Predictive modeling used for CM/DM interventions
Risk scores, gap scores indicate interventions
When Blue Cross Blue Shield of Michigan decided to create a proactive case management and disease management program, the Detroit-based insurer turned to predictive modeling to identify which of its members were eligible for the BlueHealthConnection program.
"We have a huge population. We are using predictive modeling because we don’t have the resources to have an impact on every member in the depth that we would like. We have developed a strategy to at least touch every member," says Jeff Powell, MS, MA, manager of outcomes measures and evaluation.
Almost 5 million members are covered by Blue Cross Blue Shield of Michigan, and 1.9 million of those are covered under BlueHealthConnection.
The insurer provides claims data and other information to an outside vendor specializing in predictive modeling to identify members who are eligible for the program.
Targeting three disease groups
The company has identified 68,000 diabetics, 3,500 members with congestive heart failure, 32,000 with ischemic heart disease, and 72,000 with asthma.
In creating its proactive case management model, the health plan started with diabetes and researched evidence-based literature to find the essential interventions that should occur with each of the populations and modifiable risk factors for each condition.
For instance, members with ischemic heart disease should manage their lipid protein levels and diabetics should monitor their hemoglobin A1C levels every three months and have regular eye and foot exams.
In establishing the model, the health plan considered the population’s age, gender, utilization patterns, comorbidities, and gaps in care, which include how often they see a physician and how often they have a recommended test or procedure.
The predictive modeling process stratifies the population and assigns a risk score and a gap score, showing if there is a gap in care.
The plan has established a numerical cutoff with risk scores and gap scores. Anyone who is above a certain point is considered high level or midlevel and is targeted for outbound calls from the case managers.
"We try to engage them and get them involved in a program," adds Michelle Fullerton, RN, CCM, manager of integrated case and disease management.
If a diabetic hasn’t recently had a hemoglobin A1C test, the case managers try to fill that gap by educating the member about the importance of the test so he or she can discuss it with the physician.
"The data we feed into the process set priorities in risk and gap scores. As we get more experience, we keep refining the predictive modeling. We want to get the right people at the right time with the right message," Powell says.
In addition, the insurer has developed its own predictive modeling for specific conditions, such as cancer. The company looked at several years of historical claims data and looked at variables to identify the likelihood of a member continuing along the cancer care continuum.
The health plan’s own predictive model is based on research that shows that certain conditions have a high association with comorbidities. Among them are back pain, benign prostate hypotrophy, and benign uterine conditions. The risk score includes many different criteria. It takes into account if a member has seen a physician for low back pain and has had recent imaging procedures.
"When we look at these, we may identify someone who has back pain and is trying to make a decision as to whether to have surgery," says Jann Caison-Sorey, MD, MHCA, FAAP, medical director for the program.
In this case, the company sends out a video that educates the member about the latest surgical techniques, contains testimonials from patients who did and who did not have the surgery, and includes other treatment options.
"The video doesn’t drive the member to one decision or another, but when they view it, they come away with an in-depth knowledge of their condition and the pros and cons of their options," Caison-Sorey adds.
Screening for depression an important key
The health plan has begun using a new screening tool for depression.
"When we looked at our population profiles, we noted an undercurrent of depression. Depression is one of those issues that often goes unaddressed, but it keeps patients from doing things to help themselves," she says.
Patients with congestive heart failure should be taking three medications a day, but if they’re depressed, they could miss a dose or fail to take their medication at all and end up in the hospital.
The case managers use the screening tool for members who have chronic conditions to identify people who may be depressed. If the screening shows possible depressions, the case managers tell the members that their answers to the questions indicate that they may need to discuss depression with their primary care physician.
When Blue Cross Blue Shield of Michigan decided to create a proactive case management and disease management program, the Detroit-based insurer turned to predictive modeling to identify which of its members were eligible for the BlueHealthConnection program.You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content