Use and Taxonomy of Social Media in Cancer-Related Research: A Systematic Review

Abstract

Little is known about how social media are used in cancer care. We conducted a systematic review of the use and taxonomy of social media in cancer-related studies, in PubMed, Web of Knowledge, CINAHL, and Google Scholar. We located 1350 articles published through October 2013; 69 met study inclusion criteria.

Early research (1996–2007) was predominantly descriptive studies of online forums. Later, researchers began analyzing blogs, videos shared on YouTube, and social networking sites. Most studies (n = 62) were descriptive, and only 7 reported intervention studies published since 2010.

Future research should include more intervention studies to determine how social media can influence behavior, and more empirical research is needed on how social media may be used to reduce health disparities.

Social media are defined as interactive Web sites that enable users to create, share, comment on, and, often, modify content.¹ The Centers for Disease Control and Prevention lists the following as social media sites: social networking sites, blogs, microblogs, RSS (Really Simple Syndication) feeds, and online forums.² A more complete list of social media types is shown in the box on the next page. With social media, the public largely controls what content is created and circulated online, including health-related information. Researchers are interested in social media because the public and health practitioners are beginning to use these sites to connect individuals to reliable health promotion information, address impediments to health literacy, enhance patients’ communication with health care providers, and invite diverse audiences to participate in research studies, such as clinical trials.³ In short, social media sites can serve as key health communication channels to increase the bidirectional flow of information.⁴ The sites provide a location for online dialogue and encourage individuals and communities to interact by providing their opinions, personal accounts, and other information related to disease prevention, early detection, treatment, and survivorship. The Centers for Disease Control and Prevention adds that social media sites enable users to share health information in a timely manner and empower users to make health decisions informed by the contributions of other visitors.²

Examples of Social Media and Other Types of Electronic Communication

Social Media	Not Social Media
Blogs	E-mail
Microblogs	Listserv
Social networking sites	Regular Web page
Social bookmarking sites	Prerecorded podcast
Online support groups (online forum)
Online message boards (online forum)
Video-sharing sites
Photo-sharing sites
Virtual worlds
Wikis
Some forms of online games

Little is known about how social media are used in health care research. Chou et al. conducted a systematic review of social media sites (Web sites hosted on the Web 2.0 platform) in general health promotion.⁵ They recommended that more intervention studies should use social media and that social media measurement tools and methods should be developed and refined for future research. Public health researchers, including those whose focus is cancer, are beginning to examine how social media are changing the health communication landscape.³

The type of information shared, purpose of use, and application of social media sites for cancer-related information may differ from the use of social media sites for general health promotion and in other health contexts. In particular, it is unclear whether the complexity of cancer treatment and survivorship and the need for provider and caregiver involvement influence how social media sites are used for health information. For example, individuals or organizations who search for general health promotion information (nonspecific to particular diseases and treatments) on social media sites may be looking for information about general positive lifestyle behaviors (e.g., diet, nutrition, smoking cessation). Individuals who browse social media sites for cancer-related information, however, may be searching for more specific content related to the different stages of the cancer continuum (prevention, diagnosis, treatment, treatment aftermath). They may also search for online social support to share or understand others’ cancer-related experiences. For example, cancer survivors and their caregivers may more actively seek support and search for information about treatment, treatment side effects, and symptom management.

Interventions for cancer survivors may also entail challenges in study implementation arising from the health of study participants. Similarly, it is unclear whether interventions that use social media sites as the vehicle for cancer communication would elicit similar efficacy or participation levels as those reported by Chou et al.⁵ Nevertheless, in light of the potential power of social media to affect cancer information sharing and seeking and the enactment of prevention behaviors, it is important to explore how these new, widely and increasingly accepted communication channels can be used across the cancer continuum.

Communication technologies that allow users to access social media sites (e.g., computers, tablets, smartphones) are increasingly available. Newer devices can bring information to wider populations, including those who have had limited computer Internet access and have traditionally have been affected by the digital divide.^5,6 Chou et al. have examined ways that social media can narrow this communication digital divide for health promotion.⁵ Similarly, it is important to examine how social media may bring cancer-related information and interventions to populations affected by the digital divide and subject to disparities in both cancer and cancer communication.

In response to this gap in knowledge, we conducted a systematic review of peer-reviewed studies specific to social media and cancer. We aimed to (1) examine the taxonomy and time line of social media sites used in cancer care, and (2) categorize the designs of studies that assessed the role of social media in the context of cancer care. Although it was not a primary aim of our review, we also examined the extent to which cancer-related articles published on social media discussed the digital divide, health literacy, and the potential impact of social media interventions on cancer disparities.

METHODS

We conducted a systematic review of articles published through October 2013 (with no specified beginning date) and followed the Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines.⁷

With the assistance of a medical librarian, we developed a comprehensive search strategy for 3 medical–health care databases (PubMed, Web of Knowledge, CINAHL) and Google Scholar. We used the following search strategy and search terms: (social media OR “social media”) AND cancer; (Web 2.0 OR “Web 2.0”) AND cancer; (Medicine 2.0 AND “Medicine 2.0”) AND cancer; (M-health 2.0 AND “M-health”) AND cancer; (E-health OR “E-health”) AND cancer; (social networking site AND “social networking site”) AND cancer; (social bookmarking OR “social bookmarking”) AND cancer; (discussion thread OR “discussion thread”) AND cancer; (message board OR “message board”) AND cancer; (blogs OR “blogs”) AND cancer; (online communities OR “online communities”) AND cancer; (RSS OR “RSS”) AND cancer; (photo sharing OR “photo sharing”) AND cancer; (video sharing OR “video sharing”) AND cancer; (wikis OR “wikis”) AND cancer; (virtual world OR “virtual world”) AND cancer; (phone application OR “phone application”) AND cancer; (interactive media OR “interactive media”) AND cancer; (twitter OR “Twitter”) AND cancer; (myspace OR “MySpace”) AND cancer; (facebook OR “Facebook”) AND cancer; (youtube OR “YouTube”) AND cancer; (Second Life OR “Second Life”) AND cancer; (Smart Phone OR “Smart Phone”) AND cancer.

