A Cognitive Behavioral Stress Intervention for Women Who Smoke

Objectives Aim 1: Examine the effect of using cognitive reappraisal and HRVb in a single session on stress task performance. Investigators expect that participants assigned to practice both cognitive reappraisal and HRVb will demonstrate greater cognitive performance and persistence on stressful tasks during a laboratory visit in contrast to a control group of female smokers. Aim 2: Examine the effect of using cognitive reappraisal and HRVb on short-term HRV, craving, and affect. Investigators expect that participants assigned to practice both cognitive reappraisal and HRVb will exhibit the greatest HRV adaptations, and demonstrate less reactive craving and affect by the end of the visit than participants randomized to a control group. III. Background and Rationale Twenty percent of all deaths in the United States, or more than 480,000 deaths each year, are attributable to smoking (USDHHS, 2014). Cigarette smoking remains the leading preventable cause of death, with 16.8% of adults currently smoking (CDC, 2015) and over 16 million Americans living with a smoking-related disease (USDHHS, 2014). Smoking remains difficult to quit; even with the best available current treatments, quit attempts are successful 35% of the time or less (Garrison \& Dugan, 2009). Smoking cessation and relapse to smoking is by far the most frequent path of cigarette use for smokers (Garcia-Rodriguez et al., 2013; Piasecki, 2006; Rafful et al., 2013). Stress has been implicated as a primary mechanism in smoking relapse (McKee et al., 2003; Baer et al., 1989; Cohen \& Lichtenstein, 1990), often triggering increases in negative affect (Shiffman, 2005; Shiffman \& Waters, 2004) and exposing difficulties in emotion regulation (Farris, Zvolensky, \& Schmidt, 2015). Evidence suggests that stress prospectively predicts smoking lapse (Shiffman \& Waters, 2004) and lapses triggered by stress progress more quickly to relapse (Shiffman et al., 1996), suggesting deficits in the ability to cope with stress. In particular, women's smoking habits show trends that are different from men's: reports have shown that women consume more cigarettes than men (Hammond, 2009; Ng et al., 2014) and are less likely to successfully quit smoking than men (Cepeda-Benito et al., 2004; Perkins, 2001; Piper et al., 2010). While stress has been implicated in relapse for all smokers, recent evidence indicates that stress is a principal factor in promoting relapse to smoking in females, in part because women appear to be more strongly predisposed to stress responses (Torres \& O'Dell, 2016). It is therefore likely that women are at particular risk for relapse due to stress. Cognitive reappraisal has demonstrated positive effects on stress responding and smoking-related outcomes. Cognitive reappraisal is the reframing of a situation in order to influence one's emotional response to it (Gross, 1998). Compared with acceptance and suppression, reappraisal as measured by the Emotion Regulation Questionnaire (ERQ; Gross \& John, 2003) is associated with overall lower craving and negative affect during craving inductions and stress tasks, as well as improved performance on a cognitive stress task (Szasz, Szentagotai, \& Hofmann, 2012). Fucito, Juliano, and Toll (2010) found that frequent reappraisal on the ERQ was cross-sectionally associated with smoking fewer cigarettes. Recent evidence has indicated that reappraising stress-related arousal improves cognitive performance and physiological reactivity (Jamieson et al., 2012; 2013). Evidence thus far suggests that the use of cognitive reappraisal in particular as a self-regulation strategy may have positive implications for smoking behavior, including cigarette craving. HRVb has also shown promising evidence on stress outcomes. Utilizing behavioral self-regulation strategies such as heart rate variability biofeedback (HRVb) to increase heart rate variability (HRV) shows promise in coping with stress. HRV is a measure of fluctuation from the mean heart rate, representing the interaction between sympathetic and parasympathetic influences on the cardiac system (Siepmann et al., 2008), and serving as a biomarker of autonomic nervous system functioning (Zucker et al., 2009). HRV biofeedback training aims to improve adaptability to and recovery from fight or flight situations by increasing HRV (Gevirtz, 2013). A single session of HRVb improves adaptability to stress as measured with improved cognitive performance in a lab-induced stress task (Prinsloo et al., 2010). Several weeks of HRVb practice have shown reductions in substance craving (Penzlin et al., 2015; Eddie et al., 2014) and food craving (Meule et al., 2012), although no studies to date have examined the effect of a single session of HRVb practice on cigarette craving. This study assesses whether combining a cognitive and behavioral approach to improve stress responding in smokers may be especially efficacious. No other study to date has examined examined