Meesters 1999.

Methods	Single‐centre, non‐blinded, randomised controlled trial (duration: 2 winter seasons 1993‐94 and 1994‐95, October‐April each season) conducted in the Netherlands. 8 dropouts (4 bright white visor light, 3 infrared visor light, 1 no light exposure)
Participants	46 adult outpatients with a history of SAD who were without symptoms at the beginning of the study and were free of drugs Bright white visor light group: n = 18, but participant characteristics of only 14 participants reported: 2 men, mean age 41 years (± 12.7), 12 women, mean age 39.5 years (± 9.3) Infrared visor light group: n = 18, but participant characteristics of only 15 participants reported: 5 men, mean age 35.4 years (± 6.9), 10 women, mean age 36.6 years (± 4.9) No light exposure group: n = 10, but participant characteristics of only 9 participants reported: 4 men, mean age 47.5 years (± 7), 5 women, mean age 39.4 years (± 8) No information about number of prior depressive episodes nor other participant characteristics
Interventions	Bright white visor light (n = 18; 30 minutes/d in the morning except on weekends) versus infrared visor light (n = 18; 30 minutes/d in the morning except on weekends) versus no light exposure (n = 10) from October until April
Outcomes	Development of depression (BDI ≥ 13, SIGH‐SAD‐SR ≥ 20), development of severe depression (BDI ≥ 22, SIGH‐SAD‐SR ≥ 40)
Notes	Study was not funded by pharmaceutical industry; however, equipment was sponsored by Bio Bright, Inc. Study was identified by searches of electronic databases
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No information about generation of random sequence provided
Allocation concealment (selection bias)	Unclear risk	No information about allocation concealment provided
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants were not blinded to treatment
Blinding of outcome assessment (detection bias) All outcomes	High risk	We assume that the outcome assessment was performed by the participants themselves: "patients stopped participating because of reasons unrelated to their illness, such as a lack of motivation to keep scoring self‐rating scales when in a healthy condition"
Incomplete outcome data (attrition bias) All outcomes	High risk	17% of dropouts were not taken into account in the data analysis. No participant characteristics (e.g. age, sex) were reported for these 8 dropouts
Selective reporting (reporting bias)	Unclear risk	We could not identify a protocol for this study. Therefore, we rated this domain unclear
Other bias	High risk	Intervention was implemented by participants on their own at their homes

BDI: Beck Depression Inventory
 SAD: Seasonal affective disorder
 SIGH‐SAD‐SR: Hamilton Depression Rating Scale‐Seasonal Affective Disorders self‐rating version