Eligibility Criteria

Articles that focused on social media across the cancer continuum were eligible for inclusion. We applied no restrictions on the year articles were published. We excluded articles if they were not written in English, did not specifically discuss both cancer and social media, only mentioned social media without actual application of social media, were not research articles (e.g., editorials, letters to the editor, review articles), or were gray literature (e.g., non–peer-reviewed literature such as dissertations, unpublished manuscripts, and book chapters). Figure 1 illustrates article selection according to our eligibility criteria.

FIGURE 1—

Flowchart of article selection and inclusion criteria for systematic review of social media (SM) in cancer research.

Four authors independently scanned the titles and abstracts of the articles found in the search to assess their eligibility for the systematic review. When we could not make a decision regarding study inclusion from the study title and abstract, we retrieved the full text of the article and evaluated the content for eligibility. Another author arbitrated and resolved any disagreements.

Data Collection

We created a standardized data extraction form, and the entire team reviewed it before use. Four authors independently extracted data from all included studies (2 authors independently reviewed each article) according to the following elements:

• 1. type of cancer,

• 2. type of social media site,

• 3. content of social media,

• 4. how social media was used in the study (e.g., purpose, such as sharing information, recruitment),

• 5. study design,

• 6. data collection procedures,

• 7. characteristics of participant population (e.g., gender, race, ethnicity, age),

• 8. key findings of the study, and

• 9. whether the study described how the use of social media sites related to the digital divide, health disparities, or health literacy.

We reviewed all extracted data for accuracy. When we identified discrepancies, an additional author served as an arbitrator to reconcile these differences. Our team met several times to discuss the data, achieve consensus, and synthesize findings.

RESULTS

The initial search yielded 1350 articles. We excluded 635 duplicates, 410 articles that did not pertain to both social media and cancer, 220 that were not research articles, 11 that we considered to be gray literature, 3 literature reviews, and 2 that were written in a language other than English. We reviewed the 69 studies that met inclusion criteria (Figure 1).

Social Media Taxonomy and Time Line

The earliest studies on social media and cancer that we found were published in 1996.⁸ Early studies primarily described content and provided thematic analyses of messages posted on online cancer forums. From 1996 to 2013, online forums, such as discussion boards or forums, message boards, online support group Web sites (n = 38; 55.1%), were the most common form of social media reported in the studies.^8–45 Beginning in 2008,⁴⁶ researchers also began analyzing the use and content of blogs and blog responses, often examining reasons for and benefits of blogging during cancer treatment and survivorship (n = 7; 10.1%).^46–52 Published studies also specifically addressed cancer-related use of video-sharing sites such as YouTube (n = 5; 7.2%),^53–57 and social networking sites such as Facebook (n = 8; 11.6%).^58–65

Only 1 study (1.4%) explored the use of a virtual world, Second Life, to demonstrate how an interactive world can help to train physicians to share medical bad news, such as a cancer diagnosis.⁶⁶ Research describing the use of social bookmarking sites (e.g., Digg.com, Redditt.com) was also less common. One study (1.4%) assessed which Web sites that featured information about prostate cancer screening recommendations users were reading and sharing on the 4 most widely used social bookmarking sites.⁶⁷ Researchers began publishing articles (n = 3; 4.3%) on the use of microblogs, particularly Twitter, in 2012.^68–70 Also in 2012, articles (n = 8; 11.4%) began to appear on a combination of social media types, such as social networking sites and blogs or online videos and support groups.^{42,62,63,71–76}

Social Media by Cancer Focus and Type

Studies most often focused on breast cancer (n = 22; 31.9%),^{8,10,12,13,16,17,22,25,26,28–30,35,41–43,45,58,60,66,74,75} and prostate cancer (n = 4; 5.8%).^36,56,67,76 Other cancers mentioned were osteosarcoma,⁷³ rhabdomyosarcoma,⁵⁵ testicular,¹¹ skin,²³ cervical,^37,53,59,65 ovarian,^21, lung,³⁴ and colorectal.^14,24 Cancer in general was the focus of 21 articles (30.4%),^{9,15,19,20,31–33,38,40,44,46–48,52,54,57,62–64,70,72} and 11 (15.9%) discussed multiple types of cancer.^{18,27,34,39,49–51,61,68,69,71}

Most articles that described social media use in breast cancer care focused on content analysis of online forums for provision of information and emotional support as well as self-expression (n = 13; 18.8%).^{8,10,12,16,17,25,26,30,35,39,41,42,45} In the articles about cancer in general, authors described how social media users often posted content about cancer treatments (e.g., prescriptions, chemotherapy, radiation) as well as disease outcomes and expectations (n = 13; 18.8%).^{9,15,18,31–33,39,40,46,48,52,54,70} For example, 1 article described the use of a social media site to encourage women with gynecologic cancers to discuss their concerns related to sexual distress, anxiety, and depression arising from their diagnosis and treatment.¹⁸ Less often, researchers discussed research about cancer prevention information on social media sites (n = 3; 4.3%).^37,53,59 Two studies explored the type of information end users posted on social media sites about the human papillomavirus (HPV) and the HPV vaccine.^37,53 In another study, researchers used a social networking site to recruit participants to take a survey that measured individuals’ HPV and cervical cancer knowledge.

