Major costs associated with depression are related to work loss and reduced productivity. Since treatment can improve functioning in these areas, economic evaluations of new antidepressants should consider their impact on productivity costs. However, clinical trials do not generally measure effects of pharmacotherapy on disability days but rather use patient or clinician assessments of functioning (e.g., HAM-D). We have developed (and previously reported) a method using HAM-D scores to estimate disability days for comparing economic impact of different treatments. OBJECTIVES: The methodology is based upon theoretical considerations. In order to strengthen its value for practical applications, the current study assesses the predictive validity of the proposed methodology. METHODS: As patients using antidepressants make clinical improvements, the methodology predicts associated gains in productivity. We compared the gains predicted for clinical trial patients to actual observations from a real-world setting. For this purpose, we used data reported by Treglia and Neslusan at ISPOR’s 5th annual international meeting, and derived from a commercial health care database, examining patterns of absenteeism for 716 employees newly diagnosed with depression from 1994-1996. RESULTS: The observed number of medical absences averaged 14.0 days during the first 2 months after pharmacotherapy began and 7.5 during the next 2 months. In comparison, our new methodology predicted 12.4 days during the 2 months of initial treatment and 5.2 days in the following 2-month period. CONCLUSIONS: The new methodology predicted similar absenteeism for clinical trial patients as that observed among employees. The different levels of absenteeism are likely due to differences between the two groups (i.e., clinical trial versus real-world). Since our results support the predictive validity of the methodology, we are proceeding to use it conditional upon further validation involving simultaneous measurement of both HAM-D scores and disability days in a sample of patients with depression.