potential additive benefits of combining both strategies. Combining both cognitive and behavioral self-regulation skills for stress adaptation is consistent with cognitive-behavioral theory upon which evidence-based treatments for Axis I pathology were developed, as behavioral (e.g., HRVb) and cognitive (e.g., reappraisal) approaches to self-management allow the individual to target both internal and external stimuli when coping with stress (Rokke \& Rehm, 2001). Beyond psychological effects of stress, the combination of such approaches has implications for informing smoking cessation treatments, as it has been previously noted that current treatments for substance addiction "are failing to address important factors that are active in sustaining \[such\] pathology, because phenomena that lead to relapse... are mediated by physiological as well as cognitive processes" (Eddie, Vaschillo, Vaschillo, \& Lehrer, 2015, p. 266). In order to further improve cessation rates, cognitive reappraisal instruction and HRVb can be incorporated at low cost into standard smoking cessation treatment, which currently consists of pharmacotherapy and smoking cessation counseling according to the Clinical Practice Guideline (Fiore et al., 2008). HRVb technology ranges from free smartphone applications (e.g., Elite HRV) to sophisticated devices available for purchase online (e.g. EmWave), making it an affordable and feasible addition to enhance current smoking cessation treatment. As no studies to date have explored the effects of HRVb in a single session on smoking craving, this study will be the first to address the potential use of a relatively new technology for its short-term effects on variables that increase lapse risk, such as affect, craving, and stress responding. This pilot study will assess the efficacy of practicing HRVb using a device that is available to the general public that has not been previously tested for its short-term effects on craving. Combining HRVb training prior to a stressful task, followed by cognitive reappraisal during a task, has also never been explored for its additive effects above the use of either strategy alone. Both strategies would allow for a more comprehensive approach to managing stress responding, both cognitively and physiologically. Moreover, promising research on self-regulatory strategies like cognitive reappraisal has not compared it with HRVb to assess relative efficacy in improving stress responding, or in assessing its relative efficacy in improving smoking-related outcomes. IV. Procedures 1. Research Design This study is a single-session experimental trial using two interventions (heart rate variability biofeedback and cognitive reappraisal of stress) as independent variables and examining smoking craving, affect, and heart rate variability as dependent variables. The laboratory visit includes a baseline questionnaire, pre-test and post-test measures of smoking craving, affect, and heart rate variability. Participants will be randomized to practice both interventions, or to a control group that undergoes all procedures excluding the two interventions (independent variables). 2. Sample Participants will consist of 60 English-speaking females ages 18-65. As this is an experimental study with both between-subjects and within-subjects measurements, 60 participants will be sufficient for a power level of .8. 3. Measurement/Instrumentation Objective Measures Heart rate variability will be measured as one of the primary outcome measures of this study. As heart rate variability is susceptible to several other environmental and mental factors that affect the sympathetic nervous system, additional measures need to be taken to ensure the validity of the heart rate variability data. As secondary measures, blood pressure and breath carbon monoxide will be recorded. Specifically, the following measures will be used for data collection, all of which are non-invasive: * Heart Rate Variability (HRV). HRV will be examined using electrocardiogram (ECG) measurement with Biopac Acqknowledge software. Electrodes will be placed on the torso of each participant by a trained, same-sex research staff member in order to detect heart rate, cardiac impedance, and respiration rate. HRV data cleaning will follow standard guidelines (Bernston et al., 1990). * Blood Pressure (BP). An inflating arm cuff will be used to detect blood pressure. * Respiration. To control for compensatory breathing that may lead to hyperventilation when reducing breathing rate during biofeedback practice, respiration belts will be attached around participants' ribcage to monitor breathing rate and tidal volume. 