An observational international cohort study will be conducted. A questionnaire will be administered using a snowball sampling technique, which relies on the referrals and network connections of the initial participants. Snowball sampling was selected due to its ability to efficiently recruit individuals engaged in smoking behaviors and those transitioning to alternative tobacco products, where social networks influence participation. This method also allows for streamlined recruitment without relying on preexisting registries [14].

To be eligible for inclusion, individuals must be aged between 18 and 99 years. This criterion ensures that the study population exclusively consists of adults, encompassing a broad spectrum of age groups, from early adulthood to the later stages of life. This diverse demographic range aims to capture a comprehensive representation of adult participants, allowing for a thorough examination of the intervention’s effects across various life stages. The study will be initially conducted in Italy and later in Turkey and Greece.

Participants who consent to be recontacted for inclusion in the cohort will undergo biannual assessments. To ensure continuity in the biannual assessments, participants who consent to follow-up will receive scheduled reminders via email, social media, or phone (if contact information is available). Additionally, engagement strategies such as periodic study updates will be implemented to encourage ongoing participation.

Individuals who face challenges in accessing and completing the questionnaire and those who experience difficulties reading or responding to the questionnaire will be excluded from participation. Additionally, individuals who are unable to accept the privacy and participation policy will not be eligible for inclusion in this study.

Participants who express their willingness to be recontacted will be part of the cohort for this study. The cohort group will be asked to complete a short version of the questionnaire during these follow-up assessments. This approach permits the gathering of valuable information over an extended period and assessing any changes or trends in the data that may emerge over time.

The study was carried out according to the Declaration of Helsinki guidelines and was approved by the Bioethics Committee of the University of Turin (protocol 0075071) on February 2, 2024.

All participants will provide informed consent before participation, with the option to withdraw at any time. Consent to participate and to collect and process personal data is given electronically (ticking the corresponding box) on the project website, after reading detailed information on data collection and processing, and before accessing the questionnaire, in accordance with the European Commission General Data Protection Regulation (679/2016).

Both the informed consent form for study participation and the data processing consent form have been approved by the institutional ethics committee.

Data will be fully anonymized to ensure privacy and confidentiality. No compensation will be provided for participation. No identifiable personal data will be included in the manuscript or the supplementary materials.

A feasibility assessment was performed using a fixed model multiple regression and correlation analysis approach, parameterized according to a noncentral F distribution, which can consider the effect size of the population means. This accounts for differences in expected sample size (ie, a larger effect size requires a smaller sample size to achieve the same level of power). The noncentrality parameter (λ) has been established using the following equation:

λ = f²(u + v + 1)

Where u indicates the df for the numerator (ie, the number of independent variables) and v is the df for the denominator (ie, df of error variance). This parameter is used to measure the magnitude of the treatment effect. It indicates how far the alternative hypothesis distribution is from the null hypothesis distribution. Since it is possible to describe this scenario as a function of the effect size (f²) and df, everything can be rewritten as an F test that involves the λ using the following equation [15]:

Where PV_S represents the proportion of Y (ie, a quantitative dependent variable) variance accounted for by some source (S) in the sample, and PV_E represents the proportion of error variance. Finally, the sample size has been calculated as a function of the unknown λ values in terms of the given power (1 – β), u and v values under the significance criteria of .01 and .05, respectively [16].

To account for multiplicity in answering, the sample size was adjusted by incorporating the design effect by considering the intraclass correlation (ICC) and average cohort size (m).

n_adjusted = n × {1 + (m – 1) × ICC}

The sample size calculation was based on a noncentral F distribution, incorporating the noncentrality parameter (λ) to account for the expected effect sizes. The choice of λ was informed by prior studies on the relationship between smoking behaviors, lifestyle habits, and transitions to ATNPs. Specifically, effect sizes were estimated based on the existing literature examining behavioral and health-related changes associated with smoking transitions. Given the variability in expected effects across different lifestyle factors, a conservative approach was taken in setting λ values to ensure adequate statistical power while minimizing the risk of type 2 errors. Sensitivity analyses were also conducted to assess the robustness of the sample size determination under different assumptions regarding effect sizes and ICC. These considerations strengthen the validity of our study design and ensure that the sample is sufficiently powered to detect meaningful associations.

Therefore, the final sample size is presented in Table 1, with a final choice of 9171 responses, corresponding to a reasonable and common ICC value for such situations.

Baseline and demographic characteristics of the study participants will be summarized using medians (IQRs) for continuous variables and absolute and relative frequencies for categorical ones. To examine the associations between smoking habits, dietary patterns, and lifestyle behaviors, univariable and multivariable analyses will be conducted using generalized linear models, with the appropriate distribution selected based on the outcome variables. Confounding variables such as age, gender, and socioeconomic status will be adjusted for using multivariable regression models, informed by prior literature. Interaction effects will be evaluated through interaction terms within models and stratified analyses across demographic subgroups, with likelihood ratio tests used to determine their significance. All analyses will be carried out using R software (R Foundation for Statistical Computing).

Data will be collected using a structured questionnaire administered via LimeSurvey [23]. Once the data is collected, it will be exported in a format suitable for statistical analysis. Data will be securely stored on a private password-protected server with access limited to authorized personnel only. To ensure data security, the server will use encryption protocols and role-based access controls, preventing unauthorized modifications or downloads. Data will be stored in servers located in Western Europe; the data will not be transferred outside of Europe.

Regular backups of the dataset will be made to prevent any loss of data. All data management procedures will comply with GDPR and other applicable data protection and privacy regulations.

Given the longitudinal nature of the study, participant retention and potential changes in behavior over time are key considerations. To address the issue of dropouts, participants who consent to follow-up will be recontacted biannually, with reminders sent by email and social media to encourage continued participation.