JMIR J Med Internet Res Journal of Medical Internet Research 1438-8871 Gunther Eysenbach JMIR Publications Inc., Toronto, Canada v13i4e119 22182518 10.2196/jmir.1911 Original Paper Online Social Networks and Smoking Cessation: A Scientific Research Agenda Eysenbach Gunther Wicks Paul Cobb Nathan K MD 1
The Schroeder Institute for Tobacco Research and Policy Studies American Legacy Foundation 1724 Massachusetts Ave NW Washington, DC, 20036 United States 1 202 454 5745 1 202 454 5785 ncobb@legacyforhealth.org
2 3 4 Graham Amanda L PhD 1 3 Byron M. Justin MHS 4 Niaura Raymond S PhD 1 3 4 Abrams David B PhD 1 3 4 Workshop Participants 5 1 The Schroeder Institute for Tobacco Research and Policy Studies American Legacy Foundation Washington, DC United States 2 Division of Pulmonary & Critical Care Department of Medicine Georgetown University Medical Center Washington, DC United States 3 Department of Oncology Georgetown University Medical Center / Lombardi Comprehensive Cancer Center Washington, DC United States 4 Department of Health, Behavior and Society Johns Hopkins Bloomberg School of Public Health Baltimore, MD United States 5 see acknowledgements Oct-Dec 2011 19 12 2011 13 4 e119 13 08 2011 27 08 2011 19 09 2011 25 09 2011 ©Nathan K Cobb, Amanda L Graham, M Justin Byron, David B Abrams, Workshop Participants. Originally published in the Journal of Medical Internet Research (http://www.jmir.org), 19.12.2011. 2011

This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.

 Background

Smoking remains one of the most pressing public health problems in the United States and internationally. The concurrent evolution of the Internet, social network science, and online communities offers a potential target for high-yield interventions capable of shifting population-level smoking rates and substantially improving public health.

 Objective

Our objective was to convene leading practitioners in relevant disciplines to develop the core of a strategic research agenda on online social networks and their use for smoking cessation, with implications for other health behaviors.

 Methods

We conducted a 100-person, 2-day, multidisciplinary workshop in Washington, DC, USA. Participants worked in small groups to formulate research questions that could move the field forward. Discussions and resulting questions were synthesized by the workshop planning committee.

 Results

We considered 34 questions in four categories (advancing theory, understanding fundamental mechanisms, intervention approaches, and evaluation) to be the most pressing.

 Conclusions

Online social networks might facilitate smoking cessation in several ways. Identifying new theories, translating these into functional interventions, and evaluating the results will require a concerted transdisciplinary effort. This report presents a series of research questions to assist researchers, developers, and funders in the process of efficiently moving this field forward.

 Smoking cessation social support social networks addiction treatment tobacco Introduction

Smoking remains the leading cause of 443,000 preventable deaths and nearly US $200 billion in excess costs in the United States each year [1]. Smoking rates in the United States have stalled near 20%, [2] and large-scale reduction in smoking prevalence remains an urgent public health imperative. Although the evidence-based cessation interventions recommended by the US clinical practice guideline for tobacco dependence treatment [3] have been shown to double quit rates, they are largely underused [4]. Reaching the US public health goals of cutting the smoking rate to no higher than 12% by 2020 [5] will require novel approaches to create new interventions, enhance the effectiveness of existing cessation treatments, and maximize the reach and utilization of both.

The evolution of the Internet and the growth of online social networks may present a solution to the intertwined problems of effectiveness and reach of cessation interventions. Social support [6], social integration [7], and social networks [8] appear to play important roles in smoking behavior and cessation. Yet numerous tobacco treatment studies aimed at creating supportive relationships (eg, peer or buddy training) or harnessing existing social relationships (eg, spouse interventions) have generally yielded disappointing results [9-12]. The limitations of traditional treatment settings in which this work was conducted (eg, low attendance, time constraints, or type and number of available support persons) may partially explain the difficulty in leveraging social support in the cessation process.

Online social networks, by contrast, offer round-the-clock access to vast numbers of participants, potentially superseding these limitations and offering a realistic delivery model for social support. In theory, smokers might benefit not only from active, personal interactions with other network members, but also from various passive sources of social support and influence. Such interactions could alter an individual’s motivation to quit, reinforce the undesirability of smoking, assist in buffering cessation-related stressors and enhancing coping skills, and provide suggestions for eliminating smoking cues [13]. To date, there has been a wealth of behavioral science research on the role of social networks in face-to-face interactions but little published research on online social networks [14-16].

The growth of online social networks and their penetration into popular awareness has been phenomenal, with over 70% of American adults now using some form of social media or online social network [17]. As of early 2011, an estimated 150 million Americans actively use Facebook, the largest of the online social networks [18]. Intentionally created online networks dedicated to smoking cessation are smaller but have been in existence for over a decade. These types of dedicated systems—where smokers and former smokers communicate through various channels in an effort to quit and stay abstinent—are now widely used by hundreds of thousands of smokers over relatively long periods of time [16,19]. Over the years, cessation-focused online networks have evolved from simple systems for the exchange of messages to complex networks complete with multiple modes of communication (eg, chat rooms, forums, or private messaging), self-representation (eg, personal profiles, blogs, or journals), and affiliations (eg, buddy or friends lists, or private groups), through which social norms, social influence, and social support may be conveyed in real time [16,20].

Concurrent with the exponential growth of online social networks has been the rapid evolution of social network science, spurred on as improvements in computer capacity and software have caught up with theory and the burgeoning size of available data sets [21,22]. In studies of real-world networks, social network science has demonstrated that social influence flows through networks and can influence a broad range of behavioral and emotional changes, including smoking and alcohol use [8,16,23], obesity [24,25], happiness [26], and depression [27], as well as loneliness [28] and suicide [29]. Social network analysis allows for an expanded view of an individual’s social universe, taking into account not only their own connections, but also the connections of their friends and contacts and beyond. This ability to look at the social structure in aggregate allows for inferences about how topology (the network structure) both enables and drives behavior change.

Actually cutting smoking prevalence by nearly half by 2020 will require cessation interventions that can reach millions of people in consumer-friendly ways. The convergence of robust evidence for the role of social support in cessation, the growth and proliferation of online networks, and the recent advances in social network analytic techniques present an opportunity for the development and dissemination of high-impact interventions targeting smoking. The notion that online social networks present a powerful and novel approach to cessation is supported by a research in relatively disparate disciplines, including tobacco control, social psychology, and social network science, to name just a few. Leveraging the enormous potential of online social networks to reach and treat smokers will require a transdisciplinary conversation among researchers, developers, and funders that bridges behavioral, network, and computer sciences and other fields [30,31]. We sought to initiate this discussion by convening a multidisciplinary group of experts to identify gaps in knowledge and research questions regarding the potential of online social networks to more rapidly reduce smoking prevalence. Our goal was to construct a strategic research agenda to guide future collaborative work. This paper presents this agenda in the form of 34 pressing research questions and related issues, along with brief discussion.

 Methods

We invited approximately 100 experts and thought leaders (listed under Acknowledgements at the end of this article) across a range of relevant content areas to a 2-day workshop held September 30 to October 1, 2010 in Washington, DC. Participants represented a broad range of disciplines, including economics, engineering, epidemiology, linguistics, mathematics, medicine, nursing, psychology, public health, network science, sociology, software engineering, and product design and commercialization. A small number of participants were invited to give focused overview presentations to help bridge disciplinary borders and to establish a common starting point for discussion. These included presentations on the epidemiology and treatment of tobacco use; basic principles of social support theory and social support interventions in tobacco control; social network science and network-based interventions in tobacco control; the history, evolution, and current state-of-the-science of general and cessation-specific online social networks; and methodological, measurement, and analytic issues regarding social network data collection, analysis, and interpretation. Additionally, representatives from three of the largest for-profit, health-related, online social network interventions were invited to describe their programs and the lessons they had learned in managing online networks.

