Psychometric testing of a health survey for field reporting of decompression outcome.

A short, multi-item, self-administered diver health survey for measuring decompression illness (DCI)-related health status following diving was developed and tested. Seventy-six diver health surveys with subsequent diagnosis by hyperbaric specialists for DCI (yes, uncertain, no) were used for validity and reliability testing. Field testing in 21 occupational divers was used to evaluate test-retest reliability (n = 30) and typical diver health scores (n = 100). There was a significant correlation (Pearson's r = 0.50, P = 0.00001) between diver health scores and diagnosis for DCI. Cronbach's reliability coefficient was alpha = 0.73 and split half reliability was r(full) = 0.82. Test-retest reliability revised for dispersion was r(revised = 0.62. In working divers, a diver health score of six or greater indicated probable DCI with good specificity and sensitivity. The diver health survey provides a valid and reliable interval measure of DCI-related health status. For calibrating decompression models against diving outcome, an interval health score increases model degrees of freedom and may reduce response bias compared to a binary classification (yes/no DCI).