Potentially preventable readmissions are a scourge on the US healthcare system.
Each year millions of patients are discharged from the hospital, only to return within 30, 60, or 90 days.
Not only do patients, their families, and their caregivers suffer as a result, but hospitals, insurers, and the government waste billions of dollars that could be spent on other public health priorities. Many if not most of these readmissions could have been avoided if clinicians had effective, scalable, and timely methods for identifying not only which patients were the highest risk, but what steps should have been taken to mitigate that risk.
In recent years there has been a proliferation of readmission risk assessment models, yet readmission rates have barely budged. Fundamental flaws exist in most approaches in the areas of Data, Model Adaptability and Clinical Workflow Integration.
Many tools rely solely on historical patient data mined from the EHR or are disease-specific models that cannot be scaled to address all readmissions challenges. Models that rely on data collected at discharge are not timely enough to enable clinicians to take meaningful action, and ones that are not well-integrated into clinical workflow are not easily adopted.
For a readmission risk assessment tool to achieve a meaningful and long-lasting impact, these common pitfalls must be avoided at all costs. Today, I’m going to address some of the many data challenges faced when trying to risk assess patients.
Historical Data does not Predict Future Readmissions
Anybody who has ever invested in the stock market, rooted for a local sports team, or stuck with a television show past its tenth season knows that past performance gives you no guarantee on future returns. Factors beyond our control and beyond our ability to predict may cause our fortunes to turn on a dime.
Consider the Dow Jones Industrial Average: Those who had any investments around July of 2007 remember the feelings of unabashed optimism and certainty inspired by the great bull run of the early 2000s. Unfortunately, those same investors also most assuredly remember what happened shortly thereafter, when the financial crisis of 2008 erased trillions of dollars’ worth of wealth.
A recent systematic review of readmission risk models concluded that many hospitals still model their approach to identifying high-risk patients based on historical admissions, claims data, and outdated information on patient populations .
Using these old data to model and predict readmissions is dangerous. And with increasing pressure on hospitals to reduce readmissions, this approach also runs the risk of becoming extremely costly. Just ask the guy who splurged on Brooklyn Dodgers tickets in 1958, or the guy who put all his money into 8-track cassettes in 1979, or the guy who started a Hummer dealership in 2005.
Any of these folks will tell you that past performance data can not only betray you, but it may also prevent you from recognizing the obsolescence of your sources. As a result, this data may cost you a fortune.
Disease-Specific Models are too Myopic
In the Patient Protection and Affordable Care Act, Section 3025 created a program that penalizes hospitals for excess Heart Attack, Heart Failure, and Pneumonia readmissions in their Medicare populations . As a result, academic researchers and commercial enterprises built models that predicted readmissions for these diseases only .
Focusing efforts on these disease areas alone neglects other at-risk patient populations. So when the government announced new penalties for Total Hip and Knee Arthroplasty and Chronic Obstructive Pulmonary Disorder (COPD), hospitals are playing catch-up and ill-prepared to respond .
In addition, suppose that theoretically a hospital could develop risk assessment tools for the majority of their most-commonly treated diseases. Many readmitted patients have multiple chronic conditions and comorbidities. How can clinicians determine which tools to run on which patients?
This approach would be incredibly difficult to scale and maintain.
Identifying Individual Patient Needs to Improve Quality of Care Across the Continuum
By effectively identifying high-risk patients at the time of admission, nurses, physicians and case managers can better determine the most appropriate care plan required to best meet the individual needs of each patient.
Relying on historical utilization data or clinical factors solely to predict readmissions is ineffective and unreliable.
Case managers and physicians need a more holistic understanding of who a patient is, not the disease that they have. Using evidenced-based technology to fully evaluate patients at admission will allow clinicians to provide more focused personalized care, leading to more seamless care transitions and informed post-acute care providers.
The end goal is to improve quality of care for patients, while striving to reduce readmission rates.
Eric Heil is co-founder, president and CEO of RightCare, a company that specializes in reducing readmissions.
