1. Characteristics of dropout studies.

Study ID	Methods	Participants	Interventions	Outcomes	Notes
Bramanti 1989	Study design: parallel design Number of arms: 2 Experimental arm: protirelin tartrate (TRH‐T) Control arm: placebo	Geographical location: Italy Setting: unclear Number of participants: 30 Stroke criteria: acute stroke Method of stroke diagnosis: not reported Inclusion criteria: not reported Exclusion criteria: not reported Depression criteria: not reported Number included in treatment group: unclear (63% male, mean age 72.2, SD not reported of the overall cohort) Number included in control group: unclear (63% men, mean age 72.2, SD not reported of the overall cohort)	Treatment: protirelin tartrate (TRHT) 2 mg/day Control: placebo Duration: 2 weeks Follow‐up: none	Depression measured using the HDRS	Results not available in format suitable for the review
Downes 1995	Study design: parallel design Number of arms: 3 Experimental arm 1: information + counselling Experimental arm 2: information pack Control arm: usual care	Geographical location: UK Setting: outpatient Number of participants: 62 Stroke criteria: not reported Method of stroke diagnosis: not reported Inclusion criteria: 1) lived at home; 2) had an informal carer; 3) stroke increase mRS; 4) post‐stroke mRS score of 2 to 5 Exclusion criteria: 1) not living at home; 2) not having an informal carer; 3) having no increase in disability or change in lifestyle/dependency Depression criteria: HADS score > 11 Number included in treatment group 1: 22 (50% men, age not reported) Number included in treatment group 2: 22 (55% men, age not reported) Number included in control group: 18 (44% men, age not reported)	Treatment 1: information plus counselling. Egan's problem‐solving approach, individual is helped to explore concerns, clarify problems, set goals, and take appropriate action. Protocol discussed first and formulated into a counsellor/client contract. Information pack containing information on physical, cognitive, behavioural and emotional effects of stroke, carer well‐being, and local services. Treatment 2: information only: information pack containing information on physical, cognitive,behavioural, and emotional effects of stroke, carer well‐being, and local services. Control: usual care, no visit(s) or information pack provided Duration: information session consisted of 1 visit and provision of the information pack. Counselling consisted of up to 8 counselling sessions over 4 to 6 months Administered by: nurse counsellor Supervision: unclear Follow‐up: none	Depression measured using the HADS‐ Depression Anxiety measured using the HADS‐ Anxiety	Unable to isolate outcome data for non‐depressed participants at randomisation
Friedland 1992	Study design: parallel design Number of arms: 2 Experimental arm: psychoeducational support Control arm: usual care	Geographical location: Canada Setting: outpatient Number of participants: 88 Stroke criteria: all subtypes Method of stroke diagnosis: via clinical signs Inclusion criteria: 1) completed formal inpatient rehabilitation and a period of rehabilitation provided by a home care programme Exclusion criteria: 1) history of psychiatric admission, 2) previously on antidepressant medication, 3) aphasia with limited ability to communicate verbally Depression criteria: unclear Number included in treatment group: 48 (44% men, mean age 69 years, SD 11) Number included in control group: 40 (44% men, mean age 69 years, SD 11)	Treatment: psychoeducational, with participant and members of their support team; work to improve social support, establish new supports, emotional support offered Control: usual care, no visits Duration: treatment continued for 6 to 12 sessions over approximately 3 months Administered by: specially trained social support intervention therapist Supervision: unclear Follow‐up: 6 months (3 months post intervention)	Depression (difference in scores at end of treatment) measured using the GHQ‐28	Results not available
Graffingo 2003	Study design: parallel design Number of arms: 2 Experimental arm: sertraline (SSRI) Control arm: matched placebo	Geographical location: unclear Setting: unclear Number of participants: unclear Stroke criteria: unclear Method of stroke diagnosis: unclear Inclusion criteria: unclear Exclusion criteria: unclear Depression criteria: unclear Number included in treatment group: unclear Number included in control group: unclear	Treatment: sertraline (SSRI) Control: matched placebo Duration: unclear Follow‐up: unclear	Depression: unclear what measure was used	Results not available
