Hattink 2015.

Study characteristics
Methods	Design: Parallel group randomised trial. No. of participating centres: Multinational study (Netherlands, UK). Study dates: May 2013 to March 2014.
Participants	Inclusion criteria: Informal caregivers, volunteers, and professional caregivers of a person with dementia, who were sufficiently computer literate to utilize the STAR website. Exclusion criteria: Not reported. No. of participants randomised to interventions: ‐ STAR intervention (n = 27 lay people ‐ informal caregivers plus volunteers). ‐ Wait‐list control group (n = 32). Baseline characteristics: Caregivers: Mean age for participants was 53.9 years (SD = 13), most were female (71%), 63% of participants were caring 2 or more years. Care recipients: The study does not provide information.
Interventions	Experimental: E‐learning tool: European Skills Training and Reskilling (STAR) project. Control: No intervention but free access to STAR after research ended. Lenght of intervention: 2 to 4 months.
Outcomes	Primary outcomes: ‐ Knowledge on dementia assessed with the 30‐item Alzheimer's Disease Knowledge Scale (ADKS). ‐ Attitudes toward dementia assessed with 2 questions from the Alzheimer's disease survey. ‐ Approaches to dementia assessed with the Approaches to Dementia Questionnaire. ‐ Usefulness and user friendliness assessed with an ad hoc 29‐item questionnaire. Secondary outcomes: ‐ Empathy assessed with the 28‐item Interpersonal Reactivity Index. ‐ Quality of life assessed with 2 distinct questions ‐ only administered to informal caregivers. ‐ Burden assessed with 1 question ‐ only administered to informal caregivers. ‐ Sense of competence assessed with the 7‐item Short Sense of Competence Questionnaire ‐ only administered to informal caregivers. Outcomes assessment: 2 to 4 months.
Notes	Clinical trial registration code: Not reported. Funding: European Union Leonardo da Vinci Life Long Learning Programme (no. 510364‐2010) and the BAVO Foundation in the Netherlands. Conflicts of interest: None declared.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Comment: Interventions were randomly assigned by strata using randomization software.
Allocation concealment (selection bias)	Unclear risk	Comment: There is no information to judge allocation concealment.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: Interventions could not be blinded. In fact, participants randomly allocated to the STAR intervention were invited to take part in a Facebook community.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Comment: Outcome assessments were based on standardized self‐assessed questionnaires delivered online but filled with participants unblinded to interventions.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: With a high total dropouts at post‐test (39%), reasons are recorded and no relevant differences appeared between completers and dropouts.
Selective reporting (reporting bias)	Low risk	Comment: All outcomes described are reported.
Other bias	Unclear risk	Comment: The study reports data grouped for informal caregivers and volunteers together as lay people because of the relatively small number of volunteers.