Findings from previous psychotherapy research suggest that the majority of improvement takes place in the initial phase of treatment with the relative effectiveness dropping at a negatively accelerating rate. However, the evidence for this pattern of change comes from investigations of short-term treatments and it is unclear whether this also holds for more flexible and long-term psychotherapy. We provided open-ended treatments under routine-care conditions for a representative sample of 362 patients, including a large proportion characterized by severe psychopathology. Patients attended 52 sessions on average (SD = 59, range = 1-364, Mdn = 36). Our results indicated that the degree of improvement was linearly associated with time spent in psychotherapy and contingent upon the severity of psychological problems at intake. The least severely afflicted received the shortest treatments, experienced the most rapid change but demonstrated smaller overall magnitudes of improvement. More severely suffering patients received longer treatments, had slower rates of change but in general received greater overall benefits. We argue that previous suggestions of psychotherapy dosage have been less appropriate for patients suffering from moderate to severe psychopathology.