Following the overview presentations, participants were divided into small multidisciplinary working groups and tasked with developing a list of priority research questions. The guiding framework for workgroup discussions was to address the key question “What do we know and what do we need to learn that will make a difference in improving cessation outcomes?” The framing of the question was deliberately broad to enable participants from diverse disciplines and with varying content expertise to contribute their perspectives.

Participants were instructed to formulate and group research questions into four major categories: (1) a dvancing theory (developing, refining, or integrating existing theories and models from online and offline social network, social support, smoking cessation, and behavior-change domains), (2) understanding fundamental mechanisms (how online social networks produce behavior change at the individual and network level), (3) intervention aproaches valuation (methods and metrics for appropriate program, process, and outcome evaluations). To encourage brainstorming, groups were instructed to imagine they had access to the intellectual, financial, and technical resources represented by any combination of the attendees or speakers at the workshop. Each working group presented their list of research questions back to the full group for further discussion. The groups worked independently on each major theme area, and then their recommendations were synthesized and refined with feedback from the entire group. After the workshop, the planning committee met to review the findings and to summarize the general areas of research topics and priorities for dissemination.

 Results

Participants generated a large number of research questions at varying levels of granularity. Common and overlapping ideas were integrated and a subset of questions was selected for further discussion and elaboration by the report’s authors. For each key topic area, we present a summary of discussions and provide examples of the most pressing research questions or issues raised.

Several overarching themes emerged from the discussions. First, participants noted that traditional models of offline (eg, face-to-face) intervention and evaluation are often reflexively applied to online observations or interventions. While there are ways in which offline and online behaviors overlap and can reciprocally inform models, mechanisms, implementation, and evaluation, there are also important differences that require critical thinking about online networks. There is a need to challenge and test the assumptions inherent in traditional models when developing, implementing, and evaluating online interventions.

A second theme related to the mechanisms of behavior change. Numerous theory-based processes of behavior change have been described within social networks, including diffusion of information, viral spread of interventions, social support, social norms, and modeling. It is unknown whether these or other unidentified processes are important in online social networks for cessation, and if any of these may be iatrogenic (ie, promoting continued smoking rather than cessation).

A third theme centered on the appropriate use of theoretical models, empiricism, and statistical or simulation modeling techniques. Future advances in online social network interventions will likely depend on a transdisciplinary approach to develop appropriate theoretical models, test them in vivo and in silico (software modeling), rapidly iterate to determine interventions with the highest probability of effect, and perform intervention trials with appropriate research designs and end points. Such advances will require improvements in existing capacity to collect complex and large-scale longitudinal data on behaviors and interactions within online networks.

Finally, we note a common assumption during the workshop that social network interventions will increasingly take advantage of mobile delivery mechanisms—whether smart phone apps, text messages, or other formats. While few questions address this shift explicitly, we have attempted to write this summary to be agnostic toward delivery platform. Both questions and recommendations are intended to be broadly applicable, regardless of location or modality.

Advancing Theory of How Social Networks Influence Smoking-Cessation Behavior

Social network and social support models in smoking cessation [6] derive primarily from social learning theory [32] and from the study of interventions to change existing social support interactions or develop new supports (eg, from a counselor and other members of a group participating in face-to-face smoking-cessation treatments) [10,33]. Observational studies have shown that social support is associated with smoking initiation and cessation, and that smokers associate with other smokers in proximal social networks. Intervention studies, though, have yielded mixed and largely disappointing results in attempting to manipulate or harness support [6,10,33]. These findings highlight the importance of the need for a theoretical framework that permits simultaneous understanding of the observed association between social phenomena and smoking, and manipulation of the social environment to effect change. There is a need to evolve our theories and models to refine their explanatory power and their applied utility to facilitate behavior changes, such as smoking cessation [34].

The application of network theory to social networks has largely occurred in studies of real-world (ie, offline) networks [8,29,35-37] using retrospective self-report measures and cross-sectional data. Recent research on social networks has rapidly evolved using online data [16,38-42], new computational methods, and mathematical and simulation modeling [37,43,44]. Data from electronic communications networks (eg, online social networks, email systems, and telephone networks) can be collected in real time and can record communications and interactions at multiple levels and with repeated observations of intraindividual, interindividual, and contextual influences. Such methods can inform interventions and measures of process and outcome, but the proliferation of data and results calls for new or more refined models or theoretical frameworks to facilitate interpretation and application.

Theories that try to explain and change behavior in small real-world settings may not translate easily into the online world, where interactions occur on a larger scale and in a different medium. A transdisciplinary synthesis [30,45] is likely required to integrate our understanding of the nature and form of social networks, as instantiated or reflected in the online world, and their functions that serve to initiate and maintain changes in individual behaviors. Structuralist social network theories, which address how patterns of social relationships are associated with substantive topics such as health behaviors [37,45], come closest to fusing form and function, and serve as a useful point of departure for understanding how social networks and individual tobacco use behaviors intersect.

A century ago, Simmel [35] called for more than knowing how to measure characteristics of networks, such as the density of their interconnections, and recommended developing a set of assumptions about how best to describe and explain the social phenomenon of interest in its proper context. This challenge remains today as we seek to integrate social and individual theories of behavior change. Indeed, the structure of a network may induce, maintain, or strengthen a behavior not just by transmission of information, but via forces of exclusion, adaptation, or the binding together of members [46]. Looking at network causes of phenomena of interest requires asking what kinds of social networks lead to particular outcomes.

 Specific Questions

1. How well do theoretical models of social influence translate between offline and online contexts?

2. How does online social network data map onto real-world networks? Does research based on retrospective self-report with sparse observations in the real-world match with dynamic, observed behavioral data collected online?

3. How can behaviorally important ties be identified in online social networks that may be composed of large numbers of apparently weak ties?

 Understanding Fundamental Mechanisms

Online social networks exist across a broad range of health conditions and behavioral risk factors including tobacco use (eg, becomeanex.org, quitnet.com, and stopsmokingcenter.com), diet and fitness (sparkpeople.com), diabetes (tudiabetes.com), chronic diseases (patientslikeme.com), and others [16,47-49]. Research is in its infancy regarding the mechanisms through which these online social networks might or might not effect behavior change. Social support models suggest that behavior change is mediated in part through information exchange, instrumental or emotional support, stress buffering, or improved self-efficacy [6]. However, other mechanisms may be as or more important in online networks, such as exposure to new or different norms or behaviors modeled by other network members [16]. To date, the design and implementation of online networks has been largely based on offline cessation approaches, usually comprising only small groups of smokers actively trying to quit. The evolution of more effective cessation interventions will require an in-depth, sophisticated understanding of the unique aspects of online social networks and the specific mechanisms through which they effect behavior change, as well as the careful selection of evaluation strategies matched to this intervention context.

Homophily, Heterophily, and Network Topology

Homophily refers to the tendency of people to associate with similar others (“birds of a feather”), while heterophily refers to the tendency to collect in diverse groups. That homophily tends to be a driving factor in the formation of social networks [50] is an important consideration in offline networks: the tendency of smokers to associate with other smokers may decrease the impact of normative exposure to nonsmokers or former smokers within a network. In contrast, online social networks may be heterophilous, comprising individuals across the cessation continuum including individuals who have been abstinent for years [16] or current smokers who are curious but not yet motivated to quit. Research in offline networks suggests that topological factors (the pattern of ties between individuals within the network), such as clustering of smokers, affects cessation over time [8]. Other work in online networks indicates that dense connections at the individual level reinforce social signaling and increase the chance of behavior change [39]. As most existing online networks remain uncharacterized, little is known about their structure or the optimal topology to effect behavior change.

