
  • Closed-loop referrals: Connecting providers and services directly
  • HMIS for the homeless: Providing full-fledged homeless management
  • Intuitive, cloud-based design: Offering an easy and accessible tool for all users


The giveaway

On average, we could spend $5.11 on swag for each SXSW visitor. Or, we could use that money to do some good.

At our booth, Pieces Tech will donate $5.11 on your behalf – all you have to do is show up and choose which of our five partner organizations you’d like to support. Then we’ll introduce you to Pieces Iris™, which is reshaping the way care providers and local organizations work together – to the benefit of our communities.

Find us at SXSW

March 12-15, 2017 | BOOTH #708
Nonprofit donation activity and Pieces Iris™ demos

Monday, March 13, 5:20 p.m.
Pieces Iris™ on the pitch stage: We’ll be giving out a jackpot of donation coinage!


Register for SXSW

There’s still time. SXSW is spectacular and inspiring – you won’t regret it.



The Scientific Method: Putting the HVBP Program Under the Microscope By: Nadia Christensen, MD

As clinicians, we are all first and foremost scientists. We have been trained on the “scientific method” – to clearly understand the problem, the hypothesis, the interventions, the outcome, and the implications for future study. That has been drilled into us in school, in training, and in every piece of evidence that we lovingly read in our daily news feeds. We place so much emphasis on the evidence when it comes to treating our patients, however, I am not sure that we place enough emphasis on the evidence as it applies to our healthcare programs.  

The Problem: Rising Healthcare Costs

National healthcare spending has been increasing every year. Recent results released by CMS at the end of 2016 shows our national 2015 healthcare expenditure at $3.2 Trillion or $9,900 per person and accounting for 17.8% of GDP. As you might expect, the largest share of the bill is paid by the US government with Medicare picking up $646B and Medicaid paying for $545B. Close behind the US government in healthcare spending is the private insurance plans picking up $1072B.  In last place are the consumers, with a paltry of out of pocket spend of $338B.

The Intervention: Programs to Rein in Cost

With these huge dollars poured into healthcare, it is not surprising that CMS has implemented numerous programs to shift our payment system from fee-for-service to value-based care. Instead of paying hospitals and providers for the number of visits and tests they order, now payments are based on the value of care that they deliver. The value-based programs have included a variety of hospital programs including Hospital-Acquired Condition Reduction Program (HACRP), Hospital Readmissions Reduction Program (HRRP), and Hospital Value-Based Purchasing (HVBP) Program. As we head toward the new programs coming up this year, such as the Alternative Payment Models (APMs) and Merit-Based Incentive Payment Systems (MIPS), it is a good time to glance back at the programs that have been in place and consider how well they have been working.

The HVBP Program is one program that brings both quality of care and cost of care into the program. Hospitals in the program have 2% of their reimbursement withheld and then hospitals are gauged on measures such as mortality, HACs, patient safety, patient experience, efficiency, and cost reduction. Based on the hospitals total performance score, they may earn back their 2% and have the opportunity to earn an additional 2%. At the moment, the program is cost neutral as far as healthcare expenditures, so the main question becomes: Is this financial incentive for hospitals really improving care?

Is HVBP Program Improving Outcomes?

A 2015 study looking at the early effects of the HVBP program compared 12 clinical processes and 8 patient experiences measures between hospitals enrolled in the program and hospitals excluded from the program. This study looked at a 1-year time period from 2011 to 2012. The study found no significant difference in any of the clinical processes or the patient experience measures during the first implementation period of the program. The authors concluded that the financial incentives were too low to incentivize hospitals or that the complicated design might not have given them a clear target. The results suggest that we have yet to find the right measures to improve clinical process and patient experience.

In 2016, a new study investigated if HVBP was indeed improving care by comparing mortality for AMI, heart failure, and pneumonia between hospitals participating in the program and hospitals excluded from the program. This study was conducted over a 5-year period from 2008-2013. The study found that mortality trends between the two groups was small and non-significant. In looking at subgroups of hospitals, including poor performers, there was still no association between HVBP Program participation and better outcomes. The authors concluded that we have yet to find the right quality metrics and incentives to improve patient outcomes.

The Bottom Line

As we apply the scientific method to the HVBP Program, one would question if the program is targeting the right outcomes or providing the right incentives? What interventions have the high performing hospitals implemented that may be contributing to their better outcomes? Are there other factors that CMS should be targeting? For example, hospital ownership and social determinants are now shown to play a role in improving outcomes, yet they are not currently factored in HVBP program design or measures. So, as any good scientist would conclude, the HVBP “experiment,” so far, is inconclusive.  Scientific method would dictate that the next step is to put our subject back under the microscope and test new hypotheses.


