Fixed costs are higher and variable costs lower for clinical laborator
y testing than for alternate-site testing (AST). For AST to become cos
t-effective, variable costs must be reduced. Four models for accountin
g for the costs of clinical laboratory testing and AST are compared. A
common error is to omit some of the indirect costs of a cost-pertest
analysis for AST, thereby providing false claims. For a cost-per-test
analysis, we recommend that identical categories be used for both the
central laboratory and for AST, However, cost-per-test data alone do n
ot completely represent institutional costs or savings associated with
laboratory testing at any site.