Specific Questions

1. What is the role of homophily in the formation of online networks?

2. What is the role of heterophily in the provision of social support throughout the cessation process, and how does it influence cessation outcomes?

3. Can ties within online social networks be fostered or manipulated to “rewire” networks, modify topology, and drive behavior change?

4. What impact does network topology have on behavior change? For example, does having a dense local network increase the probability of making a quit attempt, cessation, and maintenance of abstinence?

 Social Diffusion

Information and behavior diffusion through offline social networks are well-studied phenomena, encompassing myriad behaviors from seed choice by farmers to the spread of smoke-free policies from city to city [51]. In contrast, inducing or manipulating diffusion through both online [40,41] and offline networks [51,52] has proved challenging in practice; deliberately causing spread of information or a behavior is easier to conceptualize than to implement. In commerce and industry, the term viral marketing refers to this deliberate seeding and resulting diffusion of a message through a targeted network, such as the promotion of a new product [53]. While viral marketing is a common practice, there is little academic literature on its use for health topics or for online approaches for health behavior change. Nonetheless, deliberate seeding and diffusion may allow for the dissemination through created networks of specific information (eg, information about a new cessation medication), interventions (eg, a quit smoking app through Facebook), smoking/cessation norms, and other health behaviors.

Specific Questions

1. How does information spread through an online social network? Are there identifiable patterns of information spread that can be leveraged in intervention research?

2. Can key participants in a network be identified and targeted to foster information diffusion or make it more efficient?

3. What are the drivers of the viral spread of an application, concept, or innovation through online networks?

4. How does network topology affect diffusion? Can social network measures and concepts such as centrality or clustering be used to predict or alter diffusion?

 Social Norms and Modeling Behavior

Despite the fact that members do not know each other at the outset, created online networks can develop their own language and norms [54]. Existing members may convey expectations for certain behaviors or participation in the network that guide and support new or struggling members [55]. These expectations and norms may differ from those in the participant’s offline network. For example, the public health community has worked hard to normalize the use of nicotine replacement as a cessation aid; however, most smokers do not use pharmacotherapy when quitting [4] and many have concerns about the safety of any form of nicotine [56]. Online social networks may present norms supportive of medication use, and existing users may model successful medication use behavior. Other norms such as recycling after failed quit attempts, enlisting external social support, or the use of telephone counseling are other examples of potential norms (positive or negative) within social networks.

Specific Questions

1. How are social norms established and communicated in an online social network?

2. What is the effect of online social norms when they differ from a user’s offline environment?

3. Does anonymity in online networks enhance or diminish the effect of modeling behavior and communication of norms?

4. Are norms and modeling effective mechanisms to influence “lurkers” (ie, members of a network that read other members’ posts/comments but rarely communicate with other members)?

 Network Formation, Social Integration, Retention, and Longitudinal Stability

There are numerous online communities and created social networks dedicated to health-related behavior change—some of them in existence for over a decade with thousands of members—yet it remains unclear what factors led to their growth or stability. Previous research has shown that small numbers of individuals may be responsible for approaching and “integrating” new members as they join an online network for cessation [16]. Most research has reported results from successful, stable networks [14,16,57], while projects that fail to form networks are rarely reported [58]. As a result, the factors that drive member integration and retention and network stability remain unclear. Adequate understanding of these factors is required to build new interventions and to maintain existing versions or enhance their effectiveness.

Specific Questions

1. What predicts engagement in an online social network? What demographic, smoking, psychosocial, or other characteristics are predictive of participation and integration?

2. What is the role of timing of interactions in online social network in influencing integration and participation? What forms of outreach and communications (eg, private messages, instant messaging, public forums, or blogs) drive tie formation?

3. What is the role of long-term users in network structure and network stability over time?

 Intervention Design and Approaches

The incredible growth of online social networks offers the opportunity for novel intervention designs. Created networks such as online communities dedicated to smoking cessation are a common component of modern health behavior-change systems and often center on the “build it and they will come” premise of intervention delivery. These networks generally comprise motivated individuals ready to make or maintain changes to one or more health risk behaviors. Such systems benefit from a specific focus, on the part of both the user and intervention designers. However, they generally do not yet take full advantage of the potential to proactively reach larger populations. Individuals must generally seek out and enroll in the closed system, and ultimately many registrants fail to return to the site [59], much less engage with the social aspects as designed (they become, at best, lurkers or, at worst, completely unengaged). In contrast, general-purpose networks such as Facebook offer unique opportunities and challenges, related primarily to their enormous size, including the potential for autonomous propagation (viral spread) of interventions. Certainly, intervention design decisions should be informed by relevant and sophisticated theories that specify the active ingredients and mechanisms of action, but the surfeit of potential participants in these extremely large networks ultimately allows for data-driven methods to drive the ongoing design and refinement of interventions.

Target Populations

Smoking-cessation interventions most frequently target individuals ready to make a quit attempt. Yet many people who join online cessation systems have already quit or are not ready to make a quit attempt [60-63]. Traditional social support models will need to be modified to assist these individuals and to maximize their utility in supporting others. Significant public and private resources are used each year to denormalize smoking and encourage cessation using traditional media [64,65]. As public health organizations increasingly use advertising and outreach efforts to drive utilization of online resources, it will become imperative to identify the types of smokers that may benefit from social network-based approaches to cessation. A one-size-fits-all approach is unlikely to be efficient or effective, and it is unclear how much customization or individual tailoring is needed to make an incremental addition to outcomes [66,67].

Specific Questions

1. Do smokers who are not motivated for behavior change benefit from social network interventions? What influence do social support and normative exposures have on smokers who may not be thinking of quitting?

2. Can online social networks assist smokers who have already quit to maintain abstinence? Can recruiting abstinent smokers into a network strengthen the network’s capacity for social support?

3. Are demographic or psychosocial characteristics important predictors of online social network utilization? What is the impact of age, gender, race/ethnicity, or other identifying characteristics with regard to network phenomena such as integration or tie formation?

4. How can network-based interventions capitalize on secular trends and historical events, such as a change in the federal excise tax rate, new year’s resolutions, The Great American Smokeout, or major smoking-related media stories such as the death of Peter Jennings from lung cancer? Do smokers recruited during the “surges” associated with these events differ from those who join an online social network at other times or for other reasons?

 Systems Integration

The oldest examples of online social networks for cessation are relatively siloed intervention approaches, focused largely on engaging users with other participants on a cessation-specific website and in an anonymous fashion. More recent interventions integrate online social networks into other treatment-delivery approaches, such as telephone quitlines [13,68,69]. The rapid expansion of large-scale networks such as Facebook where users are personally identifiable offers the opportunity to disseminate cessation interventions through existing networks, but without the aspect of anonymous participation. It is unknown the degree to which the advantages of leveraging an existing network where participants are identified are offset by the potential benefits of a network where members are anonymous. Integration of online social networks with other treatment modalities (eg, text messaging, health care-delivery settings, or electronic medical record systems) offers the opportunity to enhance treatment effectiveness, augment social support mechanics, and increase the reach of traditional services. At the same time, such integration introduces multiple new complexities.

Specific Questions

1. What is the best mechanism for online social networks to interface with other elements of health care or tobacco treatment (eg, telephone quitlines, over-the-counter and prescription pharmacotherapy, physician advice, electronic medical record, mass media campaigns, or policies)?

