Evaluating sources of traumatic spinal cord injury surveillance data in Colorado.

The purpose of this study is to evaluate the sources reporting hospitalized spinal cord injury cases to the statewide, population-based surveillance system in Colorado for the year 1994. Three reporting sources were evaluated: clinical contact persons, medical records departments, and a centralized statewide hospital discharge database. Two evaluation strategies were utilized; these include both measures of accuracy and estimates of missed cases. For the latter, capture-recapture techniques were used to estimate the number of hospitalized spinal cord injury cases missed by all three reporting sources. The clinical contact persons reported 84 confirmed cases, missed 80 confirmed cases, and reported 10 cases that were later determined not to have spinal cord injuries, resulting in a sensitivity of 0.51. Medical records departments and the discharge database reported 143 and 147 cases, respectively, missed 21 and 17 confirmed cases, and reported 118 and 69 cases that were later determined not to be cases of hospitalized injuries of the spinal cord, resulting in sensitivities of 0.87 and 0.90. Capture-recapture results indicate all three sources combined missed an estimated 1-5 cases, yielding a total annual incidence rate for hospitalized spinal cord injury ranging from 45.1 to 46.3 per million population.