Baldwin 1995.

Methods	Eight‐week, double‐blind, randomised study.
Participants	Outpatients meeting DSM‐III‐R criteria for major depression (there are some bipolars but less than 20% of the total randomised sample). Age range: over 18 years.
Interventions	Paroxetine: 101 participants. Nefazodone: 105 participants. Paroxetine dose: 20‐40 mg/day. Nefazodone dose: 200‐600 mg/day. The association of short half‐time benzodiazepines was allowed for insomnia in those patients who already been receiving concomitant treatment before the study began.
Outcomes	Montgomery and Asberg Depression Rating Scale (MADRS), Hamilton rating scale for depression (HDRS‐17), Clinical Global Impression (CGI). Total dropout, dropout due to inefficacy, dropout due to side effects, number of patients experiencing at least one side effect, side‐effect profile.
Notes	Funding: nefazodone manufacturer. Adverse events reported only if occurring in >= 10% of any treatment group.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "randomly assigned", no further details.
Allocation concealment (selection bias)	Unclear risk	No information provided.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Quote: "a double dummy technique was used to maintain the blind".
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Quote: "a double dummy technique was used to maintain the blind".
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Safety analyses comprised the data from all patients who were randomly assigned to treatment and received the study medication. Efficacy analyses were performed on all patients who had received study medication and who had undergone at least one efficacy evaluation during treatment. Study endpoint: 28/101 missing from paroxetine group; 28/105 missing from control group.
Selective reporting (reporting bias)	High risk	Standard deviations of change score for depression were not reported. Adverse events were reported only if occurring in >= 10% of any treatment group.
Other bias	Unclear risk	Sponsorship bias cannot be ruled out.