Vlogging Cancer: Predictors of Social Support in YouTube Cancer Vlogs

Abstract

Social support research has demonstrated the physical and psychological benefits of social support for patients, but has done little to identify successful strategies for eliciting social support. However, contemporary digital channels offer new ways to examine this issue. In particular, this study explores vlogging as a coping strategy for cancer patients and a context in which to explore predictors of online social support. A content analysis of 69 YouTube vlogs and 869 associated comments was performed. A series of multilevel binomial logistic regression analyses revealed that narrative features that position the cancer patient as protagonist—including providing an explanation of the diagnosis experience, agentive problem solving, and positive reappraisal of the situation—were associated with receiving empathic support. In contrast, moralizing pleads for audience checkups decreased the likelihood of receiving empathic support. Findings contribute to an undertheorized body of research that also has translational value for patients, doctors, and designers of supportive online spaces who might eventually recommend narrative vlogging in clinical settings.

Introduction

The effects of social support on the psychological and emotional recovery of cancer patients have been well documented by researchers in the social sciences and medical fields.^1–8 Scholars have debated the merits of different coping strategies (support seeking, confronting the problem, secluding oneself, etc.), but largely agree that social support is an effective resource for patients undergoing cancer treatment procedures. Indeed, some researchers argue that social support may be one of the most significant coping resources for inducing psychological recovery,^9,10 and in many cases it is associated with better physical health.^11–13 Although the benefits of support seeking are well documented, there is substantially less work on specific strategies that elicit support. This may be partly because of the difficulty of measuring the relationship between support seeking and support giving, as retrospective surveys are prone to recall biases, especially in highly emotional contexts,¹⁴ and recording in-laboratory interactions can be time consuming and artificial. However, many contemporary online self-presentations include a natural delay between posts and comments that provide a context for identifying behaviors and communication practices that effectively elicit support.

Based on findings from a content analysis of YouTube cancer vlogs and comments, we identify techniques for eliciting support online by presenting oneself as a protagonist in a narrative. For patients coping with a potentially life-threatening illness, sharing personal narratives through vlogs may be a beneficial coping strategy,¹⁵ helping patients regain agency over their experience and encouraging the development of social connections with the narrative audience.^16–19 Based on previous narrative research, we expect that vloggers who position themselves as a protagonist by providing narrative exposition, demonstrating positive growth, and problem solving will increase their likelihood of receiving supportive responses. We test this hypothesis with the aim of identifying communication strategies that could be recommended by healthcare providers to enhance patient support availability.

The importance of support

The cancer community is one of the most supported and supportive health communities,² which is important because of the psychological and physical benefits of social support. Receipt of social support is a significant factor in facilitating successful coping processes.²⁰ In addition, social support seems to mediate stress and well-being,³ helps stressed individuals overcome challenges,²¹ and may improve physical health.^11–13 However, the availability of social support is influenced by a number of factors, including the coping strategy employed by the patient.^22–24 Research finds that digital communication channels allow cancer patients to develop and utilize adaptive coping mechanisms, including support seeking.^6,25,26 One of the primary benefits of digital interventions is the access and availability they provide,^27,28 especially through access to weak-tie relationships.^28,29 For example, professional online interventions encourage support between cancer patients, which have resulted in positive health outcomes.^6,28,29 More generally, Internet use may help prevent a long-term decrease in social resources as patients become more isolated.^10,26 Connecting with others through vlogging may be an accessible strategy for finding support while providing patients the opportunity to share their personal cancer narrative.

The power of narrative

A growing body of literature has focused on the effectiveness of narrative construction as a coping strategy for the chronically ill, although a larger body of narrative literature exists beyond illness research.^30–32 Narrative creation helps struggling individuals identify potential causality and linearity, and develop an understanding of a difficult situation.^30,31 Narrative creation also helps cancer patients regain a sense of agency or control of their experience by providing an outlet to examine and reimagine their experiences.^16,18,19 Finally, narrative construction encourages a sense of self-reliance,¹⁹ allowing patients to negotiate agency within the medical system that positions doctors as experts and cancer patients as merely recipients of treatment and information.¹⁶ In sum, patients may benefit from narrative creation by repositioning themselves as the protagonist of their cancer story—gaining agency and control over their situation and improving perceptions of self-reliance.

