GHDP-123-19: Physician Cost Profiling – Reliability and Risk of Misclassification

Use this thread to discuss ANYTHING and EVERYTHING related to this syllabus reading.
Some possible questions include:

  • How can this reading be tested?
  • I don’t understand a specific topic/formula - Can we discuss this?
  • This reading gives me nightmares. Can we talk through it a bit?

Good luck!

I’m having a tough time understanding how the misclassification probability is calculated. It seems like if you only consider the lowest 25% of physicians as low cost then you’re going to have another 25% of physicians with a score below the median who are categorized as high cost. What am I missing?


Is it the calculation that you are having trouble with, or the interpretation of results?

I don’t think they are saying that the top 75% physicians are necessarily “high cost” physicians; just “not low cost.” The way you are describing low cost physicians would imply that half of all physicians are low cost, which could be a reasonable assumption, but I think the article is just using the 25th percentile as a more conservative threshold to define what makes a physician “low cost,” rather than using the median.

The end misclassification rate just suggests that some true “low cost” physicians had higher costs than normal and were therefore misclassified, for example due to receiving higher severity cases; while some true “not low cost” physicians were misclassified for the reverse reason.

Does that help? And can anyone else confirm whether I’m thinking of this correctly?

1 Like