The Truven Health Blog

The latest healthcare topics from a trusted, proven, and unbiased source.

Risk of Hospitalization

By Therese Gorski/Tuesday, December 13, 2016

As Anne Fischer wrote in her last blog How a Data Scientist Thinks about Risk Stratification, in order to predict risk, we need to first determine what “risk” is being measured. One important risk is that of being hospitalized.  Risk of hospitalization or admission models have become more targeted, more personal and seemingly more prevalent. Further, they are very much in line with the goals of population health and the “quadruple aim” (improving patient care, reducing costs, improving the health of populations and improving the provider’s experience). If a person has proper ambulatory or outpatient care for a chronic disease, acute inpatient admissions related to that disease should be rare.

At Truven Health, we have been developing and maintaining risk models for decades. Many risk models, such as risk of cost, risk of mortality and risk of complications, are best used in aggregate within a population subset. This evaluation is typically done at a service line or patient group level, for example, all cardiovascular patients or all patients with a specific MS-DRG.

In today’s world of increasing interest in care management and targeted outreach, individual level risk models hold much promise.  These models evaluate and interpret risk for each individual patient. Several years ago, we started a journey to develop new targeted risk models – including risk of hospitalization – to meet this increasing business need. We focused on specific diseases with the intent to have high predictive accuracy overall but particularly at the “tails”, meaning we must perform well at identifying those who have a high likelihood of being admitted in the near future. Or, in data science terms, model performance was measured by its sensitivity, specificity and positive predictive value for all patients above a specific risk threshold, with an emphasis on high sensitivity (that is, reducing false negatives).

To start, we chose to focus on diabetes, congestive heart failure (CHF), asthma, and more recently chronic obstructive pulmonary disease (COPD). These conditions run the gamut for both prevalence and admission risk, with asthma being most prevalent and least likely to result in admission versus CHF, which is least common but most likely to result in admission. These are also chronic conditions that are typically managed, at some level, which also fit our criteria.

In addition to specific diseases, we also focused on identifying risk at one-, three- and six-month intervals so that a care manager can better understand the risk at hand and be able to prioritize accordingly. Further, we report risk across several categories including “all-cause admissions”, “potentially avoidable admissions” (largely defined by AHRQ) and risk of “related admissions” which represents conditions that are considered to be related to the main condition as defined by Truven Health. Finally, along with each model’s risk score, we provide the patient attributes that are driving the risk score, whether it be a recent hospitalization, level of disease severity or even age. We believe that this additional insight gives the care manager a bit more background on the patient, helping to explain why the patient may have an increased risk value.

These models have been notably successful in terms of their predictive accuracy and our work will continue as we expand the number of diseases and work with our clients to help make the information actionable. The general trend toward person specific risk versus risk in aggregate will only grow and will become more refined as we, and the industry in general, are able to obtain and incorporate more personalized information about people.

Therese Gorski
Senior Director, Advanced Analytics

How a Data Scientist Thinks about Risk Stratification

By Anne Fischer/Tuesday, October 25, 2016

“Risk”. It’s a word we hear every day in the healthcare industry. We want to avoid risk, we want to predict risk, we want to find patients that are high risk. We want to risk stratify populations (organize people into a set number of mutually exclusive tiers of increasing risk).

My recent blog posts have centered around the concept of Population Health. Clearly the idea of risk is particularly important in this world, where the goals are to keep well individuals healthy, avoid poor outcomes for those that are already sick, and minimize costs. Understanding, assessing, and predicting risk are all essential to this effort.

But what is “risk”? If you asked a physician, an insurer, and an average Joe on the street to describe “high risk” from a healthcare perspective, you would likely get very different answers. A physician might describe someone with high risk of developing a disease, high risk of a serious disease complication, or high risk of mortality. An insurer might describe someone at risk for a high amount of spending in the immediate future. The average Joe might describe someone at high risk for impairment/inability to function in daily life. Understanding the context-appropriate definition of risk is the first step toward building analytics to support risk analysis. And the appropriate definition is always dependent on the real world application.