800mL spirobag calibration will follow standard guidelines, asking participants to inflate and deflate the bag five times in order to obtain an accurate estimate of tidal volume. * Carbon monoxide (CO). A standard carbon monoxide monitor will be used to detect carbon monoxide levels in participants' breath, as an objective measure of cigarette smoking. This will be used for eligibility purposes (establishing that participants are not light smokers) and to ensure that participants have been abstinent from smoking for at least 12 hours prior to the start of the study visit. Instrumentation One of the interventions examined in this study is heart rate variability biofeedback, which involves breathing at a slow and steady rate and receiving feedback from a device about one's heart rate. The EmWave2® HRVb device (HeartMath®, Boulder Creek, Colorado) will be used for HRVb practice and will serve as an additional measure of HRV for the HRVb group. The EmWave2® device trains the user in HRVb and provides readings of HRV, time elapsed, and varying levels of physiological coherence (Edwards, 2014). Repeated use of the device has demonstrated clinically significant improvement in performance on tests of executive functioning associated with emotional dysregulation (O'neill \& Findlay, 2014) and statistically significant increases on measures of general health and mindfulness (Edwards, 2014). Baseline Self-Report Measures After consenting procedures are complete, participants will be directed to a website (Qualtrics) where they will complete a battery of baseline self-report questionnaires. Qualtrics is an online-based, private research company that specializes in data collection. All data entered through Qualtrics will be connected only to participants' unique identifying numbers that will not be associated with any identifying information (e.g., name or contact information). Participants will fill out self-report questionnaires that assess demographic information, smoking history and behavior, clinical symptoms, emotion regulation, and other mental health-related indices. (See Attachment 7 for the study measures.) Specifically, the following measures will be administered: * Demographics. General demographic information will be collected, including sex/gender, age, height, weight, hormonal contraceptive use, menstruation pattern, sexual orientation, race/ethnicity, education/career history, and mental health history. Additionally, several items assessing DSM-IV criteria for bulimia nervosa will be included. * The Timeline Followback (TLFB). The TLFB will be used to quantify daily smoking behavior over the past 30 days, as a measure of smoking heaviness. It has demonstrated high test-retest reliability for 30-day intervals in cigarette smokers (Robinson et al., 2014). * The Fagerstrom Test for Nicotine Dependence (FTND). The FTND is among the most widely used measures of nicotine dependence, with confirmed reliability in different settings and populations that smoke cigarettes (Agrawal et al., 2011; Fagerstrom, 2012). * The Reasons for Smoking Questionnaire (RFS). The RFS has been used for over 30 years to assess smokers' motives for smoking, and has demonstrated good convergent validity and internal consistency (Currie, 2004) and adequate test-retest reliability (Tate, Schmitz, \& Stanton, 1991). * The Beck Depression Inventory-II (BDI-II). The BDI-II is a 21-item measure of depression that has demonstrated reliability and stability (Beck, Steer, et al., 1988; Beck, Steer, \& Brown, 1996). * The Emotion Regulation Questionnaire (ERQ). The ERQ is a validated measure of emotion regulation strategies with good test-retest reliability (Gross \& John, 2003). * The Perceived Stress Scale (PSS). The PSS is the most widely used measure of the perception of stress, and was designed for use in community samples (Cohen, Kamarck \& Mermelstein, 1983). Higher PSS scores are associated with failure to quit smoking, greater vulnerability to depression, and more colds (Cohen et al. 1988). * The Generalized Anxiety Disorder-7 (GAD-7). The GAD-7 is a brief, validated measure of generalized anxiety with demonstrated reliability that has been recommended for clinical practice and research (Spitzer et al., 2006). * The Social Desirability Scale (SDS). The SDS-16 measures social desirability, which is helpful in studies with objective and subjective outcomes with potential inconsistency. * The Difficulties with Emotion Regulation Scale (DERS). The DERS will be used to measure problems with emotion regulation within six domains: nonacceptance of emotional responses, difficulties engaging in goal-directed behavior, impulse control difficulties, lack of emotion awareness, limited access to emotion regulation strategies, and lack of emotion clarity. It has been shown to have good internal consistency, validity, and test-retest reliability (Gratz \& Roemer, 2004). * The Ruminative Responses Scale (RRS). The RRS will be used to measure the degree to which individuals respond in ruminative ways during depressed moods. It has been demonstrated to have strong inter-item consistency (Treynor, Gonzalez, \& Nolen-Hoeksema, 2003). * The Emotional Cascade Assessment Measure (ECAM). The ECAM will be used to measure how individuals respond to upsetting situations, and the degree to which they engage in emotional cascades (Selby \& Joiner, 2009). This measure is currently