Fowler 2005.

Methods	Within‐participant, randomised controlled trial of 3 parallel groups This was a multi‐centre study conducted in the USA and carried out in a secondary care setting
Participants	86 participants with chronic hand eczema Dropouts: 4 Inclusion criteria of the trial Between 18 and 65 years old Symmetrical hand or atopic dermatitis of moderate severity for at least 2 weeks Exclusion criteria of the trial Use of systemic treatments in the last month or topical corticosteroids in the last week before study entry Study population Gender: 52 female, 34 male Age: mean 46 years
Interventions	Intervention • Hydrocortisone butyrate 0.1% cream on the one hand vs fluticasone propionate 0.05% cream twice daily on the other hand for 2 weeks in 26 participants • Hydrocortisone butyrate 0.1% cream on the one hand vs prednicarbate emollient 0.1% cream twice daily on the other hand for 2 weeks in 28 participants • Hydrocortisone butyrate 0.1% cream on the one hand vs mometasone furoate 0.1% cream twice daily on the other hand for 2 weeks in 31 participants Duration 2 weeks
Outcomes	Primary outcomes of the trial Not defined Other outcomes Investigator‐rated severity of hand eczema on a 4‐point scale (0 = none, 1 = mild, 2 = moderate, 3 = severe) for 4 clinical signs (erythema, cracking/ fissuring, scaling, papules/vesicles) Investigator‐rated severity total sum score Participant‐rated severity of hand eczema on a 4‐point scale (0 = none, 1 = mild, 2 = moderate, 3 = severe) for 6 clinical signs (erythema, cracking/ fissuring, scaling, papules/vesicles, pruritus, burning/pain) Participant‐rated severity total sum score Investigator‐rated mean reduction in percentage of hand involvement Participants‐rated preference and cosmetic acceptability Adverse events
Notes	Three participants with atopic dermatitis participated in the study. Percentage of hand involvement was the only outcome whereby exclusively participants with hand eczema were analysed. Each intervention group had a within‐participant design. in addition, the difference in efficacy between the 3 groups was evaluated The study did include a participant‐ and investigator‐rated severity score, but we were unable to use the data. The study also included adverse events, but we were unable to use this information because only numbers for both treatment groups combined were stated Study authors were contacted on 4 March 2014 and replied 6 March 2014 Declarations of interest: 2 study authors acted as consultants Funding: the study was funded by Ferndale Laboratories, Inc., manufacturer of the study drugs. Two study authors were investigator and consultant for Ferndale Laboratories, Inc. Sample size rationale: not stated; personal communication clarified this was not conducted
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "patients were randomised in balanced cohorts to 3 parallel treatment groups" Comment: no further details given in the article; personal communication with the study author clarified that a computer programme was used to create the randomisation code
Allocation concealment (selection bias)	Low risk	No details in the article about how allocation was concealed from participants and investigators. Personal communication revealed that allocation was conducted by the sponsor, who was at a remote site. Participants were enrolled without knowledge of the expected treatment group
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "the medication[s] were dispensed to the subjects in blind‐labelled tubes that were clearly marked with the subject's identification number and the word left or right" Comment: study authors state double‐blind design. The sponsor and the study co‐ordinator had access to the randomisation code list; treating physicians and participants were unaware of this
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Double‐blind study, which includes observer blinding. Observers had no access to the randomisation code and were truly blinded (personal communication)
Incomplete outcome data (attrition bias) All outcomes	Low risk	No intention‐to‐treat analysis but per protocol (82 of 86 = more than 80%)
Selective reporting (reporting bias)	Low risk	No trial registration found; however all outcomes described in the Methods section are described in the Results
Other bias	Low risk	Baseline comparisons: no significant differences among the 3 groups in terms of age, gender, race, and eczema severity Diagnostic certainty: yes The study was completed