Tobacco use is the leading preventable cause of cancer and is associated with increased morbidity and mortality. In spite of the fact that 70% of current smokers are motivated and try to quit, the rates of natural and aided cessation are low and range from 5 to 10%. Diary and ecological momentary assessment (EMA) methodology has become vital for capturing dynamic experiences in naturalistic settings, within and across days (e.g. smoking urges). Such experiences reflect not only the complexity of the smoking-cessation process but also appear as potential, and currently under-investigated, predictors of quit success. In this project, we address an important research question of the time-varying effects of predictors captured with EMA on a point outcome. For example, strong smoking urges prior to a quit attempt may not predict quitting success;however, experiencing them immediately after a quit attempt may be detrimental. This project aims to increase the understanding of barriers to successful quitting by developing new flexible statistical tools for exploring novel research questions on time-varying effects of smoking-related processes (intensively measured with EMA) on point-prevalence smoking outcomes and applying these tools to identify contextual and psychological barriers to smoking cessation in two completed EMA studies. This project has 3 specific aims.
Aim 1 involves developing and validating the varying-coefficient regression (VCR) model for EMA data. The model will accommodate unique features of EMA including large data volume, unequal spacing between observations, varying times of measurements across individuals, and difference in the total number of assessments per person.
Aim 2 consists of applying the model to two EMA datasets for the purpose of identifying barriers to smoking cessation and critical periods of their manifestation. Specifically, we will determine a) what smoking-related processes predict quitting success, b) how the predictive potential of these processes changes over time, and c) what periods in the smoking-cessation process are critical for determining cessation success. The high quality, well-designed studies yielding these data were previously funded by NIH and capture momentary assessments of smoking urges, negative affect, abstinence self-efficacy, and presence of other smokers, in adults who attempted natural or aided smoking cessation. As part of aim 3, user-friendly software will be developed to disseminate the work and promote use of the method by researchers in a variety of cancer-related fields. This proposal addresses the need for sophisticated statistical methods for the analysis of EMA data that are becoming increasingly prevalent in cancer research. The proposal will also answer novel and unique questions about psychological and contextual barriers to smoking cessation. This knowledge will directly contribute to improving the focus, efficacy, and cost-efficiency of cessation interventions and to alleviating the burden smoking places on society and individuals. It will also serve as a foundation of future methodological and applied work.
to public health: Ecological momentary assessments (EMA) collected via cell phones and other technologies are increasingly used to study health-related behaviors such as smoking by capturing daily and momentary experiences. The proposed work will develop a novel statistical method for analysis of EMA to predict major health outcomes from factors manifesting in daily lives. This will advance current analytical practices, which take little advantage of the detailed EMA records by collapsing them into a few summaries, and promote secondary analyses of existing EMA datasets. The empirical applications of the method will identify psychological and contextual barriers to smoking cessation, which can translate into more focused and efficacious interventions and alleviation of the burden smoking places on society and individuals.
|Huh, Jimi; Shiyko, Mariya; Keller, Stefan et al. (2015) The time-varying association between perceived stress and hunger within and between days. Appetite 89:145-51|
|Dziak, John J; Li, Runze; Tan, Xianming et al. (2015) Modeling intensive longitudinal data with mixtures of nonparametric trajectories and time-varying effects. Psychol Methods 20:444-69|
|Shiyko, Mariya P; Burkhalter, Jack; Li, Runze et al. (2014) Modeling nonlinear time-dependent treatment effects: an application of the generalized time-varying effect model (TVEM). J Consult Clin Psychol 82:760-72|
|Shiyko, Mariya; Naab, Pamela; Shiffman, Saul et al. (2014) Modeling complexity of EMA data: time-varying lagged effects of negative affect on smoking urges for subgroups of nicotine addiction. Nicotine Tob Res 16 Suppl 2:S144-50|