HIMSS 2017: What you need to know

HIMSS 2017:
What you need to know

Pieces Technologies will be visiting HIMSS 2017 with one goal:

Change how healthcare organizations see their data.

With the proliferation of healthcare data and metrics comes confusion:

  • Which data really matters?
  • How can we turn big data and AI and into tangible decision-making?

We’ll show you at the event. For now, let’s start with some important HIMSS resources.

Where we’ll be at HIMSS 2017

Our CEO, Ruben Amarasingham, MD, will be all over HIMSS – most notably moderating a panel Monday and giving a special talk on Tuesday. Here are highlights to remember.


  MONDAY, FEB. 20 – 10:30AM | 320 | Chaplin Theater
Shattering the Glass Ceiling: Lessons Learned for Aspiring Female Executives
CEO Ruben Amarasingham, moderator | Learn More
  MONDAY, FEB. 20 – 10:00AM – 6:00PM | Booth 7785-07 | Innovation Zone
Exhibit Hours
  TUESDAY, FEB. 21 – 9:30AM – 6:00PM | Booth 7785-07 | Innovation Zone
Exhibit Hours
  TUESDAY, Feb. 21 – 2:00PM | Booth 7785-07 | Innovation Zone
AI Approaches for Throughput Management | Learn More
  WEDNESDAY, Feb. 22 – 9:30AM – 4:30PM | Booth 7785-07 | Innovation Zone
Exhibit Hours


VR headset giveaway

We’re going to back up our promise by allowing visitors to see everything differently – with VR. Our booth’s VR experience will surprise you, and it’ll give a stark demonstration of the concepts behind Pieces Decision Support.

Plus, by coming to our booth and chatting, you could walk away with your own VR equipment. Come see us at booth #7785-07!

HIMSS schedule and map

Let us take the complexity out of this whirlwind event. Check out the map below to locate us in the Innovation Zone, and visit the full HIMSS map and detailed HIMSS schedule to get a better handle on your experience.

Haven’t registered?

Do it! The world’s best healthcare minds are waiting for you.

Register for HIMSS

Skip the line

You can cut to the chase right now and schedule a demo of Pieces Decision Support, our solution for turning clinical data into informed clinical decisions and interventions.

Schedule a free demo at HIMSS.

Spiral Memo Pads, Creepy Hallways, and the Original Macintosh:  A Former Doctor’s Perspective.

Spiral Memo Pads, Creepy Hallways, and the Original Macintosh: A Former Doctor’s Perspective.

By Sanaz Cordes, MD


Sometimes friends, clients, students, or family ask: “Don’t you miss being a real doctor?”  I always pause for a moment when I get this question.  I’ve given up explaining that being a doctor is like being a Marine…. you kind of always get to “be” one, even after you’re no longer in active service.  I once even had a local newspaper reporter refer to me as “Sanaz Cordes, a ‘former’ physician,” in an article that I’m certain no one read.

But, I must admit that, as I work with health tech startups, I long for the ability to be teleported back to the early 2000’s (we’ll leave the backdating vague), when I was actively practicing inpatient and outpatient medicine.  When I think about the “tools” we used back then, as compared to what’s available now, I long for a “do-over.”  Some of my biggest frustrations with practicing medicine centered around the piecemeal, manual workflow of caring for patients.

At the time, the shiniest object in the room was the EMR! Almost no practices, and very few hospitals, had EMRs for charting, CPOE, or even results viewing.  I remember carrying spiral Mead “memo” pads and a pencil while walking back and forth between the hospital’s nurses’ station, the patients’ rooms, and the floor secretary’s desk dozens of times each night!  Saturday morning rounds used to take upwards of 4 hours.  I always fretted missing a critical lab result, forgetting to place an important order, leaving out relevant chart notes for consultants, or even forgetting to see a patient altogether!  And, sadly, one of these things would normally occur weekly.

It’s sometimes hard for me to comprehend the types of tools and technologies that are available for physicians today.  I spent most of 2014-2015 focusing on the physician shortage in this country.  Most of the early solutions were centered around enticing more providers into the profession.  But, I think we’re starting to see the pendulum swing.  Yes, we need more physicians and advanced practitioners, but the healthcare industry is finally allowing technology to help solve this challenge as well.  In no other industry have we seen such a “horse-and-buggy” mindset as we do in healthcare.  But, there is light at the end of this long and windy tunnel!  If you had told me 10 years ago, when I transitioned to a career in healthcare technology, about things like healthcare Artificial Intelligence and Machine Learning, it would have triggered a blank stare as my mind wandered to episodes of Star Trek: The Next Generation.  It would have seemed like a beautiful but impossible dream.

