1. Quality Rating Scale items.

No	Item___________________________________________	Components_________________________________________
1	Objectives and specification: Were main outcomes established a priori?	0 = Objectives unclear 1 = Objectives clear but main outcomes not specified a priori 2 = Objectives clear with a priori specification of main outcomes
2	Adequacy of sample size: Were there enough completers in each group?	0 = No/don't know 2 = Yes
3	Planned duration of trial including follow up?	0 = < 3 months 1 = > 3 months < 6 months 2 = > 6 months
4	Method of allocation	0 = Not randomised and likely to be biased 1 = Partially or quasi randomised with some bias possible 2 = Randomised allocation
5	Concealment of allocation	0 = Not done or not reported 1 = Partial concealment reported 2 = Done adequately
6	Clear description of treatment (including doses of drugs) & adjunctive treatment	0 = Main treatments not clearly described 1 = Inadequate details of main or adjunctive treatments 2 = Full details of main and adjunctive treatments
7	Blinding of subjects	0 = Not done 1 = Blinded but no test of blinding 2 = Blinded and integrity of blinding tested
8	Source of subjects described and representative sample recruited?	0 = Source of subjects not described 1 = Source of subjects but unrepresentative sample e.g. in‐patients/specialist settings 2 = Source of subjects described plus representative sample
9	Use of diagnostic criteria (or clear specification of inclusion criteria)	0 = None 1 = Diagnostic criteria or clear inclusion criteria 2 = Diagnostic criteria + specification of severity
10	Record of exclusion criteria and number of exclusions and refusals reported?	0 = Criteria and number not reported 1 = Criteria or number of exclusions & refusals not reported 2 = Criteria and number of exclusions and refusals reported
11	Description of sample demographic characteristics?	0 = Little/no information (only age/sex) 1 = Basic description (e.g. marital status/ethnicity) 2 = Full description (e.g. socio‐economic status/clinical history)
12	Blinding of assessor	0 = Not done 1 = Blinded but no test of blinding 2 = Blinded and integrity of blinding tested
13	Assessment of compliance with experimental treatments (including adherence to therapy)	0 = Not assessed 1 = Assessed for some experimental treatments 2 = Assessed for all experimental treatments
14	Details on side effects	0 = Inadequate details 1 = Recorded by group but details inadequate 2 = Full side effect profiles by group
15	Record of number and reasons for withdrawal by group	0 = No information on withdrawals by group 1 = Withdrawals by group reported without reason 2 = Withdrawals and reason by group
16	Outcome measures described clearly or use of validated instruments	0 = Outcomes not described clearly 1= Some outcomes not clearly described 2= Outcomes described or valid & reliable instruments used
17	Information on comparability and adjustment for differences in analysis	0= No information on comparability 1= Some info on comparability with appropriate adjustment 2= Sufficient comparability info with appropriate adjustment
18	Inclusion of withdrawals in analysis (ITT or endpoint)	0 = Not included or not reported 1 = Withdrawals included in analysis by estimation of outcome 2 = Withdrawals followed up and included in analysis
19	Presentation of results with inclusion of data for re‐analysis of main outcomes (e.g. SDs)	0 = Inadequate presentation 1= Adequate 2 = Comprehensive
20	Appropriate statistical analysis (including correction for multiple tests where applicable)	0 = Inappropriate 1 = Mainly appropriate 2 = Appropriate and comprehensive
21	Conclusions justified	0 = No 1 = Partially 2 = Yes
22	Declaration of interests (e.g. source of funding)	0 = No 2 = Yes
	Notes: 1= Details on how the allocation code was protected from those involved in patient recruitment may be achieved by having allocation done by a central independent body, or protection of code (e.g. sealed opaque envelopes). 2= Source of subjects refers to the setting in which subjects were found (e.g. inpatients, outpatients, general practice, community etc). 3= Test of integrity of blinding is normally done by asking participants to guess their allocated group. Results can be compared to those which would be expected by chance. 4= Whether or not the decision to initiate an antidepressant was based strictly on the primary care practitioners judgment that there was clinical depression warranting treatment rather than insisting that criteria for a specific diagnosis, such as major depressive disorder, be established.