Breathe Easier II: A Dyad-based Multiple Behavior Intervention

Overview

This pilot randomized control trial (RCT) will test a 12-week, multiple behavior intervention physical activity and stress management for survivors with early stage lung cancer (stages I-III) and their family members (1 survivor + 1 family member or friend = 1 dyad). The long-term goal of this research is to improve health outcomes for survivors of lung cancer and their family members. The goals of the intervention, Breathe Easier, are symptom reduction (less breathlessness, less fatigue, less stress) and change in multiple behaviors (increase in stress management and increase in physical activity, and decrease tobacco use - if appropriate). Our aim is: To conduct a 6-month, two-group, pilot randomized control trial intervention study with a pre- and post-test study design to estimate preliminary intervention effects on (a) reduction of symptoms (breathlessness, fatigue, and stress) in survivors of non small cell lung cancer (stages I-III) and family members or friends; (b) increase in physical activity behaviors immediately following the intervention and at 3-months; (c) increase in stress management strategies immediately following the intervention and at 3-months; (d) reduction in smoking behavior among participants who smoke tobacco products at study entry immediately following the intervention at 3-months.

Lung cancer is the second most diagnosed malignancy and the leading cause of cancer deaths, accounting for more annual deaths than breast, prostate, and colorectal cancer combined. Approximately 80% will be diagnosed with NSCLC, and about 42.1% will present with early-stage or localized disease (stages I-III). Improvements in screening and treatment have led to better outcomes for early-stage diagnoses. As early detection improves along with treatments, the current 5-year survival rates (57%, stages I-III; 70%, stage I) are projected to improve. Only a small percentage (5-10%) of patients with lung cancer are asymptomatic at diagnosis. Cough, dyspnea, chest discomfort, and fatigue are common symptoms. Up to 80% of long-term lung cancer survivors report smoking-related comorbidities such as chronic obstructive pulmonary disease (COPD) and heart disease. Thoracic surgery, radiation and chemotherapy may improve or worsen symptoms. After curative surgery, survivors describe persistent physical and psychological symptoms negatively affecting their QOL. The extent of the resection, co-morbid conditions, and smoking status further exacerbate symptoms. Symptoms may be intensified by smoking or withdrawal from tobacco. Nicotine withdrawal brings about a variety of physical and psychological symptoms that can temporarily intensify the symptom burden. Overall, survivors with lung cancer experience more symptom distress than survivors with other types of cancer. Multiple unhealthy behaviors often accompany smoking, including lower physical activity levels and ineffective stress-management skills. Researchers have examined the prevalence of multiple health behaviors in lung cancer survivors; four studies targeted dyads. In one cross-sectional study data from 37 dyads (mostly white, non-Hispanic) showed high rates of smoking (43%, survivors; 30%, family members (FMs)) and physical inactivity (84%, each group). In terms of readiness to change, within 6 months post-diagnosis, 63% of patients and 81% of FMs were ready to increase their physical activity, and 88% of patients and 91% of FMs were prepared to stop smoking. A qualitative study involving 26 African American dyads, within 7 years post-diagnosis, revealed fewer current smokers (15%, survivors; 8% FMs) and persistent physical inactivity (65% survivors; 42% FMs). Two more recent studies collected similar data from survivors and their caregivers (mostly white, non-Hispanic) within 1-month post-diagnosis and 6-months post thoracic surgery reported that survivors and caregivers were receptive to educational and lifestyle behavior modification programs. Evidence strongly suggests that smoking cessation and physical exercise interventions (individually administered) improve QOL and other outcomes; furthermore, exercise therapy is a well-tolerated, safe, and feasible adjunct therapy that can mitigate several diseases- and treatment-related side effects. Mounting evidence suggests that survivors with early-stage lung cancer benefit from regular physical activity. In a study of 175 long-term lung cancer survivors, the 27% who reported meeting physical activity guidelines also reported a higher QOL. Two pilot studies provided "proof of principle" and showed aerobic training (continuous activity involving large muscle groups) is safe and feasible for survivors of early-stage lung cancer. Thus, it is hoped that this research will advance scientific knowledge about these behaviors being complementary and producing a synergistic effect on behavior change and symptom management. Family, friends, and especially partners of survivors of lung cancer must cope with its physical and psychological impact on the survivor, themselves, and others. It is important to recognize that survivors and their family members react to a serious illness as a unit, and as a result they both have a genuine need for help from healthcare providers. Partners not only provide emotional and practical support, but they must also cope with their own concerns, including the uncertainty surrounding the course of the illness, fear of losing their partner and the impact of their health behaviors. More than half of the partners of lung cancer survivors report negative emotional effects of being the support person. Relationships between survivors and partners are affected by the cancer also. An emerging consensus in the literature is that when both the survivor and family member are treated simultaneously, the well-being of each individual improves. When their needs are not addressed, family members are at risk for impaired health. Interestingly, most health behavior-related interventions focus on an individual's behaviors. Yet, health behaviors tend to cluster in families and social networks. The past decade has seen a growth in interventions exploring the effects of family members (and their behaviors) on each other when a serious illness is diagnosed in one of them. A