Today, there are companies building tools that predict a bad outcome before it happens.  These AI tools provide a risk score and alert the right resource, at the right point of care, before it’s too late.  And, the tool isn’t just predicting a “one-size-fits-all” outcome based on the disease using EMR data like labs and vitals.  It’s predicting and preventing an adverse outcome for one specific patient by interpreting structured and unstructured EMR data, social data, population analytics, and “truths” about that patient across the continuum of care.  If I just let that sink in, it almost drives this “former” doctor to tears.  I remember manually carrying tubes of my patients’ blood, spinal fluid, and even less glamorous specimens down a creepy hallway in the bowels of Parkland Memorial Hospital to the lab. I feared that my patients’ tests and the results I desperately needed would vanish.  Then, I would spend the next several hours religiously checking for results on the original Macintosh dinosaur that required a whack on the case to stop flickering. 

Technologies like cognitive processing can immediately flag a chronic congestive heart failure (CHF) patient with diabetes for a pharmacist consult when he arrives in the E.D. with a broken wrist – because he hasn’t filled his meds in over 8 weeks.  The enormity of marrying ambulatory data, social behavior, and inpatient EMR data to preemptively keep this CHF patient from returning with a hypertensive or diabetic crisis is mind-boggling.  Some forward-thinking health systems that are using these tools are seeing results like a 31% reduction in readmission rates among CHF patients.

Technologies like machine learning can independently aggregate and interpret new data to reveal hidden insights that humans could never manually process.  Not only are these technologies significantly improving the quality of patient care, but they driving enormous cost and efficiency savings by allocating resources in a prioritized manner and reducing costly adverse outcomes.  These technologies are enabling personalized surveillance, prediction, action, and reporting for patients – whether they’re at home, in their primary care provider’s office, in the E.D., or even in the ICU. 

I spent some of the happiest years of my career scurrying around the ward with my tattered copies of Pharmacopeia and Sanford Guide in hand. Then, the smart phone, loaded with the Epocrates app, arrived.  It was life-changing.  My colleagues and I were convinced that it couldn’t get any better than that.  Technology had peaked.  But, fast forward a decade or so (again, no need for exact math), and the disruption of our technology-averse industry continues!  I often tell my colleagues who are still practicing to seize the new technologies available to them.  After all, embracing innovation is why we are no longer whacking that old, flickering Macintosh.

The Arrival of Natural Language Processing in Healthcare

The Arrival of Natural Language Processing in Healthcare

Dave Paige


“Writing in English is like throwing mud at a wall.”

― Joseph Conrad


I recently watched the 2016 movie “Arrival.” The film explores the idea that what you think and how you think may actually be closely intertwined. “Arrival” is a story about humanity’s first contact with aliens and how a pair of scientist find ways to communicate without a common language. As they spend more and more time with the octopus-like creatures, they get increasingly frustrated with their lack of progress and must get creative in order to effectively communicate with these new visitors to Earth. I won’t spoil it for you, but this film beautifully illustrates how powerful and difficult the use of language can be, whether it’s between a linguist and a 10-foot-tall mollusk or with each other.

In a way, Electronic Medical Records (EMRs) can be seen as big repositories of human language about patients. Patient charts are largely written observations, results, and decisions that can be done with pen and paper or, now, with EMRs – completed with a computer keyboard. Depending on the size of the practice or hospital, this means the creation of hundreds or thousands of notes a day, including Progress Notes, Admission Notes, Procedure Notes, Discharge Summaries, Consultation Notes, and more. 

These narrative clinical notes are largely entered by providers, nurses, and the rest of the care team essentially writing sentences and paragraphs. These notes are the bread and butter of patient documentation – it’s natural for medical professionals to record what they’re thinking or discovered in the same manner in which they would tell a colleague in conversation. The design of EMRs did not spring fully formed out of thin air. Like most innovations in technology, these systems were designed predominantly as charge capture repositories, but with the intention of making some improvements the day-to-day workflows of doctors, nurses, and other patient care professionals. If paper charts hadn’t existed before the invention of EMRs, they may have been designed differently to better leverage computers and perhaps better organize the intake and recording of patient data. But, for the foreseeable future, narrative documentation isn’t going anywhere.