being validated in our lab. * The Three-Factor Eating Questionnaire (EQ). The EQ will be used to assess overconsumption of food as a response to emotional, cognitive, and social cues, as food cravings and weight gain are common concerns for smokers who want to quit. Several subscales have been shown to have adequate test-retest reliability, and the disinhibition subscale predicts weight gain (Cuntz et al., 2001; Foster et al., 1998). The cognitive restraint subscale has demonstrated construct validity (Karlsson et al., 2000). The BDI-II includes an item that measures an individual's suicidality. If at any point a participant indicates they are feeling significantly suicidal, a message will pop-up at the end of the survey, informing the study staff of the individual's risk for high suicidality. Study staff will then perform a brief assessment and intervention to determine if further professional attention is required to keep the participant safe. (See "Protection Against Potential Risks" and Appendix 2 for more detailed information.) Additional Self-Report Measures At multiple points throughout the study, additional self-report measures will be used to assess smoking craving and affect, which are the subjective outcomes of the study. * The Positive and Negative Affect Schedule (PANAS) is a 20-item measure of affect that has demonstrated test-retest reliability, as well as good convergent and divergent validity (Watson, 2000). * The Questionnaire on Smoking Urges-Brief (QSU-B; Cox, Tiffany, \& Christen, 2001) is a 10-item measure of smoking urges and cravings, where a higher score indicates a stronger craving. Finally, several subscales of the Stress Appraisal Measure (SAM; Peacock \& Wong, 1989) will be used as a manipulation check for the cognitive restructuring intervention. The SAM has demonstrated reliability and good internal consistency of its subscales (Carpenter, 2016). Computer Stress Tasks Participants will first have ten minutes to complete a Mirror Tracing Task (MTT; Quinn, Brandon, \& Copeland, 1996) that measures persistence on a stressful task, which is associated with nicotine dependence treatment outcome (Brandon et al., 2003) and can distinguish cigarette smokers from non-smokers (Quinn et al., 1996). The MTT involves tracing three shapes displayed on a computer screen, using the mouse of the computer screen. A buzzing sound indicates whether the shape is being traced incorrectly. As the primary purpose of this task in the study is to induce stress, participants will be instructed to complete the task without the option of termination for ten minutes. Second, participants will complete the validated Montreal Imaging Stress Task (MIST; Dedovic et al., 2005) which involves mental arithmetic with negative feedback from the investigator, who will be a female member of study staff. The MIST was specifically designed for eliciting psychosocial stress when participant movement is restricted and has demonstrated effects on biological indices of stress, including cortisol (Dedovic et al., 2005) and brain areas associated with the limbic system (Dedovic et al., 2009) and HPA axis (Zschucke et al., 2015). Third, a computerized Stroop task (Stroop, 1935) will be implemented to induce mental stress. The Stoop task is a well-known and validated measure of mental stress and cognitive performance that involves reading words and identifying colors on a computer screen. e. Detailed study procedures Interested participants will complete a phone interview for eligibility. Participants referred from the Tobacco Dependence Program will have provided a breath carbon monoxide sample meeting eligibility criteria, which is at least 15ppm. Participants recruited from other sources will be asked to provide a breath sample during a 5-minute visit to the laboratory in order to confirm smoking status at least one day prior to the scheduled study visit. Eligible participants will be scheduled for a 2.5-hour visit and asked to refrain from alcohol, tobacco, or e-cigarette use for 12 hours prior to their appointment time. The study visit will consist of a breath CO test to confirm tobacco abstinence in the past 12 hours (indicated by a breath CO level of less than 10ppm. Evidence shows that daily smokers who have not yet smoked the first cigarette of the day exhibit expired CO levels between 9 and 11ppm (Adan, Prat, \& Sanchez-Turet, 2004). Interested participants who are not CO-eligible will be rescheduled for another study visit day. CO confirmation of study eligibility will be followed by the informed consent process, and finally the completion of a baseline questionnaire packet (detailed below). They will also be instrumented with the BioPac psychophysiology stations by a trained female study staff member. They will then be asked to conduct a baseline psychophysiological assessment for 10 minutes, which is done in order to provide a comparison for the effects of the interventions and stress tasks. Participants will be randomized to one of two conditions, blocked