Objective: The association between duration of untreated psychosis (DUP) and later outcome is not fully understood. Jonas et al. in their 20-year follow-up found that the association could be explained by lead-time bias. In this study we aimed to analyze the relationship between DUP, time since onset of psychosis and functional outcome using a similar statistical approach as the Jonas study. Method: Using data from 496 participants with first-episode schizophrenia, DUP was assessed using the IRAOS and functioning at the baseline assessment and the subsequent follow-ups (1, 2, 5 and 10 years) was assessed using the GAF-F. For premorbid functioning, the Premorbid Assessment of Functioning Scale was used and rescaled to correspond to the GAF. Results: The model with the best fit of data included both a slope and a level change. This model of level of function over time had the inflection point at the time of first treatment. This model indicated a slow decline per year until first treatment, at which point there was a sharp decrease in functioning, and after which functioning gradually improved again. Both in this model and in models accounting for potential lead-time bias, however, longer DUP was associated with a decrease in function for each additional week of DUP. This is in contrast with the Jonas et al. study. Conclusion: In this study, we did not find evidence of a lead-time bias, but rather found that onset of treatment occurs at the time when participants level of functioning was most impaired, and consequently was not at random.