You Don’t Know Mike

Published on July 14, 2016

Who is the ideal customer for your healthcare organization and the services you provide? When will they most need your services? Can you find them? Are they able to find you?

Most importantly, when are they most ready to pay attention to your message?

Who is at risk?

Imagine you are trying to prevent heart attacks for your ACO (Accountable Care Organization). The Mayo Clinicoffers this list of risk factors:

  • Men age 45 and older, women age 55 and older
  • Tobacco, primary & secondary exposure
  • High blood pressure
  • High blood cholesterol or triglycerides
  • Diabetes
  • Family history
  • Inactive lifestyle
  • Obesity
  • Stress
  • Stimulant drug use
  • History of preeclampsia
  • History of an autoimmune condition

A subset of the at-risk population can be measured through their clinical records – for example, for individuals with complete medical histories, the Framingham Risk Score estimates 10-year cardiovascular risk.

Your data is often incomplete: according to the CDC, the average consumer visits a physician 3.3 times per year, and 40% of consumers have one or fewer visits. EMR (Electronic Medical Record) data is incomplete – many medical conditions go undetected until symptoms become serious. You can boost that knowledge with health risk assessments, but as the Healthcare Intelligence Network has shown, few people fill them out (completion rates vary dramatically).


Imagine a customer named Mike. What do you know about him?Mike.jpg

Membership File

Mike is a 55 year-old man who buys insurance in the individual market, with no claims history. Mike is a new member of your ACO, with a Bronze level high-deductible plan. First red flag: Mike’s gender and age put him at a higher risk of heart attack.

HRA (Health Risk Assessment)

Mike ignored the HRA sent out as part of his enrollment. Ignoring the HRA means that Mike’s predicted “engagement” score is lower.

EMR (Electronic Medical Record)

Mike has no existing primary care relationship. His one prior visit, to a retail clinic for an unspecified virus, was paid for in cash (no claim record), and was not sent to your EMR.

The information you know about Mike is not enough to cross the threshold for outreach and intervention.

What DON’T you know about Mike – what information is missing? Is he really at risk for a heart attack?


Predictive models to the rescue!

Combining clinical history with external behavior data allows the development of predictive models that increase your ability to identify at-risk consumers. But will those consumers actually respond?

analytics_model.png

Who can you help?

Once identified, where do you intervene? Budgets are tight – you cannot afford to assign a care coordinator to each member of the at-risk population. How do you decide how to allocate your scarce resources?

Key lesson: Consumers respond and engage at different rates. Identifying who is likely to respond is crucial to achieving the best outcomes.

As referenced in this article from Healthcare IT News, we define “engagement” as a concept that combines a patient’s knowledge, skills, ability and willingness to manage their own health and care with interventions designed to increase activation and promote positive patient behavior.

Understanding who is most likely to respond is just as important as measuring who is at risk.


Back to Mike – what DON’T you know about him? What information is missing?

Clinical Data

Had his retail clinic visit been logged, you would see that Mike is 5’ 10” and over 200 pounds – Class I obese – our second red flag.

Social, Behavioral & Demographic Data

Mike has stable work, much of it hourly or project based as a contractor. He has limited benefits offered through his employer. While mostly sedentary outside of work, Mike owns a boat, and spends much of his free time hunting & fishing with friends. Mike is a weekend smoker, and has been for much of his life. His diet is not healthy, and alcohol accompanies many evenings, and most of his hobbies.

He is divorced – while his debt level is relatively low, much of his income supports his ex-spouse and two children. Mike rents an apartment, and has limited assets – he is living paycheck to paycheck – which limits his ability to cover his deductible. Mike lost his father to a heart attack at age 62, and his mother to breast cancer at 73.

Knowing this background information would raise the assessment of Mike’s risk significantly, perhaps enough to warrant intervention. Augmenting your membership and EMR data with consumer demographic data and resultant models will bring Mike to your attention.

Yet the assessment of his potential for engagement is still low. At this point in his life, Mike will ignore most reasonable attempts to change his behavior: health care, diet, and smoking.

When should you intervene?

Mike is about to have a health “wake up” call.ER_visit.jpg

Mike had a long, physical day at work and spent Friday evening at the local watering hole.

He woke up at 4am Saturday morning – “I woke up hurting. The first thing I noticed was pain in my chest. It didn’t feel like heartburn, more like a weight. My arm hurt, and neck and jaw. I was cold from sweating, and felt sick.”

Having lost his own father to an early heart attack, this event scared Mike enough to get up and drive to the nearest ER.

The good news – Mike was not experiencing an Acute Myocardial Infarction. In this case, Mike was suffering from influenza (flu shots not being part of his annual

Quit_Smoking.jpg

health regimen). Not having had the flu since he was a child, it was difficult for Mike to self-diagnose the severity of the event.

The better news – this wakeup call raises Mike’s receptivity to intervention significantly, especially for the next 30-60 days. He cares deeply about supporting his children, and wants to make positive changes to reduce his long term risks.

If you act swiftly, you can nudge Mike in several ways that will lower his risk profile – for example:

 

  • Find a primary care physician and schedule a wellness visit
  • Smoking cessation counseling
  • Diet & weight loss changes
  • Blood pressure & cholesterol treatment

Summary

There is not enough information in your membership file and clinical history to measure risk and allocate scarce resources. Augmenting patient data with consumer demographic information and predictive modeling allows you to intervene with those who are most likely to engage with your messages and services.

Add a comment

Subscribe to Email Updates