In addition to the benefits of narrative generation, sharing one's cancer narrative online may engender useful support by fostering interpersonal connections,^33,34 facilitating psychological adjustment,^9,10 and promoting recovery.^11–13 By sharing a cancer narrative through sites such as YouTube, a large support network becomes available to the patient. This expanded network may fulfill a listening or comforting role for cancer patients that family members and medical professionals alone cannot fulfill, which is an important part of the narrative process.¹⁷ We hypothesize that cancer patients who create and share an agentive narrative account of their experience may thus benefit from increased social support availability.

Vlog narrative structure

When creating and sharing a cancer narrative, the inclusion of certain narrative components may assist in forming an intimate connection with the audience (and eliciting support).¹⁸ From previous research, we have identified four narrative components specifically relevant to cancer narrative that position the patient as the protagonist. First, conveying one's personal history, particularly the moment of diagnosis, may make the vlogger relatable,¹⁸ which can increase audience intimacy and encourage support.³⁵ Second, pleading for the audience to get a checkup may similarly increase the vlogger's relatability,³⁶ and also positions them as a moralizing force. Third, problem-solving behaviors are likely to elicit support and are consistent with the protagonist behavior.³ Fourth, positive reappraisal of the situation that reflects emotional growth may additionally elicit support.³ The inclusion of these protagonist-signaling narrative features is thus predicted to promote social support elicitation.

In addition, these narrative components may operate differently within different vlog formats. Liu et al.³⁶ outlined four distinct vlog categories employed by vloggers with chronic illness. Teaching vlogs utilize personal experience to educate viewers; in personal journal vlogs, patients provide day-to-day updates on their current status in a diary format; in self-documentaries, vloggers capture situational cancer-related experiences as they occur; and in video compilations, vloggers compile images and video from multiple sources. Although Liu et al.³⁶ do not measure differences in received support between vlog formats, the authors note that personal journal and self-documentary vlogs facilitate intimate connections between the narrator and audience. As a result, vlog categories may differently affect the likelihood of supportive responses. Thus, the following are proposed:

H1: The inclusion of (a) the diagnosis experience, (b) pleads for audience checkup, (c) problem solving, and (d) positive reappraisal in vlogs will increase the likelihood of supportive comments compared with vlogs without these narrative components.

RQ1: Will vlog format affect the likelihood of receiving supportive comments?

Methods

Sample

A content analysis of 69 YouTube vlogs and 869 comments was performed. A random sample was created by collecting every third video result after searching “cancer vlog” and “cancer blog” on YouTube in August 2015, until 35 vlogs were collected per phrase. One vlog per vlogger was included and vlogs with fewer than two comments were not collected (eliminating automatically generated comments for sharing to Google+). Although this sample is not exhaustive, the collection of 35 vlogs per search term (sorted by relevance) seemed to capture a comprehensive sample, as irrelevant material (e.g., fitness videos and marketing videos) became increasingly difficult to eliminate after the 35th vlog. One vlog was eliminated as a vlogger had created multiple accounts. Comments were copied into a word document for analysis. To avoid biasing data toward vlogs that received a large number of comments, every third comment was collected for vlogs with >30 comments.

Coding procedure

Each vlog and comment was coded in full, as our two different units of analysis. All variables were tested for intercoder reliability according to Krippendorff's alpha for two coders using approximately 11 percent of the sample (92 comments). Upon achieving reliability, each coder independently analyzed half of the sample. Problem cases were collectively discussed and coded.

Measures

Control categories

Demographic control variables included vlogger gender, vlogger age, cancer subject, and cancer type, as each of these variables has been shown to be associated with receipt of social support in the literature.^{2,3,7,21,37–40} Vlogger gender (α = 1) was visually coded as male, female, or unidentifiable. Vlogger age (α = 0.765) was grouped into five categories: 0–10, 10–20, 20–40, 40–60, and 60+ years. Cancer subject (α = 1) was coded as either self, spouse/partner, or family member. Spouse/partner and family member were later collapsed into a single response because of few instances of each, creating a dichotomous category of self/other. Finally, cancer type (α = 1) was assessed according to the video title and vlog content.

Additional control variables included support network and vlog number. Support network (α = 0.843) was coded as a proxy for pre-existing support network, as offline support could not be measured precisely through YouTube. This was coded as negative support (e.g., negatively affected by family and friends), no support, positive support, or very positive support according to verbal information provided by the vlogger. Finally, vlog number (α = 1) was coded according to the number of vlogs previously published by the vlogger to account for possible changes in support as people become practiced at vlogging.

