The prediction of return-to-work in a chronic pain population: Psychological, demographic and medical variables

The purpose of the present study was to identify potential predictors of return-to-work in a chronic pain population. The discriminatory and predictive potential of psychological, demographic and medical variables was investigated.;This investigation was conducted with 418 chronic pain patients. Each subject had completed a six-week multidisciplinary treatment program.;The psychological variables were measured by a post-treatment administration of the MMPI. K-corrected T scores were used for analysis. The demographic variables included age, sex, education, marital status and number in family. The medical variables were years from injury to treatment, type of injury, medications, and surgery. Of the 418 subjects, 150 (36%) returned to work and 268 (64%) did not return to work within a two year period.;Discriminant analysis procedures were used to determine the ability of the selected variables to predict group membership. Results indicated that when analyzed separately, only psychological variables were able to significantly differentiate between the groups at the 0.05 level of significance. The 13 MMPI scales accurately classified 66.75% of the cases.;Further analysis utilized all predictor variables in the discriminant function. These variables did significantly discriminate between groups at the 0.05 level of significance. The total classification accuracy of all variables was 67.46%.;A stepwise selection procedure reduced the number of predictor variables to nine. These variables included the Hs, Ma, Sc, and L MMPI scales, three demographic variables (age, marital status, number in family), and one medical variable (non-narcotic medication). These variables significantly differentiated between groups at the 0.0001 level of significance.;The nine selected variables accurately classified 241 (89.9%) of the 268 cases who did not return to work. These variables predicted accurate group membership for 45 (30.0%) of the 150 subjects who were able to return to work. The total classified accuracy is 68.42%.;Classification procedures indicated that individuals who had lower discriminant scores on Hs, Sc, age and number in family with higher scores on Ma, Mf, L, marital status (married) and non-narcotic medication were associated with the return-to-work group. Implications for treatment and recommendations for future research were discussed.