A method for comparing survival of burn patients to a standard survival curve.

A method is presented which enables individual institutions to compare their patient survival rates to a standard survival curve. The method can be used for survival following any sort of trauma, here illustrated for survival of burned patients. The only requirements is that a standard survival curve be available which gives predicted probabilities of survival for each patient. The procedure is simple and can be easily computed with a small calculator. It requires fewer patients than does a corresponding Chi-square test, and can be used with as few as 30 patients. The power for a specified alternative can be calculated. Detailed examples are given applying the procedure to a randomly selected set of 50 patients from the National Burn Information Exchange registry, Ann Arbor, Michigan.