2. How does involving a smoker’s offline network (eg, friends, family, medical practitioners, worksite wellness, or occupational health programs) augment or diminish the effect of an online social network on cessation?

 Development Methods

There is a chasm between the rapid-cycle, diffusion-focused development methods used by entrepreneurs and industry to launch online programs and the traditional, efficacy-based development methods of behavioral and social scientists. For example, Facebook has grown literally from a dorm room project to over 150 million Americans a month in approximately 6 years. Ironically, this is typically the same amount of time between submission of a federal grant application and the publication of its main outcome paper. Shortening this timeline is critical if we are to develop effective interventions that can be deployed on a large scale to benefit public health in a timely fashion. Engineering principles of iterative development and early evaluation have been adapted in the behavioral sciences (eg, multiphase optimization strategy, or “MOST”, [70]) and provide one approach to achieve this goal. Online interventions are particularly suited to these methods; large, available target populations enable intervention variations to be tested against each other with statistical significance in rapid sequence or in factorial models, in theory improving effectiveness and tightening research and development timelines [67,70].

Specific Questions

1. Can engineering models, such as MOST, speed development time and/or increase efficacy of network-based interventions?

2. What process and outcome metrics are most appropriate during intervention development and refinement? Participant engagement? Retention? Network integration? Quit attempts? Early abstinence?

 Evaluation

Several high-quality randomized controlled trials of Internet cessation programs have been conducted [13,57,62,67,71-76]. Yet, to date, there have been no published reports of tobacco intervention trials that link social network structure or dynamics to either social support metrics or more distally to cessation outcomes. Not all of these trials have included social network components, but among the ones that did, there are several reasons for this gap in the literature: the difficulty of constructing appropriate assessment and intervention protocols, the difficulty in maintaining participants in social interventions, and the challenge in disentangling social processes from other features of many Web-based interventions (eg, tailored materials, expert systems, or access to counseling staff). There is a critical need for the identification of appropriate research designs, data collection methods, and evaluation strategies to determine the impact of social processes within online interventions that may drive cessation and abstinence.

Research Design

The use of randomized control trials in research to evaluate online social network-based interventions presents a number of challenges. Among these are selecting a feasible, ethical, and rigorous control condition [77,78]; avoiding contact between participants randomly assigned to different conditions; and managing the attrition observed across virtually all online interventions [59]. Alternative evaluation designs used in eHealth research, such as practical clinical trials, pragmatic randomized controlled trials, and nonexperimental and quasi-experimental designs [77,79-81], may be appropriate as well for social network interventions. Given the size of data sets that are generated from online social network interventions, automated systems for the categorization and extraction of data (such as natural language processing and sentiment analysis, data mining, and pattern recognition) may also play important roles in exploratory analyses. The use of varied methods and data sets will make consistent and standardized reporting of results increasingly important as the field advances.

Specific Questions

1. Given that alternative Internet interventions are a mouse-click away, what are the important considerations in selecting a rigorous and appropriate control group and evaluating contamination (ie, exposure to the intervention arm among control participants)?

2. Other than randomized control trials, what rigorous research designs can be aptly used to optimize online social network interventions? Are there specific research designs that are best used at specific phases of the development–dissemination–implementation continuum?

 Data Collection and Analysis

Online interventions and social networks in particular are part of the “big data” problem [22], an emerging issue where the quantity of behavioral and other process data exceeds the capacity for traditional analysis. Academic computational social science—the collection and analysis of these data—lags behind other fields such as physics and biology, as well as the corporate capacity of Google and Facebook in managing big data [21]. The two primary challenges inherent in big data are adequately defining and capturing the appropriate data, and conducting effective and efficient analyses. Data collection methods such as ecological momentary assessment, mobile tracking data, content and sentiment analysis, and observation of online interactions can provide granular information about behavior with minimal impact on the user or their friends and contacts. These methods can generate much richer—and also more complicated—representations of social networks that contain information about the weight of ties, their valence (positive or negative), and the presence of hidden or latent ties [82].

Specific Questions

1. How can novel data collection methods such as ecological momentary assessment, passive tracking data from websites, or data from mobile devices be used to gather network-level data without affecting individual behavior or the network itself?

2. What new techniques and analytic methods will be required for analysis of “big data” and increasingly complicated network representations?

 Expanded Outcomes and End Points

Traditionally, research has evaluated the impact of an intervention only on the individuals enrolled in a study. Bolstered by evidence from both offline [8,23-29,51,52] and online [39-42,53] studies, network theory suggests that behavior change may diffuse through a network. Successful intervention with an individual smoker may have positive externalities (a term for collateral effects, drawn from the economic literature) that ripple through the network [83] causing other smokers to quit or to cut back on their smoking, or resulting in changes in attitudes or other beliefs [8]. For example, a quit attempt by an individual enrolled in a program might prompt a close friend to also attempt cessation. Success of the friend would not normally be part of a traditional analysis, but becomes critical from a network standpoint, particularly since interventions may be specifically designed to elicit this effect. Given that evaluating changes in behavior among individuals outside the purview of a research study may be difficult or impossible, alternative end points, outcomes, and evaluation strategies become imperative [83].

Specific Questions

1. What end points or surrogate outcomes will permit the evaluation of externalities in online network interventions?

2. What are the ethical implications of observing or even inducing behavior change in individuals that have not consented to participate in a research study?

 Modeling to Inform Design and Evaluation

The use of mathematical predictive models in public health, and tobacco control in particular, has recent support [84-86]. Their use to design, refine, or evaluate behavioral interventions for cessation is less defined, but the opportunities are compelling. Previously, models have been employed to examine how best to optimize the multiple modes of delivery of smoking-cessation interventions, as well as to capitalize on context, such as multilevel influences of restrictive polices, mass media, and increased sales taxes [85,87-89]. In silico techniques such as agent-based modeling, where powerful computers simulate autonomous users interacting within a network over time, can be used to predict responses to intervention design changes [90]. Under certain circumstances they can also be used to disentangle behavioral outcomes from network processes and potentially contribute to evaluation [34]. Such techniques not only may play a valuable role in accelerating intervention development and evaluation, but also may help to determine the potential impact of interventions prior to time consuming and costly promotion and implementation–dissemination efforts.

Specific Questions

1. How can mathematical and computer-driven simulations of various kinds (eg, dynamic systems models or agent-based models) contribute to intervention development, refinement, or evaluation?

2. How can existing systems models inform work with online social networks? How might existing systems models be affected or informed by large-scale social networks (such as Facebook)?

 Discussion

An increasingly interconnected online social Web provides incredible opportunities to shift behavior, affect health, and meet public health challenges. Despite promising starts in individual fields, it will take further rigorous and transdisciplinary research and development to meet the potential described in this report. Tackling the questions posed here, structuring research protocols, and developing appropriate analytic techniques will require true collaboration across multiple fields and divergent disciplines [30]. Success may lead to interventions with the capacity to reach large populations, augment existing treatment modalities, and effect behavioral change in novel ways.

While we have focused on tobacco use and smoking cessation, the same questions and approaches may apply to virtually any behavior change of interest. Interventions need to be informed by and should inform theory, model testing, and protocols for refinement. As we gain experience working in transdisciplinary teams and refine our models, we will have a clearer picture of the new measures needed for empirical data collection and testing of models to identify the mechanisms, pathways, and key processes that influence intermediate and final behavior-change outcomes of interest. Such iterative approaches will also lead to ways to validate self-report measures and integrate or triangulate the tracking of online activities with observational data and social network and support activities that are conducted offline.

