Three Core Questions from Actuaries
When health management providers take "intervention effect reports" to insurance negotiations, both sides often talk past each other. The provider says "participants improved X%"; the actuary asks "how do you prove this improvement was caused by your intervention, rather than by these people already being healthier?"
- Question 1: Is there selection bias among participants?Traditional reports cannot answer this; PSM reports provide pre/post-matching SMD comparison.
- Question 2: Is health improvement attributable to the intervention itself?Traditional reports only show correlation; PSM reports provide ATT+95%CI.
- Question 3: How should project effects and uncertainty be evaluated?Traditional reports are often qualitative; PSM reports can present effect estimates, confidence intervals, balance diagnostics, and limitations as inputs to program evaluation.
| Assessment Dimension | Traditional Report | PSM Evidence Report |
|---|---|---|
| Selection Bias Control | ❌ Untreated | ✅ SMD <0.1 |
| Causal Attribution | ❌ Simple comparison | ✅ PSM |
| Effect Size Report | ❌ p-value only | ✅ ATT+95%CI |
Key Takeaway
Insurers don't ask "is it effective?" They ask "can the effect be causally attributed, converted into claims numbers, and used to support actuarial decisions?" PSM causal evidence reports directly correspond to this evaluation framework.