The LMWH strategy dominated UFH. Routine use of LMWH resulted in savings of approximately $89 per patient.

The Monte Carlo simulation showed the 95% confidence interval ranged from savings of $7 to $373 per patient with LMWH.

The lower rate of HIT accounted for much of the savings. In the univariate sensitivity analysis, the model was moderately sensitive to the baseline rate of HII, the odds ratio of HIT with LMWH and the cost of HIT. The multivariate sensitivity analysis confirmed that LMWH dominated UFH.

"Worst-case" scenario modelling showed that LMWH remained dominant even when it increased the risk of DVT, had less effect on the rate of HIT, and was substantially more costly than UFH.

In the threshold analysis, it was discovered that LMWH would result in equivalent or greater costs to UFH only in the following scenarios:

the cost of HIT treatment fell to 15.3% of the original estimate;

LMWH costs were nearly doubled to $39/day per patient;

LMWH increased the rate of DVT by 25% or more, relative to UFH;

LMWH decreased the rate of HIT by 26% or less, relative to UFH; or

the baseline rate of HIT with UFH decreased to 0.41% or less.