July 2009 FEATURE ARTICLES
Customer Relationship Management
Analytics Assist Communications
Successful marketing departments develop customer profiles that contain information on how and when their customers wish to receive communications.
By Larry Mosiman
More than 40 percent of the American population, or 130 million people, have a chronic illness such as diabetes, congestive heart failure, depression or asthma. These illnesses account for the majority of U.S. healthcare expenditures, driven by hospitalizations, emergency room visits, poor medication compliance and deviations from treatment plans.
A further opportunity that exists in the field of disease management is the use of analytics to dive deeper into the data to uncover hidden trends and nuggets of information.
They can be prevented when the appropriate steps are taken at the right time by bringing intensive care management resources to those afflicted.
Disease-management plans depend on their members to hold up their end of the agreement. Whether this is making and keeping appointments or staying on schedule with medications, compliance with the plan — and ultimately the success or failure of the plan — is largely up to the member. To improve the probability of success, some plans have adopted a process of sending periodic flyers or other communications that are designed to remind and influence the member to stay in compliance. These communications can be improved with a more-targeted message and method of communication.
Successful marketing departments develop customer profiles with information that tells how and when their customers want to be communicated with and what messages will be relevant, and then target their campaigns to follow these preferences. Much like marketing to consumers, disease-management programs could develop profiles that indicate how, when and where to communicate with their members, as well as a response system to acknowledge when the member complied with the plan directive.
To successfully communicate with members and expect that these communications will help attain the desired outcome of plan compliance, first have an understanding of the individual members. For plan members, the required information includes not only the basics such as contact information and plan details, but it could also include behavioral insight or medical insight that is analytically derived.
Determining the best channel of communication for the member’s particular lifestyle is essential. Is it their home phone, cell phone, text message, e-mail, regular mail? The use of analytical models can also help in understanding the response rates for each channel or other behavioral attributes of the member.
If which segments of the plan population have a high response rate to voice messages is understood, for example, call center resources can be saved by focusing efforts only on this segment. At a more advanced level, if pending issues are analytically determined, such as the segment of the population likely to relapse and be re-admitted, high-cost issues can be addressed prior to their occurrence. Developing this member information database is not technically difficult, but depending on the organization and the complexity of accessing the data sources, it can be time-consuming.
The next step is to ensure that the disease-management program can interact with its members over the preferred channel at the time and frequency required. There should be checks and alternative communication paths to account for changing situations, and, more importantly, the rate of compliance (i.e., success or failure of each communication) can be recorded so improvements can be made over time.
If the results of member interactions are captured and analyzed, not only can individual improvements be made to the communication methods, but also the effectiveness of the overall plan can be analyzed and its performance reported. This data will show which segments of the population are in compliance, and why.
For more information on
SAS analytics solutions
A further opportunity that exists in the field of disease management is the use of analytics to dive deeper into the data to uncover hidden trends and nuggets of information. For example, analytics could:
1) Gain insight into the segments in the entire book-of-business using any clinical, laboratory, claims, HRA or other demographic data available. Limiting any prior assumptions about how diseases, procedures, co-morbidities or services relate to one another, unsupervised learning algorithms could cull through the population and surface issues that were previously unknown. This takes the concept of data-driven decision making and applies it to disease management by surfacing utilization or cost issues present in the population, rather than having management generate preconceived hypotheses about relationships in the population. Consider the possibility of discovering pockets of unexpected, high-cost members who are driven by off-label prescribing, such as antinausea medication intended for chemotherapy patients prescribed to members suffering from morning sickness. This relationship might not be uncovered without the aid of analytics.
2) Create a catalog of predictive models whose targets are appropriate to the issue at hand. For example, if analysis shows that members with chronic diseases who have mental health co-morbidities are far costlier than expected, a predictive model that anticipates portions of the population likely to be non-compliant with their mental health prescriptions could allow for early and effective targeted outreach.
3) Look at how diseases progress through a utilization pathway to gain a greater understanding of the typical pathways that people experience when undergoing treatment for a particular condition. What types of laboratory services are utilized and when? Might there be instances in which a service that would have been inconsequential six months ago may now indicate that there is a treatment regression or uncertainty about the diagnosis?
Timely, targeted communications can play a key role in helping with the intensive-care management of diseases. To be effective, the behavior and preferences of individual members should be understood, and the interactions should account for these behaviors and preferences. There are challenges to overcome when identifying, stratifying and targeting members to drive improved plan compliance, but there are systems in place that can facilitate, record and report the success or failure of these efforts — providing a means for continuous improvement.
Larry Mosiman is worldwide product marketing manager for SAS customer intelligence solutions, Cary, N.C.