Humans can obviously read and execute decisions based on individual notes, but the manual time and effort to do so is drastically inefficient. One study found that when hospitalists were reviewing notes, much of the content received little attention or was read very quickly. But, even if providers and care teams could ingest these notes with 100% accuracy and speed, one of the heralded advantages of EMRs has always been that patient data would be digital so computers and their ever evolving algorithms can ingest and interpret them.

As a result, we’ve seen entire new industries spring up and succeed in doing just that, such as via Population Health reporting tools to analyze hundreds of patients at a time, or Clinical Decision Support tools that provide recommendations at the point of care – just to name a few.  However, up until recently, the bulk of this kind of data analysis has been limited to the relatively small amount of discrete data (approximately 20%) found in the EMR, which are typically entered in specific, discrete formats, such as the required selection of specific ranges of numbers for vitals or laboratory data fields.

One of the most exciting areas of recent innovation in healthcare technology has been in the field of “Natural Language Processing,” or NLP, which uses cognitive computing algorithms to allow a computer to “read” unstructured text and pick out key words and phrases, in context to “understand” its meaning. This allows computers to tap into the vast, previously unexplored swaths of note data that is simply unreadable by standard tools limited to ingesting only discrete data. Not surprisingly, a vast amount of useful clinical data is found in progress notes, nursing notes, and other free text notes that are not redundantly also documented in discrete fields. And, with Artificial Intelligence, not only can NLP be used to extract written out thoughts and findings about patients, but it can be leveraged to find patterns and run analysis to lead to discoveries that the doctor or nurse didn’t even realize when writing those notes! All this at lightning speed compared to any cost-prohibitive, manual attempts to do this work by hand.

NLP can be used in a variety of applications, such as helping advertisers read social media posts to improve their ad targeting or to help a computer compete in Jeopardy!. And there have even been successes in blending the line between social data and clinical data, such as a study that combined tweets about asthma with data taken from air-quality sensors and EMR data to predict with 75% accuracy if the Parkland emergency department staff could expect a high, low or medium number of asthma-related visits that day. NLP is being used by many young, innovative companies as a powerful tool to help provide real-time, personalized clinical decision support to identify for medical risks like sepsis or COPD, before they occur. This is an emerging and exciting area of study, and the ability and accuracy of NLP combined with machine learning will likely become ever more powerful and beneficial to uncover hidden gems of insight from patient data that was previously unable to be explored.

Yes, language (especially the English language) can be messy, but with the emergence of NLP there’s hope that computers are starting to bridge the gap to a better understanding of human language, which will allow humans to better understand each other and improve patient care. 


CMS Dollars Part III:  HACRP Made Simple  The Unwanted Visitor

CMS Dollars Part III:  HACRP Made Simple The Unwanted Visitor

Nadia Christensen, MD

Unfortunately, it happens all the time. You take your elderly father into the hospital for a routine surgery such as a total hip replacement, and while being treated for the hip, he ends up with a Hospital-Acquired Condition (HAC). Think of an HAC as your bothersome Aunt Ethel, an unwanted house guest who shows up when you least expect her, bringing in suitcases full of complications, and leaving a trail of havoc in her wake. Simply stated, an HAC is a potentially avoidable infection or complication that occurs while a patient is hospitalized for an unrelated condition. Years ago, people referred to these conditions as Healthcare-Associated Infections (HAIs) or nosocomial infections. However, today many preventable conditions, which are not infections such as pressure ulcers or DVTs, have been added to the list and term broadened to HACs.

Trends in HAIs and HACs before HACRP

In 2013 the AHRQ showed a 17% decrease in HACs.  This reduction was linked to a prevention of 50,000 deaths and a cost savings of $12 billion dollars. Despite this encouraging trend, an estimated 10% of hospitalized patients were still experiencing one or more HACs.

On the infection front, in 2014 the CDC estimated that on any given day, 1 in 25 hospitalized patients would develop at least one HAI. Although this number had seemed very high, the CDC, in its annual National and State Healthcare-Associated Infections Progress Report, found that rates of HAIs are actually decreasing. There was a 50% decrease in central line-associated bloodstream infection and a 17% decrease in surgical site infection after abdominal hysterectomy. Furthermore, there was a 13% decrease in hospital-onset MRSA blood infection and an 8% decrease in C. difficile infection.

Although the trends in HAIs and HACs were going in the right direction, from a legislative perspective, efforts were made to further decrease HAC-related mortality, morbidity, and cost. The Hospital-Acquired Condition Reduction Program (HACRP) beginning in fiscal year 2015 required CMS to reduce payments to hospitals who are in the lowest 25% of all hospitals in HAC performance score.