Even when the application is understood, there is still considerable work to be done to identify the appropriate data and characteristics that lead to poor outcomes. Consider a discharge nurse who sees hundreds of patients a month as they prepare to depart from the hospital. Most knowledgeable hospital staff are aware that the most experienced discharge nurses will be able to tell you, with a high degree of accuracy, who is likely to show up back in the hospital in the near future. Multiple studies have tried to quantify the drivers of this type of “nurse’s intuition”. How do they know?

In 1964, United States Supreme Court Justice Potter Stewart used the now infamous phrase: “I know it when I see it” to describe his threshold test for obscenity in the case of Jacobellis v. Ohio. A discharge nurse might say much the same thing when asked to describe a patient at high risk for readmission. I know it when I see it. Characteristics such as illness burden, past behavior, social situation, self-care ability, home support, and others are often referred to, but the reality is that it’s the entire picture, and often a bit of an ambiguous “gut feeling” thrown in for good measure.

So how does Data Science fit into this picture? Our challenge as Data Scientists is to turn “I know it when I see it” into a measurable mathematical formula, so that everyone “knows it” even without seeing it in person. It involves extensive experimentation with different data sources, variables, and modeling techniques, as well as building in the capability for models to evolve and learn over time. At Truven Health Analytics, my team is exclusively focused on developing and testing new models, using various kinds of data that are readily available to us. In future blogs, we’ll describe some of these models including risk of developing diabetes and risk of admission. Truven Health, an IBM Company, now is positioned to move deeply into this space and develop these types of risk models by bringing together traditionally disparate data sources, clinical knowledge, and cutting edge modeling techniques.

Anne Fischer
Senior Director, Advanced Analytics

The Five Key Components of ACO Analytics

By John Azzolini/Tuesday, August 30, 2016

Accountable Care Organizations (ACOs) were created to provide financial incentives for providers to control costs and improve the quality of care. As they continue to advance, it is important for both providers and payers to ensure that risk is being appropriately shared between the two. This creates a unique set of challenges in determining the best way to design, manage and evaluate these programs. Whether you are running an ACO or contracting with one, data is integral to determining the best model. Without the proper data, those providing the care, and those paying for it, are flying blind.

What’s more, not all ACOs are created equal, with three general types of models accounting for the bulk of ACOs: employer-sponsored, employer-direct contracted, and those leveraging existing insurer relationships. The analytic tools used to evaluate performance will depend upon which type of relationship a payer has with the ACO.

The ACO Analytic “Tool Box”

The five analytic methods listed below are key for ACOs managing program performance, and for employers and health plans assessing the value they are obtaining from these programs.

1.       Attribution

All measurement depends on a connection made between the ACO and/or its providers and enrollees. As a result, we need to uncover who the enrollees are, and for whom the ACO is bearing risk.

Often, explicit patient assignment does not exist. Where it does, the evaluation models need to incorporate it into analytic databases. In the cases where it doesn’t, the ACO needs to perform that attribution based upon the observed pattern of care received by the patient population.

2.       Population Health Management

There are multiple tools available to identify and stratify patients, such as predictive modeling, where risk scores based on age, gender, and diagnosis are employed. Other methods employ biometric or health risk assessment information. Examples of these include Health and Longevity Scores, Health and Productivity Indexes, and Health Status/Opportunity Scores, that can be used to segment patient risk levels.

3.       Network Management

If an ACO is at financial risk for the management of individuals, it’s imperative to know where people are receiving health services, what kind of utilization is taking place out of network, and where those out-of-network services are being given.

Many beneficiaries are not locked into the ACO network, which makes knowing whether these services are being given by high quality, efficient providers paramount.

4.       Program Evaluation

It’s important for everyone involved through the continuum of care that an assessment be made on the effectiveness of the ACO. As anyone who has been involved in care evaluation can tell you, there are a host of methodological pitfalls that can throw a wrench into measuring program evaluation. Controlling for differences between populations – specifically those who use the ACO and those who do not – is exceedingly important to determine the effectiveness of that ACO.

5.       Quality Measurement

In addition to evaluating ACOs on the basis of financial performance, establishing core quality measures for ACOs enables us to glean insights we would otherwise not have. Metrics such as potentially-avoidable admissions, screening rates, and specific process and care measures give us a baseline for quality measurement that is imperative in defining how well the ACO is performing.