greater emphasis is needed on interventions that target families, not only the survivor. This proposed research represents the next step toward understanding the best ways to improve multiple behaviors and therefore reduce the burden of symptoms among survivors of lung cancer and their family members. The long-term contribution of this research is to advance scientific knowledge about improving outcomes for the survivors of localized NSCLC (stages I-III) and their family members. Theoretical Framework: Individual and family self-management theory (IFSMT) is a descriptive mid-range theory developed based on conceptual and theoretical perspectives from Systems Theory, Social Learning Theory/Social Cognitive Theory, and Social Support Theory and research related to self-management (SM) of chronic illnesses. IFSMT proposes that successful change in the self-management of behaviors (physical inactivity and stress management) after a diagnosis of lung cancer requires active involvement by the survivor coupled with family support. IFSMT encompasses three broad dimensions: context, process, and outcomes. The context dimension can directly impact individual and family engagement in the self-management process and, therefore, the outcomes. Within the process dimension, self-regulation is the process used to change behavior. Social support enhances the capacity to change and includes influence and collaboration among individuals and family members. In this study, the intervention impacts the self-management process by promoting self-efficacy, self-regulation, and social support; enhancing these three factors, in turn, will lead to improved proximal outcomes (increased stress management and physical activity and reduction in tobacco smoking, relapse prevention) and distal outcomes (reduction of symptoms of dyspnea, fatigue, and stress) for all participants. Design: This study will use a 6-month, longitudinal, two-group, repeated-measures, attention control pilot RCT design. During an enrollment period, the research team will provide information about the study, execute an informed consent process, and collect baseline data. After baseline data is collected, participants will be instructed about the wearing of a wrist accelerometer and responding to Ecological Momentary Assessment (EMA) surveys on their cell phones for a 7-day period prior to randomization. All participants will be instructed on how to keep an accelerometer log and how to respond to mobile EMA surveys. At the end of the 7-day period, accelerometers are returned and completed data is confirmed, the randomization process will take place. The randomization strategy will be determined by the statistician. Based on a grid developed by the statistician each envelope identified by study number only will contain the identification of the study randomized condition (Breathe Easier II \[intervention\] or attention control. Statistical Analysis As a preliminary step, baseline demographic characteristics and medical histories will be compared between groups (intervention and control), for survivors and family members separately, as well as aggregated across the dyads, using t-tests (for continuous variables) and chi-squared tests (for categorical variables). This is only a preliminary step, dyadic analysis follows. Additionally within-dyad correlations in baseline measures may differ across dyads. For example, whether the association between baseline sleep and physical activity within-dyad differs across pairs. A dyadic analysis will be used as follows: using a series of longitudinal mixed effects models, the effect of intervention vs. control on changes in symptoms (shortness of breath and fatigue) after baseline will be estimated. Specifically, mean symptom at follow-up (end of intervention, 3- and 6-month follow-up) will be simultaneously regressed on group, time, group x time, baseline value, and covariates not balanced by randomization. In addition, models will control for smoking status. Interactions between smoking status at study entry within the intervention group will be explored, as it is hypothesized that the effect of the intervention will be greatest among those who were smoking at study entry. Models will include a subject-specific intercept to adjust for repeated measures of the outcome within participants. Models will be run for the study pool (survivors and family members, considered as dyadic units) with clustering by dyad. This will adjust standard errors for the dependence of outcomes within the dyad by including an interaction term between participant (survivor vs. family member) and treatment group (intervention vs. control). As this is a pilot study, the interest is in estimating effect sizes and corresponding 95% confidence intervals, rather than strict statistical hypothesis testing. Similar analyses will be run to estimate intervention effects on physical activity behavior (exercise capacity, minutes of moderate-intensity activity both objectively measured and self-reported, and sedentary time) and to explore intervention effects on changes in exploratory variables (context and process measures) over time. Finally, using a series of longitudinal regression models implemented with generalized estimating equations with robust standard errors and clustering by dyad, the effect on smoking outcomes at follow-up (abstinence) will be estimated. Specifically, binary indicators of abstinence on treatment assigned (intervention vs. control) and potential covariates (including variables not balanced by randomization) will be regressed using binomial errors, a logit link function, and a working unstructured correlation to accommodate within-subject correlation. Coefficients will be standardized, and models run with the sub-sample of dyads where at least one member smoked at study entry. All models will be run on the intent to treat the sample with all randomized participants included. Changes over time in secondary measured variables (e.g., process measures, sleep) and potential moderators of the intervention effect, will be explored using a similar analytic strategy to that described above. Moderator analyses will be explored by including main effects of the potential moderators (in the model of physical activity behavior for example), as well as the interaction with the treatment group.