Gritz 1992.

Methods	Setting: HMO, USA Recruitment: members of HMO agreeing to participate in a Preventive Health Behaviour study and completing a baseline survey (unselected ‐ not informed that study was focussed on smoking)
Participants	1396 female smokers, average age 38, 42% smoked 15 to 24 cpd
Interventions	No face‐to‐face contact; 5 follow‐up interviews in 2 years ∙ Self‐help programme mailed in 6 weekly instalments; manuals tailored to the concerns of female smokers and addressing weight gain, social support, stress, and coping mechanisms ∙ Control ‐ no materials; same schedule of follow‐up phone calls
Outcomes	Sustained abstinence at 1 month, 6 months, 12 months, and 18 months Validation: saliva cotinine < 15 ng/mL, but due to low success in obtaining samples, unadjusted rates used. No difference in disconfirmation rates between intervention and control groups
Notes	The strictest measure of abstinence extracted gives the lowest P value for the difference between groups; all other measures show a smaller difference in quit rates
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomised; method not described
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding (performance bias and detection bias) All outcomes	Low risk	Control group participants not aware of the nature of the intervention; participants did not know study was aimed at smoking cessation Biochemical validation conducted; not used due to low success in obtaining samples, but no difference in disconfirmation between groups, suggesting differential misreport unlikely
Incomplete outcome data (attrition bias) All outcomes	Low risk	12.7% lost to follow‐up at 18 months Number in each group at baseline not stated, so losses not included as smokers in meta‐analysis Similar losses across groups, so no effect on estimate