Narrative categories

Five narrative components were coded for each vlog. Vlog category (α = 0.832), which included teaching, personal journal, self-documentary, and video compilation,³⁶ was coded according to vlog content. Diagnosis (α = 1), problem solving (α = 0.861), positive reappraisal (α = 1), and pleads for audience checkup (α = 1) were all coded as dichotomous yes/no categories according to vlog content. Diagnosis was coded in instances where the diagnosis procedure was outlined by the vlogger.¹⁸ Problem solving was coded when vloggers described creating a plan to proactively deal with cancer or where they indicated a strategy to “make things work.”³ Positive reappraisal was coded when vloggers suggested some personal growth or a positive emotional or social change because of their cancer experience.³ Pleads for audience checkup were coded in any instance where the vlogger advised the audience to seek medical advice with reference to cancer.³⁶

Support categories

For comments, each support type outlined by Neuling and Winefield⁹ was coded using a dichotomous yes/no response. Empathic support (α = 0.777) was coded when commenters acknowledged the vlogger's emotions or feelings (e.g., “I see how concerned you are”), or in cases where the commenter encouraged the vloggers to continue discussing their experience (e.g., “Please keep us updated”). Reassuring support (α = 0.761) was coded when commenters expressed hope, complement, or provided an uplifting message (e.g., “You can do this!”). Informational support (α = 1) was coded whenever commenters provided factual information within the comment text (e.g., “Vitamin B6 might help”). Finally, tangible support (α = 1) was coded when the commenter expressed interest in assisting the vlogger financially or in another tangible way (e.g., “Can I donate money?”). Support types were not mutually exclusive.

Results

Descriptives

The majority of comments in our data set included some supportive element (85.3 percent). Reassuring and empathic support were provided most frequently, whereas informational and tangible support occurred infrequently (Table 1). Using the categorical framework identified by Liu et al.,³⁶ vloggers exhibited a clear preference for the personal journal category (Table 2). Documentary and teaching vlog formats occurred with similar frequency, whereas the compilation vlog type was rarely used. This suggests a preference for vlog formats that encourage intimate connections between the vlogger and audience.³⁶ Descriptive information for control categories is outlined in Table 3.

Table 1.

Support Type Distribution Across Comments

Support type	Number of instances (N = 896)	Percentage
Reassuring	579	64.6
Empathic	320	35.7
Informational	132	14.7
Tangible	13	1.5
No support	132	14.7

Note: The cumulative percentage is >100 percent as some comments included multiple support types.

Table 2.

Vlog Category Distribution

Vlog category	Number of instances (N = 69)	Percentage
Personal journal	43	62.3
Documentary	13	18.8
Teaching	12	17.4
Compilation	1	1.4

Note: One vlog was eliminated from the sample because of a vlogger using multiple YouTube accounts, reducing the total number of vlogs to 69.

Table 3.

Control Category Distribution

Control category	Number of instances (N = 69)	Percentage
Vlogger gender
Male	24	34.7
Female	45	65.2
Vlogger age
10–20	9	13.0
20–40	49	71.0
40–60	9	13.0
>60	2	2.9
Vlog subject
Self	61	88.4
Other	8	11.6
Pre-existing support
Negative support	1	1.4
No support	20	30.0
Positive support	20	30.0
Very positive support	28	40.6
Cancer type
Blood	10	14.5
Lung	2	2.9
Breast	14	20.3
Brain	2	2.9
Cervical/ovarian	4	5.8
Bone	2	2.9
Colon	3	4.4
Testicular/prostate	4	5.8
Thyroid	6	8.7
Skin	2	2.9
Pancreatic	2	2.9
Other	6	8.7
Indeterminable	12	17.4
	M	SD
Vlog number	7.26	14.07

Note: One vlog was eliminated from the sample because of a vlogger using multiple YouTube accounts, reducing the total number of vlogs to 69. Vlogger gender, vlogger age, vlog subject, pre-existing support, and cancer type are categorical measures, whereas vlog number is continuous.

M = mean; SD = standard deviation.

