Effect of Graphic Warning Labels on Cigarette Packs on US Smokers' Cognitions and Smoking Behavior After 3 Months: A Randomized Clinical Trial

Abstract

Importance: The US Food and Drug Administration's implementation of graphic warning labels (GWLs) on cigarette packs is under challenge in US courts.

Objective: To determine whether GWLs can affect US smokers' perceptions about their cigarettes or health consequences and changes in smoking behavior.

Design, setting, and participants: This study was a randomized clinical trial of the effect of a 3-month, real-world experience of cigarettes with GWL packaging. Community recruitment was done from September 2016 through December 2019 of daily smokers from San Diego, California, aged 21 to 65 years, who were not ready to quit. Participants were randomized to purchase and receive cigarettes in 1 of 3 pack designs: GWL, blank, or standard US pack. Data analysis was performed from July 2020 to February 2021.

Interventions: The study manufactured GWL cigarette packs (3 versions with Australian-licensed images) and packs devoid of marketing. For 3 months, participants purchased GWL, blank, or standard US pack cigarettes that were delivered to their home.

Main outcomes and measures: Smoking-related cognitions and behavior were queried by daily and weekly interactive text messages. Smoking behavior was self-reported before and after the intervention by 96% of randomized participants and was biochemically validated on a subsample.

Results: The study sample included 357 participants (195 women [54.6%]; mean [SD] age, 39.5 [11.9] years); 116 were randomized to the standard US pack group, 118 were randomized to the GWL pack group, and 125 were randomized to the blank pack group. Over the 3 months, participants who received the GWL packs had reduced positive perceptions of recent cigarettes smoked compared with participants who received the branded US pack (mean difference, -0.46 SD; 95% CI, -0.73 SD to -0.20 SD; P < .001). Health concerns increased in all groups, with a significant increase in the GWL group vs the US pack group (mean difference, 0.35 SD; 95% CI, 0.09 SD to 0.62 SD; P = .002). Quitting cognitions increased in all study groups, with a peak mean change of 0.60 SD for GWL participants vs 0.34 SD for US pack participants (mean difference, 0.55 SD; 95% CI, 0.28 SD to 0.81 SD; P < .001). GWL participants had slightly more cigarette abstinence periods per week than the US pack group, but the difference was not significant (adjusted odds ratio, 1.06; 95% CI, 0.99 to 1.13). At 3 months, there was no between-group difference in any smoking behavior. The blank pack group was similar to the US pack group on all measures.

Conclusions and relevance: These findings suggest that the introduction of GWL packs appears to decrease positive perceptions of cigarettes and increase quitting cognitions in the short term. However, additional complementary tobacco control strategies may be necessary for GWL packs to be associated with reduced smoking behavior.

Trial registration: ClinicalTrials.gov Identifier: NCT02676193.

Conflict of interest statement

Conflict of Interest Disclosures: Dr Strong reported receiving grants from the National Institutes of Health during the conduct of the study. Dr Leas reported receiving grants from National Cancer Institute during the conduct of the study. No other disclosures were reported.

Figures

Figure 1.. CONSORT Diagram for California Smokers in Australia (CASA) Study 3-Month Intervention

GWL indicates graphic warning label. aEMA refers to ecological momentary assessment by interactive daily texting to assess short-term smoking behavior and cognition.

Figure 2.. Change in Positive Perceptions of Recent Cigarettes and Perceptions of Health Effects Reported During Ecological Momentary Assessments (EMAs) by Interactive Texting

A, Positive perceptions of recent cigarettes are rated on a 4-point scale (1 = strongly disagree; 2 = disagree; 3 = agree; 4 = strongly agree) formed from 3 questions on satisfaction, craving relief, and taste of recent cigarettes assessed during daily EMAs throughout the study. Data were available for 111 participants in the US pack group, 112 participants in the graphic warning label (GWL) pack group, and 123 participants in the blank pack group and were normalized to reflect differences from an average rating during the 1-month baseline period. B, The Health Perception Scale is a 4-point scale formed from 2 questions on how the participant’s smoking impacted their health and that of others, assessed on weekly EMAs. Data were available for 109 US pack participants, 109 GWL pack participants, and 113 blank pack participants and were normalized to reflect differences from a mean rating during the 1-month baseline period. During the run-in period, all groups received the US pack. At week 0 (randomization), participants were allocated to a study pack group. The solid line denotes the mean predicted values from a mixed-effects model with compound symmetry covariance structure; run-in slope was set to 0; randomization period slope incorporates a knot at 60 days (panel A). Vertical segments are observed mean changes in scores and their 95% CIs. Details are presented in eTable 2 and eTable 3 in Supplement 2.

Figure 3.. Change in Quitting Cognitions and Abstinence Periods Reported During Weekly Ecological Momentary Assessments (EMAs)by Interactive Texting

Quitting cognitions were assessed on a 4-point scale during weekly EMAs. Weekly abstinence was the proportion of participants with at least 1 daily 4-hour interactive text period in which no smoking was reported. For both variables, data were available for 109 US pack group participants, 109 graphic warning label (GWL) pack group participants, and 113 blank pack group participants. During the run-in period, all groups received the US pack, and at week 0 (randomization [R]), participants were allocated to a study pack group for the 12-week intervention. The solid line denotes the mean predicted values from a generalized linear mixed model (panel A, normal; panel B, logistic) with compound symmetry covariance structure; run-in slope set to 0. Vertical segments are observed mean changes in scores (A) or proportions (B) and their 95% CIs. Details are presented in eTable 4 and eTable 5 in Supplement 2.

Figure 4.. Change in Cigarette Consumption Reported During Daily Ecological Momentary Assessments (EMAs) by Interactive Texting

During the run-in period, all groups received the US pack, and, at week 0 (randomization [R]), participants were allocated to a study pack group for the 90-day intervention. Daily cigarette consumption is estimated from the number of cigarettes reported smoked within the last 4 hours captured by the daily interactive text messaging responses. Data were available for 115 US pack participants, 114 graphic warning label (GWL) pack participants, and 124 blank pack participants. Solid line shows mean predictive values from a mixed effects model using compound symmetry covariance structure with run-in slope set to 0 and a knot at 60 days. Vertical segments have observed daily mean changes in number of cigarettes consumed and their 95% CIs. Details are in eTable 6 in Supplement 2.