A total of 6 studies (8.7%) used social media sites to recruit participants for their research.^{23,59–62,74} In these studies, researchers located specific cancer-related Facebook groups or interest pages and either messaged individual members or posted a recruitment advertisement for their study on the group’s main Facebook page.

Social Media Types, Study Goals, and Designs

Most articles analyzed the content of social media sites (Table 1), such as online forums, emphasizing their use by participants as sounding boards to express emotions and to raise awareness about cancer, provide support for cancer survivors and caregivers, promote information sharing and problem solving, recommend specific treatments, and, less often, raise funds for cancer-related initiatives. Other descriptive studies (n = 12; 17.4%; Table 2) used surveys,^{11,23,24,28,40,43,46,59,60,63,65,74} in-depth interviews,⁴⁷ focus groups,⁴⁸ social network analyses,^20,68 a mix of measurement techniques,^{8,9,13,22,30,39} Web metrics,^72,73 and frequencies⁶² to describe how users employed social media to obtain and receive cancer information. It is important to note that only 1 of these studies reviewed the accuracy of information posted on a social media site (YouTube); it reported that roughly 75% of all videos had either fair or poor information about prostate cancer screening.⁵⁶

TABLE 1—

Social Media Cancer Studies With Descriptive Design and Content Analysis Focus

Type of Cancer/Study	Participants	Social Media Platforms/Study Goals	Key Findings
Online forums
Breast
Barker and Galardi10	BC survivors and current patients	Examine reactions of women with BC to the 2009 US Preventive Services Task Force recommendations for mammography screening from 5 BC discussion boards.	Women felt betrayed by new BRCA screening guidelines, believing they aimed to save insurance companies money rather than save lives.
Bender et al.12	White BC survivors	Identify the characteristics and levels of use of online communities for female BC survivors.	Retrieved 111 Web sites: 65% had a broad focus, 34% were exclusive to BC, 1% catered to specific disease characteristics. Most sites were American, nonprofit, and moderated by community members.
Chen16	Unknown	Characterize and compare online discussion forums for 3 conditions: BC, type 1 diabetes, and fibromyalgia from social networking site DailyStrength.	Clusters for each condition differed but all fit a set of common categories: generic, support, patient centered, experiential knowledge, treatments/procedures, medications, condition management.
Civan and Pratt17	BC survivors, current patients, and caregivers	Use online BC communities as a window to understand the prevalence and characteristics of informational support that peers exchanged and its role in facilitating patient expertise.	Support centered on collaborative problem solving for future and emergent events. Action strategies, knowledge, and information resources peers recommended reflected patients' expertise and offered a source for others to learn from as they faced similar situations.
Kenen et al.26	Women aged 30–49 y	Examine the physical, psychological, and social consequences of prophylactic mastectomy during decision-making, recuperation, and reconstruction process from posts on the FORCE Web site.	Themes included self-image and interpersonal relationships, feeling loss of natural breasts, perception of self–other stigma, shifts in friendship and workplace relationships, intrusion into mother–children relationship, and coming to terms with the decision.
Mao et al.35	BC patients	Understand the frequency and content of side effects and associated adherence behaviors discussed by female BC patients on BC message boards related to AIs, with particular emphasis on AI-related arthralgia.	Few AI-related posts (18%) mentioned ≥ 1 side effect (joint/musculoskeletal pain, hot flashes, osteoporosis, weight gain). In studies reporting AI use, 13% of patients reported discontinuing and 38% mentioned switching AIs. Many community members offered support and advice for coping with AI-associated arthralgia.
Shim et al.41,a	Medically underserved African American and White women with BC	Examine how women with BC benefit from written disclosure of insights and emotions in online support groups.	Overall, disclosure of insights in online groups had stronger effects on enhancing health benefits than did disclosure of emotions.
Sillence42	BC patients	Examine how peers exchange advice in an online BC support forum to better understand online groups as a resource for decision-making.	Advice was offered in ∼40% of all support group messages. Direct and hedged types of advice were the most common responses to requests for advice. Requests for opinions or information and problem disclosure were the most common types of advice solicitation.
Yoo et al.45	BC patients	Explore how expression of emotional support in CHESS, an online BC support group for midlife women, changes over time and what factors predict this pattern of change.	Individual expression of emotional support increased in first 12 wk of study, then decreased. Age, living situation, comfort level with computer and Internet, and coping strategies were important factors in predicting emotional support patterns.
Keeling et al.25	Women aged 40–65 y	Demonstrate construction of tacit knowledge with an online BC community for midlife women and how consumers exercise their “voice” within complex professional services.	Most medical information reported from patient perspective; participants negotiated knowledge in the absence of a formal normative hierarchy.
Ovarian: Gill and Whisnant21	Unknown	Illustrate role that online communities (http://medhelp.org) play in helping patients and their families deal with ovarian cancer.	Users participated in information seeker–information provider process, sharing both technical and emotional self-management knowledge while building trust and support.
Cervical: Penţa and Băban37	Romanian men and women	Explore public's constructions of HPV vaccine as expressed on Romanian Internet discussion forums.	Positive and negative discourses found; both used appeals to authority to provide power to messaging. Themes included mistrust, helpfulness of vaccine, uncertainty.
Breast and ovarian: Kenen et al.27	Women at high risk for BC and ovarian cancer	Understand the unique health concerns and emotional support needs of women with hereditary risk who post to Facing our Risk of Cancer Empowered discussion board.	Women sought emotional and informational support about prophylactic oophorectomies. Online exchange generated sense of trust and feeling of being in control during BC treatments.
Breast and lung: Lu et al.34	Patients with chronic conditions (cancer or diabetes), caregivers, and a few providers	Determine (by text clustering method) differences in interesting topics among different types of disease discussion boards (BC, lung cancer, diabetes) on Medhelp.org.	BC discussion boards were more likely to discuss examinations; lung cancer boards focused more on symptoms. BC patients discussed complications less often than lung cancer and diabetic patients.
