| Abstract: | A predictive model for employment after spinal cord injury was developed. The study population consisted of 154 spinal cord injured persons who were treated at our hospital between 1973 and 1979, and followed for seven years after injury. Demographic, social, and injury severity data were abstracted from each subject's hospital record. Motivation to work, employment history, and sources of postinjury financial support, were assessed by a vocational rehabilitation counselor. The study population was divided into four groups: persons continuously unemployed after injury, homemakers, students, and those employed at some time during the seven-year follow-up period. Stepwise discriminant analysis was used to develop a predictive model that ultimately included seven variables: gender, motivation to work, whether the patient's last job required ambulation, race, educational level, a functional ability score, and whether the patient had children. The model correctly classified 82% of those persons who were continuously unemployed, 100% of homemakers, 63% of students, and 72% of employed subjects. Overall, 79% of subjects were classified correctly. The most important classification errors were between the unemployed and employed groups. Seventeen percent of employed patients were incorrectly classified as unemployed, and 11% of unemployed patients were incorrectly classified as employed. Although there are other determinants of postinjury vocational status, individual potential can be assessed by means of a comparatively small set of predictor variables. |
