Bullock 2009 (AvC).

Methods	Randomised controlled trial (2 x 2 factorial design) evaluating nurse delivered telephone social support (‘‘Baby BEEP’’) to improve a range of maternal health outcomes, including smoking during pregnancy. Study conducted in 21 rural Women, Infant and Children Nutritional Supplement (WIC) clinics in a Midwestern state, USA, from January 2002 to July 2006.
Participants	Inclusion criteria: Women attending rural WIC clinic who reported smoking at least 1 cigarette per day, spoke English, were 18 years or older, and less than 24 weeks' gestation. Exclusion criteria: Not further specified. Recruitment: When a woman attending a WIC clinic reported current smoking, staff explained the availability of a smoking cessation study and asked permission to provide her name and telephone number to the Baby BEEP research team. If the woman agreed, a nurse from the research team was assigned to contact her to arrange a face‐to‐face visit to explain the study and request written consent. 1420 referrals from WIC clinics, 932 eligible, 695 (75%) randomised (C = 171; I1 (booklets) = 179; I2 (social support) = 175, I3 (social support + booklets) = 170. Baseline characteristics: > 90% 'ready to quit this pregnancy'. Fagerstrom scores: C = 4.8, I1 (Booklets) = 5.0, I2 (SS) = 4.9, I3 (SS + booklets) = 4.7 Mean age: 22 years, 95% white, 63% high school diploma, 70% in relationship. Psychosocial assessments indicated participants experienced high levels of perceived stress and depression and low levels of support generally and from partners. Progress + coding: Low SES as women recruited from WIC clinics.
Interventions	A: Control: Quit Smoking for Good pamphlet from the American Heart Association and instructed that a member of the research team would call each month to arrange a saliva sample, measure exposure to tobacco smoke and ask some questions for 2 more interviews. Intervention (3 arms): B: I1 Serialised Pregnancy‐Smoking Cessation Booklets (Booklets):Eight booklets comprised a program called ‘‘Stop Smoking! A Special Program for Pregnant Women’’ adapted to a 7th grade reading level. The first booklet was given to the woman at the recruitment visit without counselling, and the 7 remaining booklets were mailed at weekly intervals. C: I2 Nurse‐Delivered General Social Support (SS): scheduled weekly telephone call and 24‐hour access to the nurse for any additional social support needed. The research nurse’s role on the calls was to use empathetic listening skills and provide social, emotional and/or informational support in response to each woman’s individual needs, such as stressors she was facing and ways she could manage her stress responses. The nurses kept logs of all conversations so that they would be able to follow‐up on issues of importance on subsequent calls and as a measure of treatment integrity. All participants in these intervention study groups were encouraged to call the nurse any time they felt stressed or the need to talk, and they were also provided with a refrigerator magnet and a business card with their nurse’s first name and a toll‐free number. The nurses received 40 h of training for the telephone support intervention. Each research nurse was given information about a variety of community resources available. D: I3 SS + Booklets: This study ID included comparisons with the control group and arm C (SS). Main intervention strategy: Maternal health intervention with smoking component: Social support (single) compared to a less intensive intervention. Intensity: Frequency (C = 1, I = 6); Duration (C = 1, I = 4). Intervention provided by project staff: Efficacy study.
Outcomes	Biochemically validated point prevalence abstinence at 28‐32 weeks' gestation* (late pregnancy) and 6 weeks post‐delivery (0‐5 months postpartum*). Baseline perceived stress scale, prenatal psychosocial profile, mental health index 5; readiness to stop smoking; Fagerstrom Test for Nicotine Dependence. Subgroup analysis for patterns of quitting and associations with partner smoking.
Notes	Process evaluation to follow‐up phone calls. Low attrition rate suggested as indicator of acceptability. The sample size of the control group is split in comparison 'all interventions vs all controls' to avoid double counting of participants in analysis.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Assignments were prepared individually for each nurse, were computer generated using SAS.
Allocation concealment (selection bias)	Low risk	Opaque, sealed envelope, prepared by the principle investigator that contained the study group assignment.
Incomplete outcome data (attrition bias) All outcomes	High risk	Attrition: 9 had a spontaneous abortion (C = 2, I1 = 3, I2 = 3, I3 = 1) or non‐viable infant (C = 0, I1 = 4,I2 = 1, I3 = 4) and were excluded from the analysis in this review. Those who dropped out and were lost to follow‐up for other reasons were included in the final analysis as continuing smokers (C = 7, I1 = 11, I2 = 11, I3 = 7). However, 165 women were lost to lab error in analysing their saliva samples and were not included in analysis. Only 530/695 (76%) randomised participants were included in this analysis. C = 128, I1 = 141, I2 = 132, I3 = 129 (n = 530) included as denominators in this review.
Selective reporting (reporting bias)	Low risk	All primary outcomes reported.
Other bias	Low risk	No other bias detected.
Biochemical validation of smoking abstinence (detection bias)	Low risk	165/695 sample lost. Self‐reported abstinence in remaining women biochemically validated using salivary cotinine (30 ng/mL or less classified as non‐smokers).
Blinding of participants and personnel (performance bias) All outcomes	High risk	The nurses who collected samples when they conducted the follow‐up interviews in late pregnancy and 6‐weeks postdelivery were aware of the study group assignment.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	The laboratory was blind to study group assignment while running the cotinine analyses. The assistants who collected the monthly saliva sample may or may not have been blinded to the study group but the rule was to treat all the women the same way.
Incomplete implementation	High risk	Percentage of calls completed in each of their caseloads ranged from 58% to 80% (p400).
Equal baseline characteristics in study arms	Low risk	Characteristics appear equal.
Contamination of control group	Low risk	Care‐providers not involved in provision of the intervention.