Hadidi 2014	Study design: parallel design Number of arms: 2 Experimental arm: problem‐solving therapy (PST) Control arm: weekly telephone calls	Geographical location: USA Setting: inpatient Number of participants: 22 Stroke criteria: first time diagnosis of ischaemic stroke < 48 hours Method of stroke diagnosis: not reported Inclusion criteria: 1) Mini‐Cog score of 3; ≥ 50 years of age; 2) able to read and write in English Exclusion criteria: 1) previous history of mental health problems; 2) diagnosis of severe aphasia as identified by a speech pathologist; 3) haemorrhagic stroke or transient ischaemic attack; 4) medical instability requiring transfer to critical care Depression criteria: CES‐D score measured at baseline but patients recruited regardless of their CES‐D score. If CES‐D score > 10, or suicidal ideation the primary physician was notified Number included in treatment group: 11 (18% men, mean age 73) Number included in control group: 11 (45% men, mean age 69)	Treatment: one‐on‐one problem solving therapy sessions lasting 1‐2 hours. Therapy entails providing patient information on impact and guidance to enable the patient to: identify and define the problem; brainstorm all potential solutions; select the most appropriate and feasible solution; create and implement a SMART (Specific, Measureable, Achievable, Realistic and Timely) goal; evaluate and re‐view progress in follow‐up sessions Administered by: a doctoral nursing student who received PST training through a 13‐ module online program adapted from a standard 3‐day in person training Supervision: principal investigator who had undergone in person PST training Intervention fidelity: not reported Control: weekly telephone calls to assess CES‐D and FIM scores Duration: once per week for 10 weeks Follow‐up: 3 months	Depression measured using the CES‐D Impairment measured using the FIM Leaving the trial early	Unable to isolate outcome data for non‐ depressed participants at randomisation
Kim 2017	Study design: parallel design Number of arms: 2 Experimental arm: escitalopram (SSRI) Control arm: placebo	Geographical location: South Korea Setting: inpatient Number of participants: 478 Stroke criteria: ischaemic stroke or intracerebral haemorrhage Method of stroke diagnosis: confirmed by MRI or CT Inclusion criteria: 1) > 20 years, 2) had an acute ischaemic stroke or intracerebral haemorrhage within the previous 21 days, 3) mRS score of 2 or greater at the time of screening, 4) agreed to participate Exclusion criteria: 1) history of diagnosed depression or other psychiatric diseases before the index stroke, 2) severe dementia, 3) cognitive dysfunction (stages 5‐7 of the Global Deterioration Scale), 4) aphasia, 5) on antimigraine or antiepileptic medication, 6) suicidal thoughts (a combined MADRS score > 8, 7) pregnant or lactating, 8) participation in another clinical trial Depression criteria: MADRS Number included in treatment group: 241 Number included in control group: 237	Treatment: oral escitalopram (SSRI) 10 mg/day Control: placebo Duration: 3 months Follow‐up: 6 months	Depression (frequency of moderate or severe depressive symptoms) measured using the (MADRS) Impairment measured using the NIHSS Functional capacity measured using the mRS Activities of daily living measured using the BI Adverse events	Unable to isolate outcome data for non‐ depressed participants at randomisation
Leathley 2003	Study design: parallel design Number of arms: 4 Experimental arm 1: social support Experimental arm 2: psychological support (cognitive therapy based problem solving) Experimental arm 3: social support and psychological support Control arm: usual care	Geographical location: UK Setting: outpatient Number of participants: unclear Stroke criteria: unclear Method of stroke diagnosis: unclear Inclusion criteria: unclear Exclusion criteria: unclear Number included in treatment group 1: unclear Number included in treatment group 2: unclear Number included in treatment group 3: unclear Number included in control group: unclear	Treatment 1: social support (information, practical advice, service liaison) Treatment 2: psychological support (cognitive therapy based problem solving) Treatment 3: social support and psychological support Control: usual care, no visits Duration: unclear Administered by: unclear Supervision: unclear Follow‐up: unclear	Depression (difference between groups at end of treatment) measured using the GHQ‐12 total score	Results not available in format suitable for this review