Embrace the Risk

Risk is a fact of life in healthcare; it always has been. But in this new landscape, the ways in which both providers and payers are sharing that risk has undergone a drastic shift. Everyone will assume risk, but as we’ve outlined above, the key is to understand and properly allocate that risk between providers, patients and payers. The data is there; to guide these decisions, the key is employing the appropriate tools to establish this balance.

John Azzolini
Senior Consulting Scientist

New Bundled Payments Research: Identifying the Most Predictive Risk Factors for TJR Readmissions

By Bob Kelley /Thursday, June 2, 2016

As part of our continuing research into bundled payments for commercially insured total joint replacements (TJRs), we recently turned our focus to patient readmissions — which we already know vary in cost and occurrence by geographic area and type of post-acute care a patient receives.

Logical thinking tells us that if we could accurately identify TJR patients who are at high risk for readmission due to certain clinical factors, we could reduce or prevent readmissions. And any change in TJR readmissions could then impact TJR value-based bundled payments. The goal, of course, would be to identify at-risk patients early in the care process, so that they could be targeted for special care or intervention programs.

But can we identify the top clinical risk factors?

To find out, we used TJR claims data from the Truven Health MarketScan® Commercial Database to analyze 84,648 simulated bundles. Each bundle was characterized by more than 35 data attributes, including patient demographics, presence of readmission, post-acute care usage, payments by component service type, comorbidities, utilization of healthcare services in the six months prior to the surgery, and more.

 Across several models, we found that the most predictive risk factors for TJR patient readmissions were:

  • Hospitalizations in the six months prior to the TJR
  • Emergency room visits in the six months prior
  • Unique condition/disease diagnoses in the six months prior

More specifically, when we looked at odds ratios estimated in a logistic regression model, additional significant TJR readmission risk factors were:

  • Number of comorbid, high-stage chronic conditions
  • Number of high-stage acute episodes in prior six months
  • Comorbid cardiovascular disease
  • Comorbid cerebrovascular disease

In a future blog post, we’ll take a look at whether or not using these clinical insights to identify and target at-risk patients provides better results than a random selection of patients.

Full details on this study, and additional findings, are available by downloading our latest research brief, Total Joint Replacements in the Commercially Insured Population: Predicting Risk of Readmission.

Bob Kelley
Senior Research Fellow, Advanced Analytics


Congratulations to the 2016 Advantage Award Winners

By Mike Boswood/Wednesday, May 4, 2016

At our annual customer conference last week, I had the honor of presenting the 2016 Advantage Awards to clients who documented significant results from projects that required innovative thinking and analytics, and produced greater value in healthcare.

Georgia Health Information Network (GaHIN) was the overall winner this year. GaHIN created a “network of networks” wherein patient information remains with the treating provider and flows only when there is authorization. Caregivers can now instantly access information from a multitude of sources through their EMR, enabling more informed decisions about treatments and avoiding unnecessary, expensive tests.

In addition to the overall winner, the following organizations were honored with Advantage Awards for their impressive results:

  • DeKalb Regional Health System: achieved reduced lengths of stay and cost savings from a systemic commitment to greater quality.
  • IASIS Healthcare: reduced readmissions through near-real time analytics and by improving the transition of care from hospital to home.
  • Liberty Mutual Insurance: exceeded enrollment goal by enabling members to understand their likely costs if they switched to the CDHP or stayed in a PPO.
  • Lockheed Martin: used analytics to understand costs of various treatments for lower back pain and to identify targeted intervention opportunities.
  • Prime Healthcare Services – Centinela Hospital Medical Center: turned a negative margin positive and improved patient satisfaction, with pervasive quality-oriented management.
  • State of Michigan: reduced invoice-to-claim discrepancies by over $30 million by automating the comparison process.
  • Trinity Mother Frances Hospitals: realized over $14M in cost reductions by guiding all departments to operate off standardized definitions, metrics, benchmarks and goals.

We support our clients’ commitment to building greater value in healthcare with our expanding range of solutions and services; we anticipate that as part of IBM Watson Health we will be able to provide ever more compelling solutions to the evolving challenges facing healthcare across the U.S. and globally.

Our congratulations to the success of all this year’s Advantage Award winners.

Mike Boswood
President and CEO
Truven Health Analytics