Qlik Healthcare User Group

3 Posts authored by: Jeff Wu

We are in some exciting discussions with Epic, so while we wait for the possibilities of that news, we're going to skip the Hyperspace Integration Post for some Emergency Medicine talk.


In his New York Times Bestseller, Blink - The Power of Thinking Without Thinking, Malcolm Gladwell writes about the remarkable power that humans have to make very quick and accurate decisions in very challenging, information heavy or information lacking environments. He gives several illustrations of this from as wide a variety of fields as fire fighters being able to determine whether a house's foundation will be sturdy enough to enter, to art critics being able to instantly spot a forgery.  He even relates a story of psychologists developing a process that can accurately determine whether a couple will be divorced within five years simply by observing fractions of seconds of their facial expressions during a conversation. This 'gut-instinct' that we have, Gladwell describes as the psychological principle of 'thin-slicing'. 



Either I shouldn't take this new job or that burrito doesn't like the idea of moving to Kansas.


The most interesting chapter to me was Gladwell's narrative of Cook County Hospital's Emergency Department during the late 1990's.  One of the country's busiest trauma centers and the basis of the television drama series ER had a very big problem.  Their Emergency Room was being overloaded with patients.  With an astronomical quarter million patients being seen in their ED annually, the lines were endless with wait times necessitating patient families pack multiple meals.  Patient evaluations we're being done in hallways on gurneys and triage of the sickest patients was being delayed by the sheer volume of waiting patients. 


As with many ED's one of the most common reasons people would come in was for a suspected Myocardial Infarction.  As a facility with very limited resources, Cook County simply didn't have the capacity or staff to admit every patient that had a suspected heart attack.  Additionally, a heart attack is a problem that is incredibly tricky to nail down.  The tests that were typically run and a patients individual make-up all contributed to the possibility of whether that patient was having a heart attack or not.  Furthermore, the only test that could conclusively tell if a person was having one took several hours to be resulted.  Since it was too expensive to house every patient that fell into a 'gray area', the hospital created short-stay and observation units for patients who potentially were having a heart attack. Unfortunately, this just diverted the problem to a different area of the hospital and the incidence of false admissions remained high.





What Cook County needed was a way to more accurately and more quickly triage these patients.  For that they turned to the research of a cardiologist named Lee Goldman.  In the 1970's Dr. Goldman looked at the volumes of information and case studies of heart attack patients and worked with mathematicians to devise the most relevant and important risk factors that contribute to an actual MI.  What Goldman developed was coined the 'Goldman Algorithm' and it boiled down the possibility of a MI in a patient to four key factors. (1) Their ECG (2) unstable angina (3) fluid in lungs, and (4) systolic blood pressure '< 100'. At the time, Goldman's contemporaries scoffed at the thought that medicine and treatment of such a serious problem could be simplified so much.  When Cook County implemented Goldman's algorithm, the results spoke for themselves.  The Goldman alogorithm was a whopping 70 percent better at identifying true MI's and also significantly safer and more accurate in determining the severity of the problem as well as recommending a governed progression of treatment.


This flies in the face of the practical 'defensive medicine' that many physicians are employing today.  While only an estimated 10% of the patients admitted for a possible heart attack were actually having one, somewhere in the other 90% were the patients that would've been sent home based on the Goldman Algorithm with potentially disastrous results.  Furthermore, many detractors to this simplified method of diagnosis argue that the aims of medicine shouldn't be to keep patients out of the hospital but to not miss the correct diagnosis of someone who actually needs help. 


What if the insights from massive data volume, and the power that can be drawn from the 'gut instinct' (thin-slicing) are not mutually exclusive?  What if we could transform large amounts of data into easily digestible snippets for physicians to make use of more accurately in their judgments.  When IBM developed Watson, it wasn't just to be a gimmick to show that computers were better at storing and analyzing massive amounts of data.  (We all know our future metal overlords are good at that)  What it was meant to highlight was the possibility of a tool like this in assisting in the decision-making  process.  The lessons from Watson and the power of thin-slicing can be coupled today in QlikView!


Last fall, one of QlikView's customers, Allina Healthcare based in Minneapolis was gracious enough to present a webcast on analytics they'd done in QlikView.  One of the key subject areas they focused on was readmissions reductions.  The challenge they faced was a familiar one - they didn't have the resources to devote to every inpatient admission but they also knew that the data existed to provide powerful insight into how they could effectively target patients that had a high probability of readmission. Echoing the same efforts of Dr. Goldman, Allina recruited a particle physicist who used QlikView to develop a predictive model for readmission.  Each and every patient was assigned a risk-score and inter-disciplinary resources were focused on patients that fell into the high-risk category for readmission.  In just the first 2 months since implementing workflow changes, they saw a reduction in readmissions by more than 10% and it continues to trend down.


