Medical Insurance Cost Correlation Analysis
A correlation study of what drives medical insurance charges, set in a scenario where a London investment firm is weighing a BMI-linked employee insurance scheme. It separates statistical signal from assumption and states plainly what correlation can and cannot support.
The Challenge
Across 1,000 anonymised records, BMI showed a strong, highly significant correlation with insurance charges (r = 0.709, p < 0.001). Age was weak (r = 0.080) and number of children showed no significant link (r = 0.031).
The scenario: an insurer's competitive quote varies the cost per employee by BMI, so staff outside an 18-24 BMI range would pay an extra monthly contribution. Before adopting it, the firm needs evidence of whether lifestyle factors actually correlate with insurance costs, and needs to be able to explain that evidence to employees who may see the scheme as unfair.
Approach
Results
BMI was by far the strongest correlate of insurance charges among the variables tested. Age showed a real but much weaker relationship, and number of children showed no meaningful link, a useful negative finding that rules out a factor that might otherwise be assumed relevant.
Correlation is not causation: higher BMI moving with higher charges does not show that BMI causes those costs. The dataset holds only three candidate variables, so unmeasured factors could explain part of the relationship.
What It Means
The evidence supports BMI as a relevant cost factor and rules out number of children. On its own, it does not justify charging employees by BMI: any scheme built on it would need to address the fairness concerns the scenario raises before adoption.