Colorectal: Beusterien et al.14	Patients with advanced cancer	Explore impact of CRC treatments in patient Web forums.	Common themes were treatment side effects, impact on life, and emotional distress. Emotional impacts exemplified resilience and positive coping strategies.
General
Buis and Whitten15	Unknown	Determine how social support content in online support communities (Association of Cancer Online Resources and Yahoo!) for different cancers varies according to cancer survival rate.	Emotional support was more common than informational support. Communities with high survival rates had more emotional support content; those with lower survival rates had more informational support content.
Love et al.31	Young adults with cancer	Understand patients' exercise and nutrition habits, identify needs, and assess how notion of works helps define their efforts.	Nutrition and exercise needs affected illness trajectories, including management of illness, everyday life, and biographical and communication work.
Love et al.32	Adolescents and young adults with cancer	Identify topics discussed in an online forum.	Themes were exchanging support, coping, describing experiences, enacting identity, and communicating membership.
Love et al.33	Young adults with cancer	Describe topics, needs, and concerns related to nutrition and exercise posted on an online young adult cancer forum.	Participants described several functions achieved through nutrition and exercise: fighting cancer, losing weight, obtaining nutrients despite treatment, promoting general health and mental health, and enjoyment.
Schultz et al.38	Cancer patients and caregivers	Review traffic to University of Texas MD Anderson Cancer Center Web site and message board.	Most posters (60%) were cancer patients. Of nonpatients, 22% were spouses and 69% were family members of cancer survivors. BC patients most often posted about treatment, and lung cancer patients most often posted about support.
Street et al.44	Predominantly White, middle-aged Australian men and women.	Examine value of facilitated telephone and online support groups for palliative care.	Participant narratives focused on dying with dignity or on deferring discussion of death/dying to focus on the present. Participants reported these forums to be safe and accessible.
Video-sharing sites
Cervical: Ache and Wallace53	Unknown	Analyze how YouTube videos portray HPV vaccination.	YouTube videos more often positively portrayed HPV vaccine.
General: Chou et al.54	Cancer survivors	Describe authentic personal cancer stories through a linguistically based narrative analysis of YouTube videos.	Common themes were orientation of being diagnosed with cancer, series of complicating actions after diagnosis, evaluation of cancer diagnosis, results and resolution of diagnosis event.
Rhabdomyosarcoma: Clerici et al.55	Adolescent cancer survivors and parents of living and deceased pediatric cancer patients	Investigate availability and content of YouTube videos about pediatric neoplastic diseases (rhabdomyosarcoma and soft tissue sarcoma)	Of 149 videos analyzed, 25 were considered useful. Only 1 video was produced by a doctor; 82.5% were produced by patients or their families, and these mostly commemorated a child's death.
Prostate: Steinberg et al.56	Unknown	Analyze PC videos on YouTube for information content and bias.	Search found 14 videos on PSA, 5 on radiotherapy, and 32 on surgery; 73% had fair or poor information content. Bias for surgery, radiotherapy, or PSA testing present in 69%; none were biased against treatment or PSA testing.
Blogs
Breast, colorectal, Hodgkin lymphoma, ovarian, sarcoma, and melanoma
Keim-Malpass and Steeves51,b	Late adolescent and young adult women	Analyze cancer patients' online illness blogs.	Themes were living in the middle between sick and healthy/living and dead, establishing a new normal, urgency and transition into the unknown.
Keim-Malpass et al.49	Late adolescent and young adult women	Examine complementary therapy use described in online cancer blogs.	Themes were awakening (understanding the new normal), establishing new identities (that incorporated loss), the good stuff in life, and release.
Keim-Malpass et al.50	Late adolescent and young adult women	Analyze cancer survivor blogs.	Themes were pain and fatigue, insurance and financial barriers, infertility, and post-traumatic symptoms (including anxiety).
General: Kim and Gillham52,b	Women and men aged 20–29 y	Explore experiences of and better understand cancer patients through PlanetCancer.org blogs.	Themes were physical burdens, future prospects, isolation (physical and psychological), guilt, mortality, images of cancer, creating a positive attitude, health care, online social interaction, and cancer survivorship.
Microblog
Cervical and breast: Lyles et al.69	Unknown	Ascertain whether cervical cancer and BC screening is discussed on Twitter, whether sufficient volume exists for meaningful analysis, and nature of content.	Study coded 203 Pap test and 271 mammogram messages (> 75% from individual accounts); ∼25% of messages discussed personal experiences (attending appointments, negative sentiments about procedure, results); 8% of Pap test and 18% of mammogram messages promoted screening.
General: Sugawara et al.70	Japanese men and women	Examine recent Twitter usage to evaluate role in lives of today’s wired cancer patients.	Most tweets focused on greetings, treatment discussions, and other psychological support.
Social bookmarking sites
Prostate: Friedman et al.67,b	Older adult African American men	Examine content of articles related to PC screening information tagged on 4 major social bookmarking sites.	Most articles did not discuss screening guidelines for high-risk individuals; > 50% of the sites reviewed also encouraged individuals to seek more information about PC or cancer screening.
Social networking site
Breast: Bender et al.58	Unknown	Characterize purpose, use, and creators of Facebook groups related to BC.	Groups most often created to raise funds, raise cancer awareness, or promote external BC-related Web sites.
Multiple channels
Breast: Quinn et al.75	Unknown	Assess daily new BC-related posts on the Internet.	BC-related topics had significantly more posts/d than others. Most posts were on media sites. Accuracy levels were high, but significantly lower in posts on blogs or discussion forums. Anonymous posts were common and less likely to be accurate. Use of discussion forms increased over past 5 y.
Colorectal and breast: De la Torre-Díez et al.71	Unknown	Characterize purpose and use of Facebook and Twitter groups concerning CRC, BC, and diabetes.	Search found 216 BC groups, 171 CRC groups, and 527 diabetes groups. CRC and BC groups mostly focused on prevention, diabetes groups on research.
Prostate: Sillence and Mo76	Young adult and midlife men	Explore types of decision-making processes users exposed to on PC online message boards.	Many messages (49.6%) reported nonsystematic decision processes, most often deferral to doctor and proof of cancer removal. For systematic processing (36.5%), messages mostly considered treatment outcomes and side effects. Processes did not vary between blogs and online forums.