Given the rapid evolution of the field of online communications and smoking-cessation interventions, and the numerous disciplines involved, we will need more agreement and standardization on metrics. For example, assessing norms and answering questions about their impact on behavior will require the development and validation of new instruments to determine active norms in an online social network and their importance. This work will be a necessary precursor to any efforts to modify existing norms or introduce new norms into existing or evolving networks. Ultimately the refinement of theories, models, and interventions would benefit from the development of standardized measures not only for norms, but for virtually all metrics mentioned in this report. Such measures would ideally have good reliability and validity across different projects, organizations, and even disciplines. Establishing a set of core measures that should be used across studies of online social networks will help test and improve both internal and external validity and will enhance theory testing by ensuring robustness, generalizability, replicability, consistency, and convergent validity across studies.

There are several limitations to this report. The recommendations presented are dependent on the individuals present at the conference and the structure provided by the organizers. Different participants or a different structure undoubtedly would have produced different questions and topics. The research priorities and recommendations presented here are but one set of views that we hope will serve to stimulate additional dialogue and research efforts. Addressing the questions posed in this report will present significant, but not insurmountable, challenges around personal privacy and the ethical treatment of research participants and their social contacts. Behavioral and biomedical researchers have traditionally thought about the impact on individuals, but social network interventions will challenge us to draw on the experience of public health professionals, social marketers, and sociologists as we increasingly target networks.

Networks and technology evolve on their own timeline, independent of the needs, funding, or aims of researchers. The study of rapidly evolving networks will require investigators and funders to tighten their timelines through the entire process (from idea, to funding, to execution, to publication). The traditional models of funding research via federal grants such as those in place at the National Science Foundation or National Institutes of Health in the United States are notoriously slow compared with industry and entrepreneurial interests. Network science and online interventions are changing rapidly and the traditional funding models must adapt as well. In 2009, Lazer and colleagues voiced concerns that research on large-scale networks “could become the exclusive domain of private companies and government agencies” [21], an outcome they noted would not be in the public interest. Developing and maintaining a strong academic research program is imperative and will require adjustments by funders, researchers, and publishers of scientific research.

Research efforts designed to address the topics and questions in this report may help identify mechanisms to significantly decrease the burden of tobacco related disease in the United States and elsewhere. The core ideas and themes developed here for smoking cessation may also apply—recognizing differences in context—to a variety of behaviors (eg, obesity, substance abuse, or adherence to medical recommendations) that could directly or indirectly improve the well-being and quality of life of our society. It is important to recognize that the powerful forces and rapid transmission of information across networks may also be used inappropriately or destructively (both intentionally and unintentionally) as well as for doing good. Ultimately, we hope that the kinds of research efforts encouraged in this paper will give rise to a new generation of interventions to help people quit smoking and stay quit, delivered and spread through a variety of social networks—networks that we recognize today, and networks that will develop tomorrow.

 Multimedia Appendix 1

Conference introduction, Dr. David Abrams.

 Multimedia Appendix 2

Conference introduction - Dr. Saul Shiffman.

 Multimedia Appendix 3

Theme
 1: Social
 support, 
health
 behavior
 and
 smoking 
cessation - Dr. Robin Mermelstein.

 Multimedia Appendix 4

Theme
 1: Social
 support, 
health
 behavior
 and
 smoking 
cessation - Dr. Thomas Valente.

 Multimedia Appendix 5

Conference keynote - Dr. Nicholas Christakis.

 Multimedia Appendix 6

Theme
 2:
 Online
 social 
networks - Dr. Nathan Cobb.

 Multimedia Appendix 7

Theme
 2:
 Online
 social 
networks - Dr. Noshir Contractor.

 Multimedia Appendix 8

Theme
 2:
 Online
 social 
networks - Dr. Nathan Eagle.

 Multimedia Appendix 9

Theme 
3:
 Intervention 
approaches - Mr. Dave Heilmann.

 Multimedia Appendix 10

Theme 
3:
 Intervention 
approaches - Mr. Trevor va Mierlo.

 Multimedia Appendix 11

Theme 
3:
 Intervention 
approaches - Mr. Chris Cartter.

 Multimedia Appendix 12

Theme
 4: 
Methods, design
 and
 analysis - Dr. Linda Collins.

 Multimedia Appendix 13

Theme
 4: 
Methods, design
 and
 analysis - Dr. Tom Snijders.

The workshop was funded by Healthways Inc., the National Cancer Institute of the National Institutes of Health, and the Johns Hopkins Bloomberg School of Public Health in conjunction with the American Legacy Foundation. External sponsors had no role in workshop planning or in the writing or publication of this report. The authors wish to thank all of the participants for their time and effort and extremely valuable contributions. We especially appreciate the speakers and planning committee without whom the workshop would not have been possible.

 

List of workshop participants:

David Abrams*, PhD; Schroeder Institute for Tobacco Research & Policy Studies, Legacy

Jas Ahluwalia, MD; University of Minnesota

Larry An, MD; University of Michigan

Eric Asche; Legacy

Audie Atienza, PhD; National Institutes of Health

Erik Augustson*, PhD, MPH; National Cancer Institute, NIH

Cathy Backinger*, PhD, MPH; National Cancer Institute, NIH

Cathy Baker, PhD, RN; Case Western Reserve University

Carla Berg, PhD; Emory University

Greg Bloss, MA; National Institute on Alcohol Abuse and Alcoholism, NIH

Georgiy Bobashev, PhD; RTI International

Janet Brigham, PhD; SRI International

Joanne Brown, ARNP; University of Kentucky

Taneisha Buchanan, PhD; University of Minnesota

David Buller, PhD; Klein Buendel, Inc.

M. Justin Byron*, MHS; Schroeder Institute for Tobacco Research & Policy Studies, Legacy

Chris Cartter; MeYou Health, LLC

Damon Centola, PhD; Massachusetts Institute of Technology

Nicholas Christakis, MD, MPH, PhD; Harvard Medical School

Nathan Cobb*, MD; Schroeder Institute for Tobacco Research & Policy Studies, Legacy

Sheldon Cohen*, PhD; Carnegie Mellon University

Trevor Cohen, MBChB, PhD; The University of Texas

Linda Collins, PhD; Penn State University

Noshir Contractor*, PhD; Northwestern University

Mary E. Cooley, PhD, RN; Dana Farber Cancer Institute

Laurel Curry; Legacy

Lowell C. Dale, MD; Mayo Clinic Tobacco Quitline

RaeAnne Davis, MSPH; North American Quitline Consortium

Nathan Eagle, PhD; The MIT Design Laboratory, Massachusetts Institute of Technology

Jason Fletcher, PhD; Yale University

Susannah Fox; Pew Internet Project

Judy Freeman; Schroeder Institute for Tobacco Research & Policy Studies, Legacy

Gabe Garcia, PhD; University of Alaska Anchorage

Joe Gitchell; Pinney Associates, Inc.

Amanda Graham*, PhD; Schroeder Institute for Tobacco Research & Policy Studies, Legacy

Anne M. Hartman, MS, MA; National Cancer Institute, NIH

Colleen Haydon, MSW, MPH; CYAN/Project UNIFORM

Dave Heilmann; SparkPeople, Inc.

Thaddeus Herzog, PhD; Cancer Research Center of Hawaii

Kimberlee Homer Vagadori, MPH; California Youth Advocacy Network

Thomas Houston, MD; University of Massachusetts Medical School

Yvonne Hunt, PhD, MPH; National Cancer Institute, NIH

Kevin Hwang, MD, MPH; The University of Texas Medical School at Houston

Caroline Joyce; Legacy

Ross Kauffman, PhD; Indiana University

Katie Kemper, MBA; Mayo Clinic

Tom Kirchner, PhD; Schroeder Institute for Tobacco Research & Policy Studies, Legacy

Amy Knowlton, ScD; Johns Hopkins Bloomberg School of Public Health

Emily Z. Kontos, ScD; Harvard University, School of Public Health

Sanjay Koyani, MPH; FDA/Center for Tobacco Products

Janna Lacatell, MBA; Healthways Inc.

