Friday, August 24, 2018

Is this moral hazard or adverse selection, and why did BCG get it wrong?

Moral hazard can look very similar to adverse selection.  For example, if you see Volvo's running through stop signs more frequently than other cars, you could infer one of two things:

  • Moral Hazard:  the cost of running a stop sign is lower in a Volvo, so Volvo drivers run more stop signs; or
  • Adverse Selection: bad drivers are more likely to buy Volvo's so they are more likely to run stop signs.  

In the blog post below, a BCG study compares Medicare Advantage plans to traditional fee-for-service Medicare:
  • Single-year mortality rates fell from 6.8 percent in the traditional fee-for-service sample to 1.8 percent 
  • Patients in the Medicare Advantage plans had shorter average stays in the hospital (about 19 percent shorter.)
  • Patients in the managed plans were more likely to receive preventive care ...For example, diabetic patients in the fee-for-service sample had an average of 11.5 amputations per 1,000 patients; those in HMO plans with global capitation had only 0.3. 

and mistakenly infers moral hazard:
What distinguishes Medicare Advantage plans from traditional fee-for-service plans is the degree to which they use mechanisms designed to encourage the delivery of cost-effective quality care. Three critical mechanisms are financial incentives that are aligned with clinical best practices, a selective network of providers, and more active care management that emphasizes prevention to minimize expensive acute care.

 when the answer is likely adverse selection.  From colleague Larry Van Horn:

Those individuals who choose a Medicare Advantage (MA) plan are historically of a different risk profile (lower), and to the extent that results are case mix adjusted the severity of the MA population is overstated. In short, tough to make the case on the BCG study with which I am well aware.

No comments:

Post a Comment