Sumner 2016.

Methods	Setting: Illinois and Missouri, USA; worksite employees & spouses Recruitment: Quote: "Participants called a toll free number (866‐902‐QUIT) to initiate enrolment. [...] Both organizations promoted Call‐2‐Quit through multiple channels including health fairs, employee web sites, employee news, promotional posters, fliers, and department managers. Each organization promoted Call‐2‐Quit to help smokers adapt to tobacco control policies implemented during the trial. In 2006, the hospital system implemented health insurance discounts of $10/month for employees who committed, during open enrolment in November, to pursue several health promoting activities. Smokers obtained the discount by “enrolling” in a qualifying smoking cessation program, such as Call‐2‐Quit"
Participants	518 employee and spouse smokers, 34% M, av. age 46.5, av. cigs/day 12.9. Participants were seeking treatment as they called the toll‐free number
Interventions	1. Directive telephone coaching ‐ Directive coaching included the following distinctive features: ‐ Calls scheduled about 1 week apart, except calls #4 and #7 ‐ Fixed topic schedule 2. Nondirective telephone coaching ‐ Nondirective coaching included these distinctive features: ‐ 7 calls planned over 90 days, as convenient to smoker and coach ‐ Quit date set according to individual preference ‐ Coach offers topics at each call, smoker selects 1, or may choose a novel topic There were up to 7 weekly calls, for 15 ‐ 20 mins each
Outcomes	Self‐reported abstinence at 12 m (7‐day PP) Validation: mailed saliva cotinine assays or witnessed cheek swabs attempted, but low return rate
Notes	New for 2018 update Funding: "The Centers for Disease Control grant number R01 DP000098 funded this study." Declarations of interest: none declared
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "New families were randomized to directive or nondirective coaching mode in a 1:1 ratio, based on a randomization table, when the first member enrolled. Consent to randomization was required to participate."
Allocation concealment (selection bias)	Low risk	Quote: "After baseline data were entered, members of a previously randomized family were assigned to the family coaching mode."
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Outcome validation attempted but low return rate. Similar levels of personal contact in different study arms
Incomplete outcome data (attrition bias) All outcomes	High risk	After 12 months of follow‐up the proportion lost to follow‐up was larger than half the initial sample, although similar across arms