This naturalistic study presents outcomes for three therapy types practiced in psychiatric public health care in Sweden. Data were collected over a 3-year period at 13 outpatient psychiatric care services participating in the online Quality Assurance of Psychotherapy in Sweden (QAPS) system. Of the 1,498 registered patients, 14% never started psychotherapy, 17% dropped out from treatment, and 36% dropped out from data collection. Outcome measures included symptom severity, quality of life, and self-rated health. Outcomes were studied for 180 patients who received cognitive–behavioral, psychodynamic, or integrative/eclectic therapy after control for dropout representativity. Among treatment completers, patients with different pretreatment characteristics seem to have received different treatments. Patients showed significant improvements, and all therapy types had generally good outcomes in terms of symptom reduction and clinical recovery. Overall, the psychotherapy delivered by the Swedish public health services included in this study is beneficial for the majority of patients who complete treatment. Multilevel regression modeling revealed no significant effect for therapy type for three different outcome measures. Neither did treatment duration have any significant effect. The analysis did not demonstrate any significant therapist effects on the three outcome measures. The results must be interpreted with caution, as there was large attrition and incomplete data, nonrandom assignment to treatment, no treatment integrity control, and lack of long-term follow-up.