Bahmer 1992.

Study characteristics
Methods	Design: randomised, placebo‐controlled, parallel Duration: 3 months Interval of assessment: baseline, 6 weeks, 3 months This was a randomised controlled trial
Participants	Number randomised: 12 (7 in the active group and 5 in the placebo group) Sex (M/F): 6/6 total (3/4 in the active group and 3/2 in the placebo group) Age of participants: active group = 31 (mean), placebo group = 27 (mean) Unit of allocation: whole person Country and setting: Saarlandes, Germany; single university dermatology clinic Inclusion criteria of the study (specified) Participants having atopic dermatitis Having used topical steroids Diagnostic criteria: not reported Severity of condition: "mild to moderate", glandol group ADASI score = 1.63, control group ADASI score = 1.89 Exclusion criteria of the study Participants using systemic steroids or antimetabolites
Interventions	Treatment group: glandol (BO), 3 capsules twice daily, each capsule 500 mg Placebo group: palm kernel oil, 3 capsules twice daily, each capsule 500 mg
Outcomes	ADASI score by clinician (method of scoring by mapping) Skin condition by participant (method of assessment subjective: participant diary) Serum lipids (lab)
Notes	Previous treatment was not stopped ("no limitations regarding other treatment were required") Assessment of compliance was not undertaken
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The trial mentioned randomisation, but did not give a specific method
Allocation concealment (selection bias)	Unclear risk	The trial did not give a method for allocation concealment
Incomplete outcome data (attrition bias) All outcomes	Low risk	There were no dropouts or losses to follow up (7/7 glandol and 5/5 placebo completed)
Selective reporting (reporting bias)	Low risk	The trial reported all outcomes
Other bias	Low risk	All randomised participants were included in the analysis in the groups to which they were randomised. We contacted the author
Blinding of participants and personnel (performance bias) All outcomes	Low risk	This was adequate as both the participant and assessor were blinded