The dose-response effect refers to the relationship between the dose (e.g., length, frequency) of treatment and the subsequent probability of improvement. This systematic review aimed to synthesize the literature on the dose-response effect in routine psychological therapies delivered to adult patients with mental health problems. Twenty-six studies were eligible for inclusion. Different methodological approaches have been used to examine the dose-response effect; including survival analysis, multilevel modelling and descriptive cluster analyses. Replicated and consistent support was found for a curvilinear (log-linear or cubic) relationship between treatment length and outcomes, with few exceptions such as eating disorders and severe psychiatric populations. Optimal doses of psychotherapy in routine settings range between 4 and 26 sessions (4–6 for low intensity guided self-help) and vary according to setting, clinical population and outcome measures. Weekly therapy appears to accelerate the rate of improvement compared to less frequent schedules. Most of the reviewed evidence is from university counselling centres and outpatient psychotherapy clinics for common mental health problems. There is scarce and inconclusive evidence in clinical samples with chronic and severe mental disorders.