Smoking cessation information seeking in Türkiye following plain packaging, COVID-19, and the varenicline recall: a 10-year interrupted time series analysis

Background Türkiye became the first country to fully implement all World Health Organization (WHO) MPOWER measures and introduced mandatory plain packaging for tobacco products in December 2019. Shortly thereafter, the COVID-19 pandemic and the global recall of varenicline disrupted smoking cessation efforts. We aimed to quantify the associations of these three sequential events with smoking cessation information-seeking behavior in Türkiye over a 10-year period. Methods Weekly Relative Search Volume (RSV) data were extracted from Google Trends (Web Search and YouTube Search) for Türkiye between December 2015 and March 2026 ( N = 536 weeks). Eight of 17 candidate Turkish search terms were retained after a prespecified validation step and grouped into four thematic clusters. A multiple interrupted time series (ITS) regression with Newey–West standard errors was used to estimate level and trend changes at each event, adjusting for seasonality. Effect sizes, Granger causality tests, an exploratory counterfactual projection, and extensive sensitivity analyses were performed. Results In the primary weekly model, plain packaging was associated with a positive but statistically non-significant immediate level change (β = 20.11; 95% CI − 3.47 to 43.69; p = 0.095). In pre-pandemic sensitivity analyses, post-policy search volume exceeded seasonally matched windows of the four preceding years (Cohen’s d = 1.50; +61.8%; p < 0.001). An exploratory counterfactual projection estimated a 10,697-unit excess over the projected pre-policy trend (95% bootstrap interval 4,415–16,850). COVID-19 onset (β = −29.91; p = 0.002) and the varenicline recall (β = −11.91; p = 0.001) were associated with significant immediate declines; the recall was followed by a positive recovery trend (β = 0.165; p < 0.001). Pharmacological-cluster increases during the pandemic and after the recall were largely attributable to the dual-indication term “bupropion”. Web searches Granger-predicted YouTube searches ( p < 0.001, robust to differencing); the reverse was non-significant. Conclusions Plain packaging implementation was followed by increased cessation-related search interest, although this association cannot be fully separated from the closely following COVID-19 period. Health crises and medication supply disruptions were associated with suppressed general cessation-related searching. Because search interest reflects information seeking rather than quit attempts, the implications for primary care are hypothesis-generating.