Abstract
BACKGROUND
Objective
To examine birth weight change caused by adding financial rewards for smoking cessation compared to no rewards for pregnant women. To estimate the average expected birth weight change for those who quit because of rewards.
METHODS
This study updates a previous systematic review and refocuses the outcome from smoking cessation to birth weight.
Eligibility Criteria
Trials with an experimental design allowing treatment effects to be attributed to rewards were included. Trials involving non-pregnant participants, or with no report of magnitude, treatment duration, timing or where most rewards were contingent on another behaviour (e.g., treatment attendance) were excluded.
Information sources
Medline, PsycInfo, Embase, Cochrane (Central Register of Controlled Trials, Tobacco Addiction Group Specialised Register and Database of Systematic Reviews), and PubMed searched to 5th December 2023.
Risk of bias
Risk of bias and certainty of evidence used Cochrane ‘Risk of bias 2’ and GRADE assessments.
Synthesis of results
Primary analysis estimated Intention-To-Treat (ITT) mean birthweight difference when randomised to offer of rewards versus control. Within-trial estimates and standard errors were derived from mean, standard deviation, and sample size data provided, or from publications. Pooled ITT estimates used common (fixed) and random effects models. Secondary analyses used trial team supplied data to derive Complier Average Causal Effect (CACE) estimate of smoking cessation on birth weight, and a standard error. Estimates were pooled using common and random effects models. Similar analyses were applied to low birth weight (<2500g), birth weight for gestational age z-scores, and small for gestational age (<10 th percentile).
RESULTS
Included studies
Primary analysis included 8 trials (2351 participants) from the UK (2 trials, 1475 participants); France (1 trial, 407 participants), and the US (6 trials, 469 participants). Secondary analysis included 7 trials as data retrieval from one US trial (51 participants) was not possible.
Synthesis of results
Primary ITT analysis (2351 participants) estimated a mean 46.3g (95% CI: 0.0 to 92.6) birth weight increase when offered financial rewards for smoking cessation. Secondary CACE analysis (2239 participants) estimated a mean 206.0g (95% CI: -69.1 to 481.1) increase for smokers who quit because of rewards. There was no effect on low birth weight (<2500g), or birth weight adjusted for gestational age, though less babies were born small for gestational age, particularly if cessation was because of rewards (CACE risk difference -17.7%; 95% CI: -34.9% to -0.4%).
DISCUSSION
Limitation of evidence
Sample size led to imprecision - maximum 2351 participants. A single trial of 3712 participants would give 80% power at 5% significance to show a 46g increase from 3.1kg to 3.146kg with 0.5kg standard deviation in both groups. Consistency - trials where smoking cessation increased (7 of 8) all showed a mean birth weight increase. In one trial smoking cessation fell as did birth weight. Bias is unlikely as 3 of 4 trials with no birth weight data showed increased cessation consistent with higher mean birth weight.
Interpretation
Trials of contingent financial rewards for smoking cessation have previously been shown to more than double pregnancy quit rates. We have uncovered a significant (46g) population level increase in mean birth weight, driven by a clinically important mean increase (206g) for those who quit because of financial rewards associated with a reduction in Small for Gestational Age births.
OTHER
Funding
Review update - The U.S. National Institute of Health, National Institute of General Medical Sciences Center of Biomedical Research Excellence Award P30GM149331. Data retrieval, synthesis and analysis – Scottish Cot Death Trust.
Registration
https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42024494262
Full Text Availability
The license terms selected by the author(s) for this preprint version do not permit archiving in PMC. The full text is available from the preprint server.