Abdullah 2015.

Methods	Country: Shanghai, China Setting: community (households) Type: RCT
Participants	318 households with smoking parents or caregivers who had children aged 5 years or younger at home
Interventions	Intervention: • Counselling, conceptualised on the basis of the protection motivation theory developed by Rogers 1975 • Smoking hygiene intervention (SHI) with brief advice to quit SHI: • Keeping child away from household members' and other people's smoke • Avoiding smoking in the car or in closed areas near the child • Not taking the child into smoky environments • Enforcing a strict no‐smoking policy at home and in the car Control: • Placebo intervention included counselling on child development issues • No SHI or second‐hand smoke (SHS) exposure reduction or quit smoking counselling provided by the study counsellor. When queries on smoking or SHS were raised by participants, they were given the hotline number of the Shanghai CDC's smoking cessation clinic.
Outcomes	Child exposure: Primary outcomes at 6 months: • Participant‐reported improvement in smoking hygiene in the household (smoking restriction by household members at home) • Reduced exposure of child to SHS inside the home measured by mean number of cigarettes per week • Reduction in children's urine cotinine concentrations Secondary outcomes: • Total SHS exposure to child from all smokers inside and outside the home • Household members smoking cigarettes around the child • Smoking behaviour of household members (reducing the mean number of cigarettes smoked daily, making a quit smoking attempt for at least 24 hours, and quitting smoking) Child illness: Respiratory illness incidence among children as reported by key household members Target behaviour change: Secondary outcome at 6 months: • Smoking behaviour of household members (reducing mean number of cigarettes smoked daily, making a quit smoking attempt for at least 24 hours, and quitting smoking). Verified by CO measure
Type of intervention	Community‐based
Notes	Conflict of interest: none declared Source of funding: Flight Attendant Medical Research Institute (FAMRI), USA, grant 072233_CIA; and American Academy of Pediatrics, Julius B. Richmond Center of Excellence
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random numbers were computer‐generated by the project manager (not counsellors) before participant recruitment.
Allocation concealment (selection bias)	Low risk	Counsellor opened a serially numbered, opaque, and sealed envelope to reveal the random assignment of each smoker to intervention or control group.
Incomplete outcome data (attrition bias) All outcomes	High risk	Large dropout rate; more than 40% of the households in each group were not available. This was the result of many households relocating to a new residential area, farther from the original study area. Analysis does not appear to be intention‐to‐treat.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Not blinded but objective measure (cotinine)
Other bias	High risk	In addition to dropout rate: • Small possibility of cross‐contamination between intervention and comparison groups • Dosing (i.e. contact duration and frequency) of the intervention was not equal for the intervention and comparison groups • Social desirability bias due to interview format