[Construct validity, reliability, and cutoff point of the Subjective Symptoms of Fatigue Test in Mexican workers]

To evaluate the construct validity and the internal reliability of the Subjective Symptoms of Fatigue Test (SSFT), as well as to establish its best cutoff point.

The SSFT was developed in Japan to measure fatigue in a population of workers; the test includes 30 items in three spheres. Construct validity was explored through differences among groups, using a population of 1399 workers from eight industries studied in Mexico City in 1999. We used the Theorell and Karasek's control-demand model as our nomological network, from which we derived two work hypotheses. Workers were divided into nine groups according to their level of control-demand; fatigue differences were assessed using analysis of covariance. The internal reliability level was tested with Cronbach's alpha. The cutoff point was obtained by means of the expected linear relation between the labor demands and fatigue.

The distribution of SSFT values was similar to the expected values, increasing proportionally with increasing demands or with decreasing control. The differences between groups were statistically significant. Cronbach's alpha was 0.89, while the value per sphere was always above 0.70. The cut-off points most closely correlated were 6 and 7 affirmative answers.

The observed score obtained in the distinct groups was as expected, in agreement with the study hypotheses. The internal reliability of the test was adequate. The proposed cutoff point for the general population was seven affirmative responses.