Note. AI = aromatase inhibitor; BC = breast cancer; CRC = colorectal cancer; FORCE = Facing our Risk of Cancer Empowered; HPV = human papilloma virus; PAP = Papanicolau; PC = prostate cancer; PSA = prostate-specific antigen.

^aArticle references Internet access digital divide.

^bArticle references cancer health disparities.

TABLE 2—

Social Media Cancer Studies With Descriptive Designs and Various Foci

Measurement	Type of Cancer (Study)	Participants	Social Media Platforms/Study Goals	Key Findings
Online survey	General (Seckin40)	Adult cancer survivors	Examine whether online self-help group interactions encourage patients’ sense of control over their illness and lead to higher satisfaction with health status.	Online support increased perception of control over cancer, which led to higher levels of satisfaction with health status. Individuals with less education and lower income levels reported more benefits from electronic coping support.
	Breast (Kim et al.28,a)	Medically underserved African American and White women with BC	Use online support groups (CHESS) to examine interplay between emotional support giving and receiving and the effects of emotional support giving and receiving on patients’ psychosocial health outcomes.	Those receiving high social support were less likely to have BC concerns and more likely to reframe their problems positively as well as to adopt positive strategies for coping.
	Breast (Tercyak et al.43)	Adult partners (male and female) of women at risk for hereditary BC	Understand the role of a Web-based message board and chat in meeting psychoeducational needs among partners of hereditary BC survivors.	Partners had a moderately high need for psychoeducational interventions across content areas. Partners tended to prefer printed self-help and interactive online resources.
	Skin (Hawkins et al.23)	Adult patients and parents of children with xeroderma pigmentosum	Use online support group to recruit participants to complete a survey on symptoms and QOL.	All participants reported positive effects on lifestyles and QOL from participation in online support group.
	Colorectal (Hwang et al.24)	Older adult cancer survivors	Assess the acceptability of narratives to promote CRC screening among members of an online weight loss community (sparkpeople.com).	Among members who were up to date with CRC screening, 39% were interested in sharing narratives. This positively correlated with African American race and negatively correlated with age, marital status, and CRC worry.
	Testicular (Bender et al.11,b)	Canadian young adult male cancer survivors	Examine met or unmet support care needs of survivors and their use of Internet, social media, and online communities.	Participants had more met than unmet needs; 62.5% had ≥ 1 unmet need. Most respondents (76%) used social media and 26% had used an online support community. Social media users were more likely to speak English as a first language and to have unmet needs.
Survey and content analysis	General (Fernsler and Manchester9)	Adult cancer survivors	Use forum to post cancer-related information.	Respondents used the cancer forum to connect with others sharing similar cancer experiences, to seek information, to encourage other survivors, and to seek emotional support
	Breast (Weinberg et al.8)	Adult BC patients	Investigate the use of online message boards that emotional and informational support.	Participants logged in to the BC bulletin board 158 times. Online approach was similar to face-to-face support groups. Participants discussed medical condition/treatment, shared personal experiences and concerns, and offered support for other members.
Survey and in-depth interviews	Breast (Bender et al.13)	Canadian adult BC survivors	Explore the role of online communities from the perspective of survivors who are face-to-face support group facilitators.	Online communities were used by 31.5% of participants, mostly during cancer treatment, for information, symptom management, and an emotional outlet to address unmet needs during periods of peak stress or uncertainty.
Survey and Web metrics	Breast (Han et al.22,a,c)	Medically underserved African American and White women with BC	Explore the social and psychological characteristics that predict different levels of engagement on an online BC support group message board (CHESS).	Many women do not use online support groups. Of participants, 44.6% posted messages, 32% lurked on the sites, and 23.4% did not log in to the system. Women who posted were more often were younger and lived alone. Nonusers and lurkers had significantly more social support than those who posted online.
Social network analysis	General (Durant et al.20)	Patients, caregivers, researchers, and providers	Determine whether gender-preferred social styles can be observed in user interactions in an online cancer community.	Men preferred to socialize in large, less intimate groups. Women were more likely to form a highly intimate bond on a women-specific cancer forum.
Focus groups and content analysis	Breast (Lieberman and Goldstein30)	Adult BC patients	Examine emotions expressed by women participating in online bulletin boards.	Expression of negative emotion did not predict psychosocial improvement. Expression of fear and anxiety was associated with lower QOL and with depression.
In-depth interviews and content analysis	Breast and prostate (Seale39)	Adult cancer survivors	Compare interviews and Internet postings involving people with BC and PC and discussion of sexual health.	Patients who disclosed information on the Web forum provided more detailed health and sexual health information online than during in-depth interviews.
Blogs