These are the types of opportunities that really lend themselves to QlikView.  Our tool can help make sense of the enormous volumes of data to allow care providers to make significantly more informed decisions for their patients and for the business. 


Let's send patients home and know it was the right decision to do so.

Wu's News - Meaningful Use

Posted by Jeff Wu Feb 11, 2013

Meaningful Use is the hot topic in US Healthcare.  President Obama's 2008 HITECH act incentivizes healthcare providers to adopt new technology in the hopes of modernizing an aging industry and decreasing the rapidly escalating costs of healthcare.


For the first years of Meaningful Use, healthcare providers are incentivized with a stimulus reimbursement for their investment in EHR technology. After, they are actually penalized for not doing so.


MU 1.jpg
First the carrot, than the stick


The problems come in the implementation and the submission process for Meaningful Use.  Meaningful Use will occur over several stages for healthcare providers in which they will need to demonstrate increasingly more robust use of EHR technology through an attestation process. Stage I and stage II metrics for Meaningful Use have been published by the Office of the National Coordinator (ONC) but have already undergone significant changes as measures are more accurately defined and more organizations complete their submission process and find exceptions.


For EHRs that are fully certified (as opposed to modularly certified) for Meaningful Use, customers have had significant issues with the standard MU reports coming from their EHR. Take for example one customer that realized that they were failing the Core Measure of Medication reconciliation for all patients upon admission to the hospital.  They ran the reports over and over, and realized that the number being returned from the report was well below the threshold and they could not figure out why. After some deeper analysis into their EHR reports, they realized that the reports, by default, took into account all patients admitted - including newborn babies.  An exception to this calculation was that newborn babies most obviously don't have any medications to reconcile upon admission (in this case delivery of the baby). They manually updated the report and their numbers skyrocketed.


By extension, another customer scrambled to implement a high-powered fully certified EHR by the end of June 2012 to insure that they had 90 days of data in 2012 to report for the full Medicare reimbursement amount.  Their hope was to start receiving stimulus money in December to help pay for the expensive implementation of the EHR.  In October, they ran the EHR's default reports for the first time so that they could be eligible for the stimulus and were deficient for almost every measure. What they realized was that several of the exceptions (like the one mentioned above) were required to be manually updated in the reports in order to get correct numbers. The work was more significant than they anticipated and to date, they are still unable to submit their first round of providers.  Additionally, if their numbers aren't sufficient in 2012, they'll also lose out on the maximum reimbursement.


MU 2.jpg
I told you it wasn't Alopecia!


The large problems with most EHRs is that they are not built for flexible reporting.  They are built as a documentation system with analytics as an afterthought. While most of the major vendors are in the process of making this a more significant component of their long term strategy, the simple matter is the needs and problems most organizations face now cannot be met with the standard tools provided. This will become even more apparent as we move into the second and third stages of MU.


That's where QlikView comes in. Our bread and butter is to make large and complex sets of data digestible and flexible for consumption at every level. From the aggregate down to the transactional, we allow direct and valuable insight into data. Nowhere has this been more apparent to me than in healthcare. QlikView is especially valuable with regards to Meaningful Use, which requires very discrete calculations for metrics that fluctuate regularly. In fact, the Meaningful Use stage I requirements were just adjusted in a release from CMS three weeks ago. Where static reports from an EHR may get run weekly or even monthly, QlikView can run Meaningful Use reports daily or faster, giving direct care providers the ability to complete care or documentation for metrics they otherwise would be deficient in later. Many QlikView customers are doing just this.


MU 3.jpg


Another big benefit is that there is no wait for an organization to have their EHR release new reports for any newly released metrics. They can be done as soon as the metrics are released from CMS, allowing a jumpstart into the attestation process.  For the changes to stage I released recently, most EHR vendors are likely not close to completion of updates to their MU reports, meaning their customers must wait until they're done to see if they are meeting the new metric requirements.  For many of the Meaningful Use Stage II metrics, they are simply a higher threshold of stage I metrics. Before you are even ready for stage II you can see how you're doing.


MU 4.jpg


On Thursday, February 21, Chris Elfner from Bellin Health and I will be presenting a webinar on QlikView and Meaningful Use.  We'll discuss how QlikView can be used as a flexible tool for providers to make sure they hit their metrics prior to the end of a reporting period. Additionally, we'll discuss how QlikView has been used by many customers to simplify the otherwise tedious process of attestation.