by age and level of motivation to quit smoking in the next 30 days. Participants ages 18-39 will be randomized separately from participants ages 40-65. Blocking randomization by age group will reduce age confounds within the sample of adult women, as age is correlated with smoking behavior and heart rate variability (Zhang, 2007). Motivation to quit smoking in the next 30 days, on a scale of 1-10, will be assessed in the initial phone interview to block randomization by level of motivation to quit (1-5 versus 6-10). Participants will be randomly assigned to one of two study conditions reflecting the presence or absence of the interventions: 1) HRVb and cognitive reappraisal instructions, or 2) Sitting quietly while engaging in a neutral control task and no reappraisal instructions. HRVb will involve paced breathing as guided by a moving light on a portable biofeedback device (described below). The neutral control tasks will involve pressing a computer key whenever a shape appears on the computer screen, which will occur approximately every minute. To maintain a neutral control condition for 30 minutes, a neutral nature video will supplement the task to prevent significant participant boredom or frustration. Cognitive reappraisal instructions will explain the function of the stress response as an adaptation to effectively overcome challenges, following an example by Jamieson, Nock, and Mendes (2012). Remaining study visit procedures are detailed in Figure 1. Debriefing Following the completion of computer tasks, 15 minute recovery period, and removal of psychophysiological recording equipment, participants will be debriefed about the study and its purposes. They also will be provided with a Debriefing/Resources Sheet with information and contact information about mental health. Participants will then be compensated for their participation. The study staff member will review the Resources Sheet with the individual and provide the participant with a physical copy to take home with them. The Resource Sheet (attachment 9) will include information on resources specific to Rutgers and central New Jersey. Protections Against Potential Risks Every effort will be made to minimize any risks associated with participation. Our procedures are generally non-invasive and pose little to no physical risk to the participants. As such, the risks involved with participation in this study are minor, though there are a few concerns. 1. Questionnaires. Some of the content of the questionnaires for this proposed project may be distressing for participants to answer (e.g. questions about substance use, self-injury, etc.), although risks from their use is likely to be uncommon as these measures have commonly been used in previous studies without incident. Though distress while answering questions from self-report measures is not extremely likely, participants will be informed that they are not required to answer any questions that they do not wish to answer, and that they may end their participation in the study at any time. 2. Suicide Risk. Although the risk of suicidal participants in this study is relatively low, it is imperative that any participant identified at high suicide risk, when answering the BDI-II during the study visit, be approached with appropriate evaluation and intervention procedures. Participants whose risk is determined to be serious (defined as endorsement of suicide plans or preparation, or imminent suicidal intent) will be referred to appropriate mental health providers for suicidal crisis intervention. If a participant endorses imminent suicide risk during an assessment, he or she will be voluntarily escorted to Rutgers University Counseling Center for further assessment. If necessary, Rutgers Campus Police will be contacted for transport of the participant for further mental health assessment, particularly in cases where imminent suicide risk is indicated and the participant refuses to undergo additional screening. The suicide risk assessment procedures, as well as the appropriate actions to be taken according to suicide risk are described in the Appendix 2. These potential actions will also be described in the informed consent form so that participants are aware prior to the clinical interviews. In the case that a participant exhibits suicidal ideation, the assessor will contact Dr. Selby to ensure that appropriate safety measures are taken prior to the participant leaving the lab. Importantly, research has found that conducting a suicide risk assessment on people who are not suicidal will not cause them to experience suicidality, and furthermore those who are experiencing suicidal ideation frequently report feeling better following a suicide risk assessment. This assessment approach has been used successfully in previous research to protect human subjects who were potentially experiencing suicidal ideation. 