Interviews	General (Andersson et al.47)	Swedish adult relatives of cancer patients	Highlight relatives’ experiences of blogging when a family member is in the end-of-life phase of cancer.	Themes were delivering information about cancer, vulnerability in writing the blog, sensitivity of information, not knowing who viewed the blog, blogging to ease burden of cancer, and blogging to build strength.
Focus groups	General (Chiu and Hsieh48)	Adult cancer patients and caregivers	Explore how cancer patients’ posting on and reading blogs, social networking sites, and online support groups play a role in their illness experience.	Themes were wanting to be remembered after death, writing for relief and gaining strength, writing to support fellow patients, influence of other cancer patients, finding other survivors, and a sense of belonging.
Survey	General (Chung and Kim46)	Adult cancer patients and their companions	Examine use of blogs related to gratification outcomes.	Four areas emerged as gratifications of blog use: prevention and care, problem solving, emotion management, and information sharing. Participants found blogging most helpful for emotion management and information sharing.
Microblog
Social network analysis	Breast and prostate (Himelboim and Han68)	Unknown	Assess density and interconnectedness of social networks of individuals who post information on Twitter about BC and PC.	Authors categorized groups who frequently posted about BC and PC as core communities and groups who more sporadically discussed these cancers as visiting communities. The different types of communities exhibited distinct patterns of information flow, information sources, and user interactions. Regardless of group classification, communities that discussed BC were significantly larger than PC communities. Visiting communities were not common in BC clusters but played an integral role in PC clusters.
Social networking sites
Survey	Cervical (Zaid et al.65,b)	Women with history of small- or large-cell cervical cancer	Determine feasibility of posting a survey on Facebook for epidemiological and QOL research.	Use of social media for epidemiological and QOL research with patients with rare gynecologic tumors appeared feasible. Social media permitted research to be conducted efficiently and rapidly.
	Cervical (Gunasekaran et al.59)	Young adult women in Australia	Assess and investigate predictors of cervical cancer and HPV knowledge among participants recruited on Facebook	Most respondents (> 50%) knew that HPV is a common virus. Predictors of high HPV knowledge were older age, higher SES, Australian birth, older at sexual debut, awareness of HPV vaccines, self-reported HPV vaccination status.
	Breast (Kapp et al.60)	Women aged 35–49 y	Assess feasibility of recruiting cancer patients via Facebook.	Few women (3.2%) clicked on ad and proceeded to survey; none completed survey.
	Colon and skin (Leighton et al.61,b)	Young adults and adult genetic counselors	Recruit participants via Facebook to participate in study of perceptions and understanding of direct-to-consumer genetic testing.	General public and genetic counselors differed significantly in interpretation of genetic test results in 3 of 4 chronic disease mock scenarios.
Multiple channels
Web metrics	General (Justice-Gardiner et al.72,a,c)	Racially and ethnically diverse populations	Use blogs, microblogs, and social networking sites to increase awareness of and access to existing Spanish-language cancer resources for cancer survivors.	Web site traffic increased in all areas from 2008 print and radio campaign to 2009 campaign with the addition of social media.
	Osteosarcoma (Lam et al.73,b)	Unknown	Assess quality, readability, and social reach of blogs, YouTube, Facebook, and social bookmarking sites used to post information on pediatric osteosarcoma.	Only 12.5% of social media included adolescent-specific content. Only 1 Web site achieved readability targets and provided adolescent-specific content.
Survey	Breast (Paxton et al.74,a,c)	Adult African American BC survivors	Use social networking (blog, Facebook, Twitter) recruitment to determine physical activity intervention preferences and whether preferences differ by medical or demographic factors.	Participants were interested in interventions that used the Web, e-mail, or postal mail to promote walking and resistance activities. Intervention preferences did not differ by medical or demographic factors.
Survey	General (Song et al.63)	Racial and ethnic minority young adult cancer survivors	Explore cancer survivor identities and how these are associated with internalized stereotypes about survivorship and psychological health on a social networking site and video blog for cancer survivors (LIFECommunity).	Survivors of childhood cancer did not publicly dwell on cancer stereotypes when discussing their experiences with each other. Survey found evidence for internalizing positive and negative stereotypes. Self-perceived negative stereotypes were positively associated with depression.
Frequencies of study enrollment	General (Rabin et al.62)	Young adult cancer survivors	Assess various strategies to recruit participants for for pilot intervention, including use of Facebook and Twitter.	Most participants were recruited from mailings. Internet and social media recruitment accounted for 15% of study contacts, 7% of those screened, and no enrollees.

Note. BC = breast cancer; CRC = colorectal cancer; HPV = human papilloma virus; PC = prostate cancer; QOL = quality of life; SES = socioeconomic status.

^aArticle references Internet access digital divide.

^bArticle references health literacy.

^cArticle references cancer health disparities.