McCafferty 2000	Study design: parallel design Number of arms: 2 Experimental arm: psychosocial intervention Control arm: usual care	Geographical location: USA Setting: inpatient Number of participants: 40 Stroke criteria: unclear Method of stroke diagnosis: unclear Inclusion criteria: unclear Exclusion criteria: unclear Number included in treatment group: 20 Number included in control group: 20	Treatment: psychosocial, addresses cognitive, behavioural and family factors associated with post stroke depression Control: usual care, no visits Duration: treatment continued for 6 weeks Administered by: unclear Supervision: unclear Follow‐up: unclear	Depression measured using the GDS	Results not available
Ohtomo 1985	Study design: parallel design Number of arms: 2 Experimental arm: tiapride Control arm: placebo	Geographical location: Japan Setting: unclear Number of participants: 188 Stroke criteria: all subtypes Method of stroke diagnosis: diagnosis via clinical signs and CT Inclusion criteria: 1) > 40 years of age, high blood pressure (> 160/90 mmHg) and hypertensive changes on fundoscopy changes; 2) stable neuroleptic, minor tranquilliser, antidepressant, brain metabolic activators, cerebro‐vasodilators washed out for 3 to 7 days prior to randomisation Exclusion criteria: 1) severe aphasia; 2) severe dementia; 3) drug dependence; 4) inadequate conditions for the study Depression criteria: not reported Number included in treatment group: 141 (54% men, mean age not reported) Number included in control group: 147 (61% men, mean age not reported)	Treatment: tiapride, 75 mg daily for 1 week, dose escalation to 150 mg to 225 mg daily for 5 weeks according to clinical response Control: matched placebo Duration: 6 weeks	Depression‐ unclear what measure was used	Results not available in format suitable for the review
Ostwald 2014	Study design: parallel design Number of arms: 2 Experimental arm: home‐based psychoeducational programme Control arm: monthly mailed letter	Geographical location: USA Setting: outpatient Number of participants: 159 Stroke criteria: stroke <12 months ago, subtype unclear Method of stroke diagnosis: unclear Inclusion criteria: 1) > 50 years of age Exclusion criteria: 1) global aphasia, 2) patient or carer had comorbidity that took priority, 3) < 6 months life expectancy Depression criteria: none Number included in treatment group: 79 (69% men, mean age 67 years) Number included in control group: 80 (81% men, mean age 66 years)	Treatment: home‐based psychoeducational programme for stroke‐care‐giving dyads post‐discharge. The intervention involved home visits by advance practice nurses, occupational and physical therapists Administered by: advance practice nurses, occupational and physical therapists Supervision: not reported Control: 1 letter a month for 12 months Duration: 6 months Follow‐up: 6 months	Depression measured using the 15‐item GDS Quality of life measured using the SF‐36 Functional capacity measured using the FIM physical and cognitive subscales	Unable to isolate outcome data for non‐ depressed participants at randomisation
Raffaele 1996	Study design: parallel design Number of arms: 2 Experimental arm: trazodone Control arm: placebo	Geographical location: Italy Setting: outpatient Number of participants: 22 Stroke criteria: unclear Method of stroke diagnosis: not reported Inclusion criteria: not reported Exclusion criteria: not reported Depression criteria: ZDS Number included in treatment group: 11 (45.4% men, mean age 69.5, SD 2.3) Number included in control group: 11 (72.7% men, mean age 70.4, SD 3.0)	Treatment: trazodone 300 mg/day Control: placebo Duration: 30‐45 days Follow‐up: unclear	Depression measured using the ZDS Activities of daily living measured using the BI	Unable to isolate outcome data for non‐ depressed participants at randomisation.

BI: Barthel Index

CES‐D: Center for Epidemiological Studies‐ Depression

CT: computed tomography

FIM: Functional Independence Measure

GDS: 15‐item Geriatric Depression Scale

GHQ‐12: 12 item General Health Questionnaire

GHQ‐28: 28 item General Health Questionnaire

HADS: Hospital Anxiety and Depression Scale

HDRS: Hamilton Depression Rating Scale

MADRS: Montgomery‐Asberg Depression Rating Scale

MRI: magnetic resonance imaging

mRS: modified Rankin Scale

NIHSS: National Institutes of Health Stroke Scale

SD: Standard Deviation

SF‐36: 36‐item Short Form Questionnaire

SSRI: selective serotonin reuptake inhibitors

ZDS: Zund Depression Scale