3. Confidentiality. Procedures will be completed once for each participant throughout the project period. There is a minimal risk that confidential information could be breached. The identifying information of all participants will not be included in the same survey as the self-report questionnaires or in the same file as their consent forms (for in-person participants). Confidentiality will be maintained in the reporting of all results. The information gathered will be used only for scientific purposes and will be reported only in aggregate form of group totals or averages. All physical data forms (from in-person participants) will be kept in a locked file cabinet in the primary investigator's lab. All personnel will be carefully trained by the investigators regarding the collection and storage of sensitive information. Data entered into statistical databases will be kept in investigators' lab space, and will be kept in locked space to which only investigators and authorized personnel will have access. All electronic data downloaded from participant surveys will be stored on a secure server in the investigator's lab (further details described below). 4. Data Security. Participants will provide electronic data via online surveys as part of their participation in the laboratory. There is small risk that data could be compromised. Each questionnaire survey and electronic data set will be numbered, and the in-person informed consent form will have an attached sheet containing that number, which will be removed once data collection is complete. f. Consent Procedures Participants will be provided with the consent form and given time to read it, then the study will be explained by a trained, CITI-certified study staff member who will also answer participants' questions. Participants will initial all pages of the consent form before signing and dating the consent form with a study staff member witness. A dated and signed copy will be given to the participant. g. Internal Validity Threats to internal validity in the study have been considered and addressed through the inclusionary and exclusionary criteria for the study. Furthermore, all participants are adult women to reduce age and gender confounds in HRV data, and randomization to study condition will be blocked by age to further reduce confounds for HRV measurement. Additionally, blood pressure will be recorded to assess potential HRV confounds. During the intervention, the biofeedback device will serve as a secondary measure of HRV and a manipulation check. Cognitive reappraisal instructions will include practice exercises and questions to ensure understanding of the intervention, and subscales from the Stress Appraisal Measure will be utilized as manipulation checks. Self-report measures have been selected for their validity and appropriateness for the study sample. Furthermore, the stress tasks were specifically selected for their effects on different facets of stress, including social stress, in order to capture the potential stressors faced by real-life smokers that lead to relapse. h. Data Analysis In order to examine the effect of practicing cognitive reappraisal and HRVb in a single laboratory session on stress task performance, investigators will examine persistence on the MTT and performance on the Stroop task separately. Persistence on the MTT will be measured via time to completion or termination of the task as an outcome measure. Reaction time and response accuracy (as averages for each individual) will be examined as outcome measures on the Stroop task. Generalized estimated equations analyses will examine the effects of the two experimental manipulations (HRVb and cognitive reappraisal) and covariates including age, average daily cigarettes smoked, Body Mass Index (as calculated from demographics information in the baseline questionnaire), and demographic variables on the three outcome measures for the MTT and Stroop task. In order to examine the effect of practicing cognitive reappraisal and HRVb on short-term HRV, craving, and affect, investigators will first clean all HRV data following standard guidelines (Bernston et al., 1990). Within-subject changes in HRV will be assessed via changes from baseline HRV measurement averages. Change scores for HRV, craving, and affect will be explored as between-group differences using 2x2 analyses of variance. Multiple regression analyses will then examine the effect of cognitive reappraisal and HRVb on changes in HRV, craving, and affect, with age, average daily cigarettes smoked, Body Mass Index, and demographic variables entered as covariates. Second, investigators will explore changes over time for the multiple assessments of HRV, craving, and affect by examining within-subject changes over time. Within-subject data analysis will be conducted in Hierarchical Linear Modeling (HLM) Version 7 software (Raudenbush et al., 2011) to account for the nesting structure of the data as well as missingness of data, and allow for analysis with random intercepts. Multilevel models will include the covariates mentioned above in addition to the experimental manipulations of cognitive reappraisal and HRVb.