Stephanie R. Land, PhD; University of Pittsburgh and ReSET Center

Yuelin Li, PhD; Memorial Sloan-Kettering Cancer Center

Paula Lozano, MD, MPH; University of Washington

Patty Mabry*, PhD; Office of Behavioral and Social Sciences Research, NIH

Michael Macy, PhD; Cornell University

Stephen Marcus, PhD; National Institute of General Medical Sciences, NIH

Darren Mays, PhD; Georgetown University

Anna McDaniel, PhD, RN; Indiana University

Karen McDonnell, PhD, MSN, RN; University of Virginia

Howard Meitiner; Phoenix House

Robin Mermelstein*, PhD; University of Illinois at Chicago

Aaron Mushro; Legacy

Toben Nelson, ScD; University of Minnesota

Ray Niaura*, PhD; Schroeder Institute for Tobacco Research & Policy Studies, Legacy

Janet Okamoto, PhD; National Cancer Institute, NIH

George Papandonatos*, PhD; Brown University

Heather Patrick, PhD; National Cancer Institute, NIH

Pallavi Patwardhan, PhD; Schroeder Institute for Tobacco Research & Policy Studies, Legacy

Jennifer Pearson, MPH; Schroeder Institute for Tobacco Research & Policy Studies, Legacy

Alan Peters, CTTS-M; QuitNet / Healthways

Alison Pilsner, MPH, CPH, CHES; MMG, Inc.

Craig Pollack, MD; Johns Hopkins University

Danielle Ramo, PhD; University of California, San Francisco

Brandi Robinson, MPH; Partnership for Prevention

Heather Rogers, PhD, MPH; Concurrent Technologies Corporation

Marcel Salathé, PhD; Penn State University

Saul Shiffman*, PhD; University of Pittsburgh

Tom Snijders, PhD; University of Oxford

Karen Sodomick, MA; Phoenix House

Mike Spittel, PhD; National Institute of Child Health & Human Development, NIH

Bonnie Spring, PhD; Northwestern University

Cassandra Stanton*, PhD; The Warren Alpert Medical School of Brown University

Maggy Sterner; Small World consulting

Erin L. Sutfin, PhD; Wake Forest University School of Medicine

Samarth Swarup, PhD; Virginia Tech

Shani C. Taylor; MMG, Inc.

Hilary Tindle, MD, MPH; University of Pittsburgh

Tom Valente, PhD; University of Southern California Keck School of Medicine

Donna Vallone*, PhD, MPH; Legacy

Trevor van Mierlo, MScCH; Evolution Health Systems Inc.

Andrea Villanti, MPH, PhD(c); Schroeder Institute for Tobacco Research & Policy Studies, Legacy

J. Lee Westmaas, PHD; American Cancer Society

Robyn Whittaker, MD; University of Auckland

* Participant in the conference planning committee and/or post-workshop summary workgroup.

Dr Cobb is a consultant to Healthways Inc., which operates QuitNet, a web-based smoking cessation application using social networks.

