Imagine a standard of care where before the patient leaves the hospital, the institution already knows not just the patient’s critical data, but has also taken actions in response to the data that further reduce the risk of readmission within the next thirty days. The potential impact goes beyond delivering a more advanced quality of care. For institutions facing the radical industry shift to patient-based outcome models, predictive analytics presents an opportunity to make significant strides to decrease patient re-admission rates — and gain a stronger foothold in meeting compliance mandates the industry faces as a result of the Affordable Care Act. 

Predictive Analytics: Saving Lives

At present the healthcare system knows very little about us till we show up at the hospital or in a doctor’s office sick.  In order to become preventive, the healthcare system will have to begin tracking our vitals decades before it currently does and listening in ways it currently does not.

As I was getting ready to see my next patient, I began what has now become a familiar ritual for doctors and nurses everywhere, washing my hands, doing my part to not become a vector for disease. As I toweled off my chapped hands for the umpteenth time today, a thought occurred to me: how come we’re not better at preventing hospital acquired infections (HAIs) despite all this hand washing? Across the country doctors and nurses are scrubbing our hands until they are impossibly chapped, yet HAIs abound. We’re doing our part but not really making much of a dent. I wondered if there might be a way that our IT systems and the immense quantity of data we collect on each and every patient might lend a helping hand…

I find it hard to imagine that Accountable Care Organizations (ACO) are going to be as wildly successful as projected without an ability to actively use the data they are obligated to collect. The Centers for Medicare & Medicaid Services (CMS) require the average ACO to compile data on their patients and then use that same data to demonstrate that they’ve met the CMS requirements of data use, reporting and improving quality. If those requirements aren’t met… well let’s just say that ACO might not be around for much longer.

With the first round of CMS penalties about to come out for hospitals with unacceptable readmission rates, there will be a lot of people suddenly interested in ways to decrease their readmission rates. I’m guessing that a lot of those people will want to use the LACE index, developed in Ontario, Canada, to predict the risk of patients for death or unplanned readmission within 30 days of discharge. LACE is an acronym for Length of Stay, Acuity of admission, Co-morbities (as measured by a Charlson Score) and number of previous ED visits in the last six (6) months preceding this admission.

One of the most frustrating things for me about starting a new clinical process improvement (PI) project is just how flipping long it takes to get actionable results. Most of us use the tried and true PDCA (Plan, Do, Check, Act) method of W. Edward Deming fame. Predictive analytics short-circuits this PDCA process by looking at huge chunks of data…

Using predictive analytics and modeling tools is becoming a common and often daily occurrence in the work of caring for patients and populations in the healthcare industry.  But for someone who may be new to the field or has had limited exposure and/or experience with predictive analytics, the concept of mining data to predict the future may seem not only foreign, but a bit intimidating.

Predictive Analytics  are a powerful tool; of that there can be no doubt. However, like any tool, it only serves a purpose if it fulfills a useful function and almost as importantly is easy–to-use in fulfilling that purpose. For example, I can use a rock to drive a nail into a board, but the rock is hard to hold, difficult to use gently, easily breakable, etc. Ease of use for any tool means being able to complete a  task as easily as possible, with minimum frustration, every single time.  The same concept applies to the user interface (UI) of a seriously powerful tool like predictive analytics. The UI has to be effortless and intuitive, particularly if the end user is not a “super-user” and instead is a mere mortal who has  many other things to do in their busy life (like most of us!).

In case you didn’t know, the Supreme Court upheld key provisions of The Patient Protection and Affordable Care Act (PPACA). This decision will allow for the greatest change in our healthcare system since Medicare and Medicaid were enacted. While in some ways the envy of the world, we spend an exorbitant amount for mediocre results when taken as a whole and when compared to the rest of the world. As an Emergency Physician, I have a front-row view of the really good and the really bad of how we care for our citizens and pay for that care. Whether you agree or disagree with PPACA, we are now confronted with the addition of at least 30 million citizens to our healthcare system, not including those who will be added with the expansion of Medicaid.

All Things Predictive

Recently, I was talking with a colleague about a particularly difficult problem she was having in trying to ferret out why their in-hospital sepsis patients did not do as well as the patients that presented to the hospital in septic shock. Despite trying several tried and true performance improvement (PI) methods, the answers simply were not coming…