If you haven't yet, you can sign up here.


Up Next: You decide! Vote for the next topic in the survey here.


Howdy folks!  My name's Jeff Wu, and I'm a Solutions Architect here at QlikTech.  I joined QlikTech back in June of 2012 and prior to that was at Epic for six years serving as the Senior Application Director for the OpTime and Anesthesia applications.


Since joining I've gotten an opportunity to see what a lot of customers are doing with Qlikview and it has sparked my passion for better healthcare delivery and a desire to share my experiences in healthcare with others.


Today I'd like to talk about my favorite place in the hospital…


The Operating Room.


Go to sleep.jpg

You're getting sleepy...


In my years as an implementer, I've gone live with Epic at over 50 individual hospitals; there really is no place like the ORs. With the ED a close second, there is no place in the hospital that is more critical to have operational efficiency from start to finish. According to a recent HIMSS article, the ORs are responsible for up to 60% of a hospital's revenue. On that note, the ORs can also account for more than 60% of a hospital's margin. This means that the most minor of problems in this area can have monumental effects on a hospital's operational effectiveness. First Case Late Starts, Delays, Add-on Cases, PACU time, patient transfers, etc. can all have very negative repercussions if not addressed in a timely fashion.


Take for example, the effect of one situation:


A surgeon is 15 minutes late to pre-op his first patient of the morning. The patient is scheduled for a basic laparoscopic cholecystectomy to treat gall stones. When she arrives, the surgeon speaks with the anesthesiologist, who is hesitant to go through with the procedure due to the patient's history of smoking, asthma, diabetes, and a recent bout of pneumonia. The surgeon is annoyed, but not wanting to lose her surgical slot, orders some STAT blood tests to check the patient's ability to undergo surgery. Upon return, it is confirmed that the patient is too sick to undergo the procedure. The case is cancelled.


What sort of revenue impact did this have? Well let's see. Due to the surgeon arriving late, the case was cancelled too late to actually make use of the room time. Both the surgeon and anesthesiologist lost the revenue that would be associated with performing the procedure. To insure speedy transitions, supplies for a surgery are laid out prior to a patients transfer into the OR--all those supplies are now wasted. Any pre-induction meds mixed specifically for that patient by the pharmacy or anesthesiologist are also wasted. What started as one small problem (surgeon arriving late) can develop into a significantly more complex problem. Now, you might think that this is a very rare occurrence that such an unfortunate series of events could happen, but I guarantee that if you talk to any surgical staff, this kind of thing has happened at least once in the past month, or more probably, the past week.


All of these combine for a massive loss to the hospital. Now, there are bound to be times when physicians are late; there are bound to be times when a patient forgot that they had to be NPO, but addressing what can happen will mitigate the financial impact of what inevitably will happen.


So how can we address this?


With data.


But not just any, or some, we need all of it.  All of it is the only way we can have an informed understanding of the situation and make an informed decision to correct it. As technology has progressed, sophisticated documentation systems have evolved to capture all the components that are necessary to address these problems. The next problem is that this data lies in disparate systems that have difficulty  communicating with each other. This was what I dealt with on a near daily basis in my time at Epic. Customers came to me with problems that they had the data to understand and correct but could not associate it meaningfully to make an informed decision. Take, for example, the very simple question every OR Director or Manager has to deal with:


Based on the number of add-on cases for the day, should I open an additional OR or pay my staff overtime?


This seems like a simple question, but what goes in to making an informed decision for an issue like this is extremely complicated. Factors such as the add-on procedures, to the staff on hand, to the cost of staff, average hours per stay, average time per procedure, even the day of the week can all impact whether the person in charge says yay or nay to something as simple as opening an additional OR. Additionally, this information is in multiple places besides the EMR. In my experiences prior to QlikView, OR Directors and managers were relegated to simply making gut checks on whether they thought a new OR was needed or whether paying staff overtime was sufficient, even though they had powerful systems and loads of data to look across to make a decision.


Here's where I get excited by QlikView. Because of QlikView's inherent powers to associate disparate sources of data, the previously difficult or impossible is now a logical first step.


Surg App Add-ons.jpg

This application takes into account EMR and Timelog data to help visualize information for decision making.


And we're just scratching the surface on what's possible.


Check out our upcoming marketing data sheet on ED and OR operational efficiency coming soon!


Up Next: A review of Meaningful Use Stage I and a foray into Meaningful Use Stage II

Filter Blog

By date:
By tag: