Identifying and managing chronically ill populations: Lower costs and better outcomes start with smarter data analysis
By Chris Scheib
For data professionals well-versed in Pareto's rule and segmentation analyses, the reality is hardly a surprise. But for others, the stats can be eye-opening: Patients with chronic illnesses account for as much as 70 percent of health care costs. And the people who live with asthma, diabetes, hypertension, heart disease and other conditions are far more likely to experience catastrophic health events requiring expensive long-term treatments.
Today, more than 40 percent of the American population – 130 million people – have a chronic illness. And as baby boomers reach retirement age and costs continue to rise, the pressure will only increase on providers and payers to improve their ability to help patients better manage their chronic conditions and diseases.
The outsized importance of disease and condition management
Given how costs are disproportionately skewed to chronic patients, the timing, methods and techniques for managing these patient populations are more important than ever. The fact is, the rising costs could be better managed – and real cost savings achieved – if care-management resources are employed to ensure that at-risk patients are identified and interventions are made at the right time. Unfortunately, many payers and providers are making decisions based on preconceived hypotheses and incorrect assumptions – which proves costly.
While payers and providers have adopted disease-management and wellness management programs, they've historically engaged individuals who were already ill. Unfortunately, that's too late to have a meaningful impact. A study published in the February 2009 issue of the Journal of the American Medical Association found that of the 15 care-coordination programs it studied, none achieved statistically significant improvements in health outcomes, costs or utilization of health services.
This highlights the true opportunity: preventive services. Providers and payers need analytical models that actually predict which patients are likely to develop certain diseases and offer timely interventions that truly prevent the conditions from developing. Those models can also help determine which interventions at which times for which patients can yield the best patient compliance to reduce incidents, complications and costs.
Personalized, proactive, preventive
These predictive analyses can revolutionize the essential model of health care delivery by tapping the vast amounts of health care data and shifting the focus to personalized, proactive and preventive care to improve outcomes. To make that happen, providers and payers need to successfully address key challenges:
Uncover the hidden trends. Through careful analysis of an integrated data set, payers should identify the health characteristics of the patient population and the most effective treatments. More importantly, higher-level analytics can move beyond simple rules-based selections to help predict at-risk patients, create risk analyses and identify population types most likely to respond to treatment programs – before they wreak financial havoc.
Improve disease-management compliance. It's important to attain a 360-degree view of the health care population. Analytical models can shed light on behavioral attributes of entire segments of the population, such as those who consistently take their medication at the right times. They can also identify high-cost segments of the population to manage problem areas proactively – such as individuals likely to relapse.
Measure the performance of treatment plans. Developing a consolidated view of treatment plan compliance can be critical to successful disease management. These views can capture the results of activities and interactions to assess the impact of treatments on outcomes, report performance from either an aggregated view or a more detailed level, and show the segments that are in compliance and why. It's critical to evaluate compliance risk factors based on patterns of behavior, environment, or frequency of visits to physicians or hospitals. Finally, try to adapt customized messaging based on results of behavioral analysis to improve program adherence and outcomes.
The requirements for a software foundation for health and condition management
Superior data integration. You need to transform data – from any source – to pinpoint and segment population groups with the highest risk of noncompliance.
An analytical approach. Gain a greater understanding of the health characteristics of your entire population, and identify those (outside of chronic disease segments) who would benefit from condition management or wellness programs.
Powerful predictive analytics. Identify trends and establish triggers that generate early-warning alerts for changes in risk factors, environmental conditions or behaviors that lead to negative outcomes.
Customer management capabilities. These include automated campaigns and communications based on predictive insights to ensure that the right message reaches the right audience in the right channel – before a health event occurs.
Intuitive, Web-based reporting. See online behaviors and refine your strategy to maximize the impact of online channels.
Disease management: where cost control and clinical outcomes align
Armed with these analyses, payers and providers are ideally positioned to prevent the onset of certain conditions. That not only benefits the patient; it benefits the bottom line as well. That requires sophisticated automation and processes to communicate with and engage at-risk patients. Like the marketing automation systems popular in a sales context, these systems have a range of tools and assets at their disposal that can be applied in countless combinations. For instance, a diabetic patient might receive an e-mail reminder to make an appointment for an annual eye examination – as well as a follow-up text-message reminder on the day before the appointment.
On a more sophisticated level, a patient presenting a range of pre-diabetic symptoms or appointments – perhaps an eye exam followed by a consultation about a circulation problem – might receive a packet of information about diabetes prevention. A patient who takes daily medicine for a cardiac condition – who has stopped ordering refills – might get a letter inquiring about the lapse and encouraging medication compliance. All of these communications are carefully scripted and synchronized to comply with acknowledged or recommended clinical pathways. The result can be a reduction in clinical events and their associated costs – and an improved quality of life for patients who can sidestep or minimize the impact of diseases and conditions.
Moving beyond common chronics
Understandably, early initiatives in disease management focused on patients with chronic conditions. However, new-breed analytics can amplify the impact of disease-management programs by predicting which members or patients are likely to develop a chronic disease within a given time frame, opening up opportunities to intervene and prevent.
Just as compellingly, payers and providers can use these analytics to move beyond chronic conditions to identify clinically unique populations that might be susceptible to shorter-term conditions that can be prevented with appropriate measures and interventions.
The right combination of predictive analytics, sophisticated automation and scripted engagement can elevate disease management from a reactive cost-control measure to an illness-prevention and mitigation strategy that dramatically lowers costs and improves the quality of life for wide ranges of patients.
Bio: Chris Scheib is a Senior Industry Consultant for health care at SAS.