Lee 2007 (HMCV).

Methods	Randomised, double‐blind, double‐dummy, parallel‐group study.
Participants	Planned: Total 480, duloxetine 240, paroxetine 240. Randomised: Total 478, duloxetine 238, paroxetine 240. Completed: Total 349, duloxetine 166, paroxetine 183. Criteria for Inclusion: Patients of either sex and at least 18 years of age who met the DSM‐IV criteria for nonpsychotic Major Depressive Disorder (MDD), and had a Hamilton rating scale for depression (HDRS‐17) total score ≥15 and Clinical Global Impression (CGI) Severity score ≥4 were included.
Interventions	Duloxetine: 60 mg/day (given orally once a day as two 30 mg capsules). Paroxetine: 20 mg/day, given orally once a day as two 10 mg capsules.
Outcomes	Efficacy: The primary measure of efficacy was to assess the severity of depression and its improvement during the course of therapy, using the HDRS‐17. Secondary measures of efficacy included: presence and severity of anxiety (HAM‐A), severity of illness (CGI‐Severity), degree of improvement (PGI‐Improvement), degree to which physical complaints were bothersome to patient (SSI), and experience of overall pain (VAS). Safety: Assessment of adverse events, including serious adverse events, routine laboratory tests, vital signs, and ECGs.
Notes	For the primary efficacy analysis, a one‐sided 97.5% confidence interval of the adjusted mean HDRS‐17 total score was used to determine whether duloxetine was noninferior to paroxetine. Non‐inferiority was declared if the upper bound of the one‐sided 97.5% confidence interval was less than 2.2. Date of first patient enrolled: 17 February 2004. Date of last patient completed: 12 June 2005.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Patients meeting the entry criteria were randomized in a ratio of 1:1 to either duloxetine 60 mg/day or paroxetine 20 mg/day".
Allocation concealment (selection bias)	Unclear risk	No information provided.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Quote: "double blind, double dummy".
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Quote: "double blind, double dummy".
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Quote: "the primary analysis was based on the Per Protocol (PP) population. The analysis of treatment emergent adverse effects was based on all randomized patients with at least one post‐baseline observation and taking at least one dose of their randomzed medication (safety set). All other results are based on all randomized patients with at least one post‐baseline observation (Intention‐to‐treat population)".
Selective reporting (reporting bias)	High risk	Only adverse events occurring with frequency > or = 5% were reported.
Other bias	Unclear risk	Sponsorship bias cannot be ruled out.