Seven articles (10.1%) described the use of social media interventions for affecting health behaviors and social support (Table 3),^{18,19,29,36,57,64,66} of which 4 had an experimental design^19,29,36,64 and 3 were feasibility studies.^18,57,66 Studies with experimental designs sought to lessen depression,^19,29 increase social support,⁵⁷ and increase the quality of life of cancer survivors.³⁶ The other 3 studies that contained social media interventions were designed to (1) increase support for women who were sexually distressed subsequent to gynecologic cancer treatment,¹⁸ (2) use Facebook to increase social support for physical activity among cancer survivors,⁶⁴ and (3) use a virtual world as a mock setting for training physicians to deliver medical bad news, such as a cancer diagnosis.⁶⁶

TABLE 3—

Social Media Cancer Intervention Studies

Channel/Type of Cancer	Study	Participants	Social Media Platforms/Study Goals	Key Findings
Virtual world: breast	Andrade et al.66	Male, 62%; female, 38%; mean age, 29 y; White, 50%; Hispanic, 38%; Asian, 12%	Study feasibility of creating standardized patient avatars in Second Life to use in training physicians to deliver bad news.	Self-efficacy affective competence scores of trainees improved overall. All participants considered the experience positive and novel, although avatars were not able to portray body language cues.
Online forum
Gynecologic	Classen et al.18	Women; age range, 28–59 y; White, 74%; Asian, 11%; African American, 8%; Hispanic, 7%	Feasibility study to examine participation rates and preliminary outcomes for an online support group (GyneGals) designed specifically for women who are sexually distressed subsequent to gynecologic cancer treatment.	Participation was higher in wait-listed control group than in immediate treatment group. Exit interviews indicated that most participants were satisfied with intervention. Intent-to-treat analyses suggested a small effect for reduction in sexual distress.
General	Duffecy et al.19	Female, 87%; male, 13%; mean age, 50 y; White, 87%; African American, 10%; Asian, 1%; Hispanic, 10%	Examine feasibility and acceptability of individual Internet intervention for post-treatment cancer survivors embedded and integrated into an Internet support group (Onward) with ultimate goal of enhancing adherence and learning. Compare with individual Internet intervention alone.	Mean number of logins over 8 wk: Onward group, 20.8; individual intervention–only group, 12.5. Onward group had fewer dropouts than individual intervention–only group (3 vs 5). Among participants with depressive symptoms at baseline (n = 17 for both groups), hospital anxiety and depression decreased significantly.
Breast	Klemm29	Women; mean age, 52.3 y; White, 90%; African American, 10%	Evaluate effects of different formats of online group support (moderated vs peer led) on depressive symptoms and extent of participation among women with breast cancer.	Depressive symptoms did not differ significantly by group or over time or demonstrate a group-by-time interaction. Moderated groups read and posted significantly more messages than peer-led groups.
Prostate	Osei et al.36	Men; mean age, 67.2 y; White, 85%; African American, 7.5%; Hispanic, 5.0%	Explore effect of an online support group (Us TOO) intervention on self-reported QOL.	From baseline to 6 weeks, intervention group improved in 3 of 6 QOL measures (urinary, sexual, and hormonal health); 6 QOL measures declined in control group. Both groups returned to baseline levels at 8 wk.
Multiple channels (social networking and video sharing): general	McLaughlin et al.57,a	Male, 64.3%; female, 35.7%; Hispanic, 85.7%; Asian, 7.1%; Native American, 7.1%	Examine key factors that may influence a childhood cancer survivor’s participation in a social networking and video-sharing intervention (LIFEcommunity) tailored to their social support, social capital, and QOL needs.	Bridging social capital of participants was positively related to level of participation, sharing videos, more frequent visits to other participants’ walls, exchanging information and empathy support through sending and receiving private messages. Decreased levels of real-life support were linked to increased levels of online participation. Participants with more family support were less likely to participate in the social networking program. A higher depression level correlated with more activity in blog posting.
Social networking site: general	Valle et al.64,b,c	Female, 91%; Male, 9%; mean age, 32 y; White, 91%; African American, 9%;	Evaluate feasibility and efficacy of Facebook-based intervention to increase physical activity; compare with a Facebook-based self-help comparison condition.	Intervention participants lost weight and engaged in more weekly minutes of light physical activity than did control group. Groups did not significantly differ in weekly minutes of moderate vigorous physical activity.

Note. QOL = quality of life.

^aArticle references health literacy.

^bArticle references cancer health disparities.

^cArticle references Internet access digital divide.

Social Media and Disparities

A secondary goal of our review was to extract data about how authors described health disparities in their work. We examined how authors described health disparities among diverse populations and the role of the digital divide and limited general and health literacy, which may contribute to disparities in health communication. Of the 69 articles, 7 (10.1%) touched on cancer disparities,^{22,51,52,64,67,72,74} and these articles often discussed populations that were underserved or overlooked because of age, geography, race, or ethnicity. Three studies concentrated on young adult cancer survivors, an understudied population, who often lack services and resources specific to their needs.^51,52,64 Three studies focused on the use of social media to address African Americans' increased risk for specific cancers (prostate cancer and breast cancer).^22,67,74 One study described health and communication-related disparities, in particular how articles about prostate cancer screening that were posted on social bookmarking sites did not adequately discuss screening guidelines for African American men, a population at increased risk for prostate cancer.⁶⁷

Three studies (4.3%) discussed the digital divide—individuals’ lack of access to Internet communication—in relation to race and ethnicity.^22,28,72 One study highlighted the growing use of the Internet among US Hispanics, thus shrinking the digital divide; presenting information online may be a viable new method to promote health information access for Hispanics.⁷² Two other studies directly aimed to reduce the digital divide via social media (e.g., providing Internet-based health services) for breast cancer survivors living in rural, medically underserved areas.^22,28

Five studies (7.2%) mentioned health literacy, and in some cases site readability, in the context of using social media for cancer care.^{11,57,61,65,73} One article described the need for a high literacy level to interpret the cancer genetic testing offered by some Web sites.⁶¹ Another study reported that insufficient e-health literacy deterred participation in an online community.¹¹ Participants in 1 study were required to have a minimum of an eighth-grade reading level to participate in a Web-based intervention aimed at increasing support for cancer survivors.⁵⁷ Another study posited that participants in online support groups may already have higher health literacy levels than nonusers of such online communities.⁶⁵ In 1 study, readability was an outcome in a study assessing osteosarcoma information on the Internet and social media channels. Overall, the researchers rated the information on osteosarcoma-related Web sites as fair, but most Web sites did not reach the recommended standards on ease of understanding.⁷³

DISCUSSION

Our review was the first, to our knowledge, to assess the taxonomy and characterize the use of social media across the cancer spectrum. The earliest research in this field (1996–2007) comprised descriptive studies analyzing the use and content of cancer-related online communities such as message boards, discussion threads, and support groups. Social media intervention studies were not reported until 2010. Our findings suggest that the use of social media has evolved from simple information delivery and access via online forums to intervention studies with increased use of interactive strategies. Most social media research in our review analyzed the content of online forums where users asynchronously posted or responded to information with resource and information sharing or emotional support.