HACs Measured & Hospitals Participating

HACRP doesn’t measure all potential HACs. It measures a subgroup of HACs divided into two domains:

Domain 1 – AHRQ Patient Safety Indicator (PSI) 90 Composite (includes pressure ulcers, pneumothorax, CVCBSI, hip fracture, PE/DVT, postop sepsis, wound dehiscence, accidental puncture/laceration)

Domain 2 –NHSN Healthcare-Associated Infection (HAI) measures:

  • Central Line-Associated Bloodstream Infection (CLABSI)
  • Catheter-Associated Urinary Tract Infection (CAUTI)
  • Surgical Site Infection (SSI) – colon and hysterectomy
  • Methicillin-resistant Staphylococcus aureus (MRSA) bacteremia
  • Clostridium difficile Infection (CDI)

The two domains are measured over a two-year time period, followed by a brief period of review and correction, and the payment adjustment is applied to the following fiscal year. The time periods for the two domains are staggered – so, in essence, the measurement occurs over a 2.5-year period.

For example: for 2017, HACs were measured from 2012-2015, data review occurred in 2016, and the payment adjustment will be applied to this coming fiscal year.

For 2017, 3023 hospitals are included in HACRP. The program includes acute care hospitals but excludes long-term care hospitals, critical access hospitals, rehabilitation hospitals, psychiatric hospitals, children’s hospitals, and IPPS exempt cancer hospitals, and religious nonmedical health care institutions. In Guam, US Virgin Islands, Northern Mariana Islands, and American Samoa, it also excludes short term acute care hospitals.

Even Hospitals Dread Report Card Day

The payment adjustment for the HACRP is a payment penalty.  There is no financial incentive opportunity.  No matter how much they study for the HACR “exam”, there’s no extra credit!  To calculate if a hospital will be subject to payment reduction, each of the above listed measures is assigned a score between 1 and 10 based on the hospital’s performance – with higher scores indicating worse performance. Note, if a hospital does not have sufficient data for a given measure, the measure is excluded from the calculation.  Also, the measures in the two domains are not weighted equally. For 2017, Domain 1 has a 15% weight and Domain 2 has an 85% weight in the hospital’s total score.

This program “grades” hospitals on a curve.  Just like the old classroom days, hospitals are competing against each other on a yearly basis. As rates of HACs decrease across all hospitals, hospitals must work even harder to improve and keep their score in the top 75%. For FY 2017, hospitals with a total HAC score greater than 6.5700 are subject to the 1% payment reduction.  But, talk about a painful report card “F”! The payment reduction is applied to all the hospital’s CMS payments for the year.

HAC Penalties for 2017

CMS states that 769 of 3,203 hospitals ranked in the worst performing quartile will be subject to the 1% penalty in 2017. This is compared to 751 hospitals out of 3,211 hospitals penalized in 2016.  

HACRP, with its huge penalties totaling $364M last year, is coming under scrutiny in the year end, as clinicians question the validity of their calculations. In a recently published study, data shows that large hospitals are more likely to be identified as poor performers for measures that have a very low probability of complications. Additionally, prior studies have found there is a surveillance bias for more vigilant hospitals looking for complications.  No good deed goes unpunished, they say.

 The bottom line

There is an overabundance of programs available to help hospitals reduce HACs. On a national level, the CDC itself provides a variety of assessment tools, strategies, toolkits, and checklists that a hospital may implement. On state level, many evidence-based initiatives are available to aid in reduction of HACs. As hospitals work on many of these programs, they struggle to identify the interventions will help them pull ahead of the curve and keep them in the top 75% of hospitals.

In this new era of medicine, tools are available to hospitals that, a decade ago, clinicians would have described as figments of a sci-fi lover’s imagination.  Technologies built on cognitive computing are now available to provide real time detection of HACs before they happen. Imagine a piece of technology “smart” enough to know that an upward trending white blood cell count in a post-operative patient with a urinary catheter should be brought to a physician’s attention.   Getting this information to the right clinical team member, before the patient develops a urinary catheter-associated infection, is truly turning the traditionally reactive practice of medicine into a proactive one!  

Early identification and detection of patients needing closer monitoring and/or interventions will not only improve the quality of patient care, lower mortality, decrease LOS, but will also help hospitals keep their hard-earned CMS dollars in the bank.  And more importantly, it keeps those pesky, luggage-toting unwanted visitors out!