 Centers for Disease Control and Prevention (CDC) Smoking-attributable mortality, years of potential life lost, and productivity losses--United States, 2000-2004 MMWR Morb Mortal Wkly Rep 2008 11 14 57 45 1226 8 19008791 mm5745a3 Centers for Disease Control and Prevention (CDC) Cigarette smoking among adults and trends in smoking cessation - United States, 2008 MMWR Morb Mortal Wkly Rep 2009 11 13 58 44 1227 32 19910909 mm5844a2 US Department of Health and Human Services (USDHHS) Treating Tobacco Use and Dependence: 2008 Update 2008 Rockville, MD US Department of Health and Human Services, Public Health Services Shiffman S Brockwell SE Pillitteri JL Gitchell JG Use of smoking-cessation treatments in the United States Am J Prev Med 2008 02 34 2 102 11 10.1016/j.amepre.2007.09.033 18201639 S0749-3797(07)00666-6 Healthy People 2020 US Department of Health and Human Services 2010 12 2 2011-06-22 Tobacco Use Objectiveshttp://www.healthypeople.gov/2020/topicsobjectives2020/objectiveslist.aspx?topicId=41 5zdgFH0ss Westmaas JL Bontemps-Jones J Bauer JE Social support in smoking cessation: reconciling theory and evidence Nicotine Tob Res 2010 07 12 7 695 707 10.1093/ntr/ntq077 20513695 ntq077 Cohen S Lemay EP Why would social networks be linked to affect and health practices? Health Psychol 2007 07 26 4 410 7 10.1037/0278-6133.26.4.410 17605560 2007-09406-004 Christakis NA Health care in a web BMJ 2008 06 28 336 7659 1468 10.1136/bmj.a452 18583678 336/7659/1468 PMC2440901 May S West R Do social support interventions (“buddy systems”) aid smoking cessation? A review tobacco control Tob Control 2000 12 9 4 415 22 11106712 PMC1748387 Lichtenstein E Glasgow RE Abrams DB Social support in smoking cessation: in search of effective interventions Behav Ther 1986 17 607 19 10.1016/S0005-7894(86)80098-3 Murray RP Johnston JJ Dolce JJ Lee WW O'Hara P Social support for smoking cessation and abstinence: the Lung Health Study. Lung Health Study Research Group Addict Behav 1995 20 2 159 70 7484310 S030646039980001X Park EW Tudiver F Schultz JK Campbell T Does enhancing partner support and interaction improve smoking cessation? A meta-analysis Ann Fam Med 2004 2 2 170 4 15083859 PMC1466648 Graham AL Cobb NK Papandonatos GD Moreno JL Kang H Tinkelman DG Bock BC Niaura RS Abrams DB A randomized trial of Internet and telephone treatment for smoking cessation Arch Intern Med 2011 01 10 171 1 46 53 10.1001/archinternmed.2010.451 21220660 171/1/46 Selby P van Mierlo T Voci SC Parent D Cunningham JA Online social and professional support for smokers trying to quit: an exploration of first time posts from 2562 members J Med Internet Res 2010 12 3 e34 10.2196/jmir.1340 20719739 v12i3e34 PMC2956324 Burri M Baujard V Etter JF A qualitative analysis of an internet discussion forum for recent ex-smokers Nicotine Tob Res 2006 12 8 Suppl 1 S13 9 17491166 Cobb NK Graham AL Abrams DB Social network structure of a large online community for smoking cessation Am J Public Health 2010 07 100 7 1282 9 10.2105/AJPH.2009.165449 20466971 AJPH.2009.165449 comScore, Inc 2008 11 19 2011-06-22 Eighty Five Percent of Brazilian Internet Users Visited a Social Networking Site in September 2008http://www.comscore.com/Press_Events/Press_Releases/2008/11/Social_Networking_in_Brazil 5zdhC6oRb Facebook 2010 10 1 Facebook press roomhttp://www.facebook.com/press/info.php?statistics 5tyQEzO17 Shahab L McEwen A Online support for smoking cessation: a systematic review of the literature Addiction 2009 11 104 11 1792 804 10.1111/j.1360-0443.2009.02710.x 19832783 ADD2710 Norman CD McIntosh S Selby P Eysenbach G Web-assisted tobacco interventions: empowering change in the global fight for the public's (e)Health J Med Internet Res 2008 10 5 e48 10.2196/jmir.1171 19033147 v10i5e48 PMC2630840 Lazer D Pentland A Adamic L Aral S Barabasi AL Brewer D Christakis N Contractor N Fowler J Gutmann M Jebara T King G Macy M Roy D Van Alstyne M Social science. Computational social science Science 2009 02 6 323 5915 721 3 10.1126/science.1167742 19197046 323/5915/721 PMC2745217 Anonymous Community cleverness required Nature 2008 455 7209 1 18769385 Rosenquist JN Murabito J Fowler JH Christakis NA The spread of alcohol consumption behavior in a large social network Ann Intern Med 2010 04 6 152 7 426 33, W141 10.1059/0003-4819-152-7-201004060-00007 20368648 152/7/426 Christakis NA Fowler JH The spread of obesity in a large social network over 32 years N Engl J Med 2007 07 26 357 4 370 9 10.1056/NEJMsa066082 17652652 NEJMsa066082 Hammond RA Social influence and obesity Curr Opin Endocrinol Diabetes Obes 2010 10 17 5 467 71 10.1097/MED.0b013e32833d4687 20689421 Fowler JH Christakis NA Dynamic spread of happiness in a large social network: longitudinal analysis over 20 years in the Framingham Heart Study BMJ 2008 337 a2338 19056788 PMC2600606 Rosenquist JN Fowler JH Christakis NA Social network determinants of depression Mol Psychiatry 2011 03 16 3 273 81 10.1038/mp.2010.13 20231839 mp201013 Cacioppo JT Fowler JH Christakis NA Alone in the crowd: the structure and spread of loneliness in a large social network J Pers Soc Psychol 2009 12 97 6 977 91 10.1037/a0016076 19968414 2009-22579-013 PMC2792572 Moreno JL Who Shall Survive? A New Approach to the Problem of Human Interrelations 1934 Washington, DC Nervous and Mental Disease Publishing Co Mabry PL Olster DH Morgan GD Abrams DB Interdisciplinarity and systems science to improve population health: a view from the NIH Office of Behavioral and Social Sciences Research Am J Prev Med 2008 08 35 2 Suppl S211 24 10.1016/j.amepre.2008.05.018 18619402 S0749-3797(08)00431-5 PMC2587290 Graham AL Abrams DB Reducing the cancer burden of lifestyle factors: opportunities and challenges of the Internet J Med Internet Res 2005 07 1 7 3 e26 10.2196/jmir.7.3.e26 15998617 v7i3e26 PMC1550652 Bandura A Social Foundations of Thought and Action: A Social Cognitive Theory 1986 Englewood Cliffs, NJ Prentice-Hall Mermelstein R Cohen S Lichtenstein E Baer JS Kamarck T Social support and smoking cessation and maintenance J Consult Clin Psychol 1986 08 54 4 447 53 3745596 Steglich C Snijders TAB Pearson M Dynamic networks and behavior: separating selection from influence Sociol Methodol 2010 08 40 1 329 93 10.1111/j.1467-9531.2010.01225.x Simmel G On Individuality and Social Forms: Selected Writings 1971 Chicago, IL University of Chicago Press Berkman LF Syme SL Social networks, host resistance, and mortality: a nine-year follow-up study of Alameda County residents Am J Epidemiol 1979 02 109 2 186 204 425958 Berkman LF Glass T Social integration, social networks, social support, and health Berkman LF, Kawachi I, editors. Social Epidemiology 2000 New York, NY Oxford University Press 137 73 Paul MJ Dredze M You are what you tweet: analyzing twitter for public health Proceedings 2011 Fifth International AAAI Conference on Weblogs and Social Media (ICWSM) July 17-21, 2011 Barcelona, Spain Centola D The spread of behavior in an online social network experiment Science 2010 09 3 329 5996 1194 7 10.1126/science.1185231 20813952 329/5996/1194 Aral S Muchnik L Sundararajan A Distinguishing influence-based contagion from homophily-driven diffusion in dynamic networks Proc Natl Acad Sci U S A 2009 12 22 106 51 21544 9 10.1073/pnas.0908800106 20007780 0908800106 PMC2799846 Salganik MJ Dodds PS Watts DJ Experimental study of inequality and unpredictability in an artificial cultural market Science 2006 02 10 311 5762 854 6 10.1126/science.1121066 16469928 311/5762/854 Kossinets G Watts DJ Empirical analysis of an evolving social network Science 2006 01 6 311 5757 88 90 10.1126/science.1116869 16400149 311/5757/88 Eagle N Macy M Claxton R Network diversity and economic development Science 2010 05 21 328 5981 1029 31 10.1126/science.1186605 20489022 328/5981/1029 Newman MEJ The structure and function of complex networks SIAM Rev 2003 45 2 167 256 10.1137/S003614450342480 Abrams DB Leslie F Mermelstein R Kobus K Clayton RR Transdisciplinary tobacco use research Nicotine Tob Res 2003 12 5 Suppl 1 S5 10 10.1080/14622200310001625519 14668083 9B28JRFL6KNWBWB0 Marin A Wellman B Scott PJP Carrington PP Social network analysis: an introduction Scott J, Carrington PJ, editors. The SAGE Handbook of Social Network Analysis 2011 Thousand Oaks, CA SAGE Publications 11 25 Bock BC Graham AL Whiteley JA Stoddard JL A review of web-assisted tobacco interventions (WATIs) J Med Internet Res 2008 10 5 e39 10.2196/jmir.989 