@all, comments are an excellent, absolutely classic example of what I call a ‘failure of perspective’ – although @Eric, your company is clearly on the scent…what I mean in this case is that all the noise around data, data, data is mostly quantitative in nature re: bad care transitions. Our view at Remain Home Solutions is more qualitative data should be acquired at, say, EMS moment thru triage and care plan development for a fall trauma case for example. The why and how did they fall and are they being discharged into same unsafe home environment. We expand the perspective beyond a singular ‘healthcare’ perspective to include assessment of the physical plant (home modification) and environment the patient is supposed to be healing productively in. And we fix the problem. At scale.
@RemainHome we keep family members from getting admitted/readmitted to nursing homes/care facilities by expediting home modifications at scale with software, products, services and funding – serving caregivers, seniors & their payers by REMOTELY assessing/designing/managing home mods with proprietary mobile platform.
If hospitals were not so expensive we would not care if re-admissions occurred. A space visitor fling over would say of our health system: “it’s principle purpose– it’s utility function–is to provide jobs for 17% of the population.” I want boring drab hospitals focusing on patients only and not elaborate social events put on by the Foundation. The hospital is only a tool and re-admission to it is not criminal. It’s like Snap-Lock tools ran all the car repair shops.
@Eric – Interesting argument. Can you tell us a little more about how your tool does its magic? Took a look at your client list – which appears to include some fairly well known names in the Philadelphia area. What kind of successes are they having? Or is it too early to tell?
@John – thanks for the feedback. RightCare software uses predictive analytics to stratify patient risk at admission using a proprietary tool. The disease-agnostic algorithms leverages a decade of leading academic research at the University of Pennsylvania School of Nursing to identify patients who are the most likely to face readmissions within either 30 or 60 days and need post-acute care services. We have efficacy and case studies available on our website. In short, it’s working and patients are getting the right type of post-acute care and avoiding readmissions. Thanks again for the interest here.
Reality is too that if you look at the aggregate fiscal impact on a hospital facility right now the actual Medicare readmissions penalty just isn’t enough to move the dial (it is no where near 1% at most hospitals and most hospitals it is less than 0.1%).
Other thing is that despite there being a huge preference for Epic/Cerner & a move to towards a single enterprise HIT vendor with combined clinicals/financials there is no statistically significant differences between hospital facility performance for FY13 and FY14 Medicare readmissions penalties across HIT vendors with the exception of McKesson Horizon customers.
@Al Lewis – Agree that it is critically important to incentive good care and quality, not just lower cost care. I believe that the best way to influence a patients behavior is data and kindness. Patients respond to data too and we have to inform them with data and outcomes regarding their personalized post care plan to maximizes the chances for a positive outcome. This is common practice in therapeutic decisions already. We evaluate the data first including response rates, efficacy levels, and safety risks when making treatment decisions. @BobbyGvegas – Thanks for the comment. The definition of insanity is doing the same thing over and over and expecting a different result. It’s time we rethink how we risk assess patients based on who they are and their individual needs, not the disease they have.
The bottom line is that, whether admissions or readmissions, it is way harder to influence patient behavior and condition than one would think it is. I myself am not hugely enamored of the whole readmissions emphasis. Just as focusing on readmissions for a few conditions meant neglecting others as you observe it does, likewise focusing on readmissions means neglecting many other things. And readmissions, in the commercial population especially, are a low percentage of total admissions.
In the Medicare population, ideally the government shouldn’t be incentivizing anything other than good care, like through single payments that don’t get influenced by infections or complications or 90-day readmissions. Once a patient shows up (or in the case of ACOs, once the person is assigned), they should be the provider’s problem without government micromanagement.
Even so, with the possible exception of end of life care, you’d be amazed at how little each change to payment systems affects actual morbidity and hence utilization. After 3 decades of getting the easy cases out of the hospital, reducing hospital utiization even more has become a much more incremental slog than policymakers would ever have expected.
First time I worked for a Medicare QIO as an analyst, in 1993 (then called “Peer Review Organizations”), my first assignment was grinding up HCFA claims data via Stata, looking at 5- and 30-day acute care readmits.
21 years later CMS is STILL analyzing readmits.