Our findings highlight the possible lag between what has been reported in peer-reviewed literature about how individuals use different types of social media sites to find cancer information and support and the public’s current use of these sites for similar information and reasons. For example, social networking sites (e.g., Facebook) are among the most popular forms of social media, but they received limited attention in social media research related to cancer. However, this finding may also indicate that the public prefers to search directly for cancer-related information or to seek support on online message boards rather than on social networking sites. This preference for online message boards may be attributable to users’ preference to remain anonymous and “faceless,” as opposed to posting personal information, pictures, and other identifying information (as happens on social networking sites) when searching for cancer information. Therefore, interventions that use social media as a health communication channel possibly should focus on cancer survivors and caregivers and host online support groups or forums that allow users to remain anonymous if they wish.

Our analysis also showed that previous research focused primarily on describing social media cancer content but rarely evaluated the accuracy or usefulness of such information. Only 1 study critiqued the accuracy of cancer-related information posted on a social media site, and it found that the majority of the site’s videos contained suboptimal information. Inaccurate medical information on the Internet may lead to users misdiagnosing themselves, altering prescription medication regimens, failing to follow physician recommendations, and sharing incorrect medical information.⁷⁷ Therefore, future research should assess the accuracy of cancer information shared on social media sites.

We also assessed how researchers use social media to address cancer-related disparities and to reduce the digital divide. Although 7 studies discussed health disparities, none focused specifically on reducing cancer disparities via social media. Chou et al. found that significantly more minority than White Americans used social networking sites (45% vs 31%), suggesting that social media can provide researchers with a unique opportunity to provide targeted health information to minority populations.⁷⁸ One study focused specifically on reducing cancer-related disparities (for young cancer survivors) via social media interventions.⁶⁴ Thus, although we observed exponential growth in the use of social media in cancer care, we found a deficit of published descriptive and intervention studies addressing cancer health disparities. Furthermore, the lack of robust descriptions of demographic and other characteristics of participants in the majority of these studies precluded thoughtful assessment of the digital divide and its impact on cancer health disparities.

Although studies have predicted that social media may be used to bridge the digital divide,^5,79,80 at this time it is difficult to fully discern what impact social media will have on reaching underserved groups and reducing the digital divide. Some researchers have noted the persistent gap in broadband Internet access among populations of lower socioeconomic status, but Gibbons et al. believe that social media may be the channels through which the digital divide will be reduced. They assert, however, that more emphasis on remedying health care system, provider, and individual (e.g., health literacy) factors is needed to reduce health disparities.³ Scant attention focused on how social media could be used to address health literacy. In light of the plethora of national public health imperatives for delivering clear and acceptable information, future social media interventions that address health literacy may represent a fertile area of research.

Our systematic review revealed a paucity of literature examining intervention applications for social media in cancer care, showing that this area of cancer care research is in its infancy. Taken together with Chou et al.’s work, our findings suggest the need to determine, through empirical intervention research, how these new, widely used communication channels can be exploited to deliver cancer information and influence behaviors. Future research could evaluate the use of social media sites in interventions to enhance behaviors associated with cancer screening, treatment, and follow-up care. Because social media sites have been shown to increase the dissemination of health information and provide emotional support, future interventions could test the efficacy of social media to enhance emotional and informational support of cancer survivors. Interventions that use popular social networking sites, such as Facebook, should also be evaluated to determine their potential for promoting cancer screening. Adults aged 18 to 49 years make up the largest group of users of social networking sites such as Facebook and Twitter,⁸¹ so these sites might be useful in reaching this age group to promote cancer prevention and control behaviors and to disseminate medically sound health information regarding cancer, with a particular focus on skin and cervical cancer screening, HPV vaccination, smoking cessation, sun exposure reduction, weight management, and the benefits of exercise.

Limitations

We excluded articles not published in English. However, our search turned up only 2 articles that were excluded for this reason, likely because our search terms were biased toward studies reported in English. In addition, we did not include scientific contributions of gray literature, such as dissertations and reports published by nonprofit organizations, in our systematic review. However, we included contributions of gray literature in our interpretation of study findings.

We did not perform a meta-analysis because of the limited number of quantitative intervention studies and the variations in study design, population, intervention types, exposures, and outcomes. Such an analysis is an important focus of our continuing efforts.

Conclusions

Our systematic review highlighted the evolution in the use of social media in cancer care from online forums and message boards where the public shares cancer-related information and emotional support to multiple sites that provide cancer information and support. Researchers may need to examine the use of other social media sites (such as social networking sites) in light of the increased adoption of these channels by the public and the health industry. Intervention research in this area is in its beginning stages.

The paucity of intervention studies suggests an important and emerging area of research for improving behaviors and promoting well-being and quality of life (e.g., increasing cancer screening, motivating youth to use sunscreen, providing support during chemotherapy, reducing fatigue). Research is also needed to explore and evaluate how social media sites can make health information more accessible, engaging, and interactive for increasingly diverse audiences.