19000979 v10i5e39 PMC2630838 Weitzman ER Cole E Kaci L Mandl KD Social but safe? Quality and safety of diabetes-related online social networks J Am Med Inform Assoc 2011 05 1 18 3 292 7 10.1136/jamia.2010.009712 21262920 jamia.2010.009712 PMC3078678 Frost JH Massagli MP Social uses of personal health information within PatientsLikeMe, an online patient community: what can happen when patients have access to one another's data J Med Internet Res 2008 10 3 e15 10.2196/jmir.1053 18504244 v10i3e15 PMC2553248 McPherson M Smith-Lovin L Cook JM Birds of a feather: homophily in social networks Ann Rev Sociol 2001 27 415 44 10.1146/annurev.soc.27.1.415 Rogers EM Diffusion of Innovations 2003 New York, NY Free Press Iyengar R Van den Bulte C Valente TW Opinion leadership and social contagion in new product diffusion Marketing Sci 2010 30 2 195 212 10.1287/mksc.1100.0566 Watts DJ Peretti J Viral marketing for the real world Harvard Bus Rev 2007 85 5 22 3 Lessig L Code and Other Laws of Cyberspace 1999 New York, NY Basic Books Lampe CAC Ellison N Steinfield C A familiar face(book): profile elements as signals in an online social network Proceedings 2007 SIGCHI Conference on Human Factors in Computing Systems April 28-May 3, 2007 San Jose, CA, USA Bansal MA Cummings KM Hyland A Giovino GA Stop-smoking medications: who uses them, who misuses them, and who is misinformed about them? Nicotine Tob Res 2004 12 6 Suppl 3 S303 10 15799593 An LC Schillo BA Saul JE Wendling AH Klatt CM Berg CJ Ahulwalia JS Kavanaugh AM Christenson M Luxenberg MG Utilization of smoking cessation informational, interactive, and online community resources as predictors of abstinence: cohort study J Med Internet Res 2008 10 5 e55 10.2196/jmir.1018 19103587 v10i5e55 PMC2630836 Stoddard JL Augustson EM Moser RP Effect of adding a virtual community (bulletin board) to smokefree.gov: randomized controlled trial J Med Internet Res 2008 10 5 e53 10.2196/jmir.1124 19097974 v10i5e53 PMC2630832 Eysenbach G The law of attrition J Med Internet Res 2005 7 1 e11 10.2196/jmir.7.1.e11 15829473 v7e11 PMC1550631 Cobb NK Graham AL Bock BC Papandonatos G Abrams DB Initial evaluation of a real-world Internet smoking cessation system Nicotine Tob Res 2005 04 7 2 207 16 10.1080/14622200500055319 16036277 M751036L54TH3674 PMC2885701 Graham AL Bock BC Cobb NK Niaura R Abrams DB Characteristics of smokers reached and recruited to an internet smoking cessation trial: a case of denominators Nicotine Tob Res 2006 12 8 Suppl 1 S43 8 17491170 PMC2885695 Etter JF Comparing the efficacy of two Internet-based, computer-tailored smoking cessation programs: a randomized trial J Med Internet Res 2005 7 1 e2 10.2196/jmir.7.1.e2 15829474 v7e2 PMC1550632 Cobb NK Graham AL Characterizing Internet searchers of smoking cessation information J Med Internet Res 2006 8 3 e17 10.2196/jmir.8.3.e17 17032633 v8i3e17 PMC2018828 Vallone DM Duke JC Cullen J McCausland KL Allen JA Evaluation of EX: a national mass media smoking cessation campaign Am J Public Health 2011 02 101 2 302 9 10.2105/AJPH.2009.190454 21164094 AJPH.2009.190454 Bala M Strzeszynski L Cahill K Mass media interventions for smoking cessation in adults Cochrane Database Syst Rev 2008 1 CD004704 10.1002/14651858.CD004704.pub2 18254058 Abrams DB Mills S Bulger D Challenges and future directions for tailored communication research Ann Behav Med 1999 21 4 299 306 10721436 Strecher VJ McClure JB Alexander GL Chakraborty B Nair VN Konkel JM Greene SM Collins LM Carlier CC Wiese CJ Little RJ Pomerleau CS Pomerleau OF Web-based smoking-cessation programs: results of a randomized trial Am J Prev Med 2008 05 34 5 373 81 10.1016/j.amepre.2007.12.024 18407003 S0749-3797(08)00150-5 PMC2697448 Zbikowski SM Hapgood J Smucker Barnwell S McAfee T Phone and web-based tobacco cessation treatment: real-world utilization patterns and outcomes for 11,000 tobacco users J Med Internet Res 2008 10 5 e41 10.2196/jmir.999 19017583 v10i5e41 PMC2630835 Zbikowski SM Jack LM McClure JB Deprey M Javitz HS McAfee TA Catz SL Richards J Bush T Swan GE Utilization of services in a randomized trial testing phone- and web-based interventions for smoking cessation Nicotine Tob Res 2011 05 13 5 319 27 10.1093/ntr/ntq257 21330267 ntq257 PMC3082503 Collins LM Baker TB Mermelstein RJ Piper ME Jorenby DE Smith SS Christiansen BA Schlam TR Cook JW Fiore MC The multiphase optimization strategy for engineering effective tobacco use interventions Ann Behav Med 2011 04 41 2 208 26 10.1007/s12160-010-9253-x 21132416 PMC3053423 Pike KJ Rabius V McAlister A Geiger A American Cancer Society's QuitLink: randomized trial of Internet assistance Nicotine Tob Res 2007 03 9 3 415 20 10.1080/14622200701188877 17365773 772841762 Strecher VJ Shiffman S West R Randomized controlled trial of a web-based computer-tailored smoking cessation program as a supplement to nicotine patch therapy Addiction 2005 05 100 5 682 8 10.1111/j.1360-0443.2005.01093.x 15847626 ADD1093 Civljak M Sheikh A Stead LF Car J Internet-based interventions for smoking cessation Cochrane Database Syst Rev 2010 9 CD007078 10.1002/14651858.CD007078.pub3 20824856 Swartz LH Noell JW Schroeder SW Ary DV A randomised control study of a fully automated internet based smoking cessation programme Tob Control 2006 02 15 1 7 12 10.1136/tc.2003.006189 16436397 15/1/7 PMC2563640 Muñoz RF Barrera AZ Delucchi K Penilla C Torres LD Pérez-Stable EJ International Spanish/English Internet smoking cessation trial yields 20% abstinence rates at 1 year Nicotine Tob Res 2009 09 11 9 1025 34 10.1093/ntr/ntp090 19640833 ntp090 PMC2725004 Japuntich SJ Zehner ME Smith SS Jorenby DE Valdez JA Fiore MC Baker TB Gustafson DH Smoking cessation via the internet: a randomized clinical trial of an internet intervention as adjuvant treatment in a smoking cessation intervention Nicotine Tob Res 2006 12 8 Suppl 1 S59 67 17491172 Danaher BG Seeley JR Methodological issues in research on web-based behavioral interventions Ann Behav Med 2009 08 38 1 28 39 10.1007/s12160-009-9129-0 19806416 Eysenbach G Issues in evaluating health websites in an Internet-based randomized controlled trial J Med Internet Res 2002 12 4 3 E17 10.2196/jmir.4.3.e17 12554548 PMC1761938 Glasgow RE Enhancing the scientific foundation of internet intervention research Ann Behav Med 2009 08 38 1 46 7 10.1007/s12160-009-9134-3 19816749 Ahern DK Challenges and opportunities of eHealth research Am J Prev Med 2007 05 32 5 Suppl S75 82 10.1016/j.amepre.2007.01.016 17466822 S0749-3797(07)00045-1 Glasgow RE eHealth evaluation and dissemination research Am J Prev Med 2007 05 32 5 Suppl S119 26 10.1016/j.amepre.2007.01.023 17466816 S0749-3797(07)00052-9 Eagle N Pentland AS Lazer D Inferring friendship network structure by using mobile phone data Proc Natl Acad Sci U S A 2009 09 8 106 36 15274 8 10.1073/pnas.0900282106 19706491 0900282106 PMC2741241 Christakis NA Social networks and collateral health effects BMJ 2004 07 24 329 7459 184 5 10.1136/bmj.329.7459.184 15271805 329/7459/184 PMC487721 Galea S Riddle M Kaplan GA Causal thinking and complex system approaches in epidemiology Int J Epidemiol 2010 02 39 1 97 106 10.1093/ije/dyp296 19820105 dyp296 PMC2912489 Levy DT Mabry PL Graham AL Orleans CT Abrams DB Reaching Healthy People 2010 by 2013: A SimSmoke simulation Am J Prev Med 2010 03 38 3 Suppl S373 81 10.1016/j.amepre.2009.11.018 20176310 S0749-3797(09)00859-9 PMC2843137 Killeen PR Markov model of smoking cessation Proc Natl Acad Sci U S A 2011 09 13 108 Suppl 3 15549 56 10.1073/pnas.1011277108 21508324 1011277108 PMC3176613 Abrams DB Bonnie RJ Stratton KR Wallace RB Comprehensive smoking cessation policy for all smokers: systems integration to save lives and money Ending the Tobacco Problem: A Blueprint for the Nation 2007 Washington, DC National Academies Press Abrams DB Graham AL Levy DT Mabry PL Orleans CT Boosting population quits through evidence-based cessation treatment and policy Am J Prev Med 2010 03 38 3 Suppl S351 63 10.1016/j.amepre.2009.12.011 20176308 S0749-3797(09)00893-9 Levy DT Graham AL Mabry PL Abrams DB Orleans CT Modeling the impact of smoking-cessation treatment policies on quit rates Am J Prev Med 2010 03 38 3 Suppl S364 72 10.1016/j.amepre.2009.11.016 20176309 S0749-3797(09)00857-5 PMC2843145 Morell JA Hilscher R Magura S Ford J Integrating evaluation and agent-based modeling: rationale and an example for adopting evidence-based practices J Multidiscip Eval 2010 08 6 14 32 57