Social support predictors

To identify the effect of narrative components on the likelihood of eliciting supportive comments, a series of multilevel binomial logistic regression models were constructed that regressed each support type on narrative and control variables. However, only models for reassuring and empathic support outcomes were successfully executed because of few instances of informational and tangible support. Thus, findings are reported only for reassuring and empathic support models (Table 4).

Table 4.

Multilevel Binomial Logistic Regression Models Regressing Social Support Type on Narrative and Control Variables

	Reassuring support				Empathic support
	Model 1		Model 2		Model 3		Model 4
	β	SE	β	SE	β	SE	β	SE
Vlog category
Personal journal			0.87^*	0.42			0.52^*	0.20
Self-documentary			0.64	0.55			1.78^***	0.28
Compilation			2.03^†	1.06			1.85^**	0.65
Diagnosis			0.61	0.45			1.33^***	0.24
Problem solving			0.13	0.39			0.69^**	0.27
Positive reappraisal			0.34	0.36			0.71^***	0.20
Audience checkup			−1.10	0.68			−1.31^***	0.40
Vlogger gender	0.45	0.33	0.45	0.33	−1.24^***	0.26	−1.48^***	0.29
Vlogger age
10–20	−1.70^*	0.73	−2.21^**	0.70	−1.27^**	0.49	−3.09^***	0.55
20–40	−1.95^***	0.56	−2.35^***	0.62	−0.74	0.41	−2.46^***	0.49
40–60	−3.16^***	0.65	−3.12^***	0.67	−0.93^*	0.48	−2.26^***	0.52
Vlog subject	−0.47	0.40	−0.30	0.40	−0.32	0.32	−0.95^***	0.25
Pre-existing support
Negative support	−2.45^***	0.38	−3.18^***	0.46	0.37	0.34	−0.49	0.30
No support	−0.16	0.36	−0.26	0.37	0.43	0.26	0.50^*	0.20
Positive support	−0.25	0.45	−0.29	0.45	0.25	0.40	−0.02	0.28
Vlog number	0.000	0.006	0.003	0.005	0.003	0.004	0.01^***	0.002
Cancer type
Blood	−1.32	1.66	−1.19	1.51	0.25	0.68	0.46	0.54
Lung	−1.70	1.64	−1.08	1.48	0.52	0.84	0.64	0.48
Breast	−0.02	1.55	0.75	1.48	0.36	0.64	1.27^**	0.48
Brain	−1.96	1.56	−1.85	1.43	0.73	0.51	1.44^***	0.34
Cervical/ovarian	−1.25	1.59	−0.59	1.51	1.22^*	0.56	1.98^***	0.43
Bone	−2.03	1.70	−1.69	1.54	1.09	0.89	1.99^***	0.60
Colon	−1.51	1.60	−0.95	1.59	0.86	0.63	2.47^***	0.62
Testicular/prostate	−1.98	1.74	−1.76	1.67	1.21	0.69	2.5^***	0.55
Thyroid	−1.35	1.62	−0.44	1.55	1.19	0.71	2.69^***	0.57
Skin	−1.31	1.55	0.88	1.84	0.78	0.51	3.31^***	0.63
Other	−1.14	1.47	−0.88	1.35	1.05^*	0.43	0.58^*	0.29
Indeterminable	−1.19	1.51	−0.45	1.49	0.65	0.55	1.62^***	0.43

Note: Baseline categories included teaching for vlog category, male for vlogger gender, >60 for age, other subject for vlog subject, very positive support for pre-existing support, and pancreatic for cancer type. Coefficients for narrative components (diagnosis, problem solving, positive reappraisal, and audience checkup) represent the change in log-odds when the component is included. Statistical significance is reported according to Wald chi-square.

p < 0.05; ^**p < 0.01; ^***p < 0.001; ^†p < 0.06.

SE = standard error.

Control models (models 1 and 3) yielded several significant results. Female vloggers were more likely to elicit empathic support than male vloggers, and vloggers aged >60 years were most likely to receive reassuring and empathic support than younger vloggers. Not surprisingly, references to pre-existing support also predicted reassuring support in comments, with very positive pre-existing support most likely to elicit this support type. Finally, two cancer groupings—cervical/ovarian and other cancers—were more likely to receive empathic support than the baseline (pancreatic).

In testing H1, the addition of narrative variables affected findings for empathic support, with every predictor reporting significant findings (model 4). As predicted, reference to diagnosis increased the likelihood of empathic support in a given comment by 278 percent (χ² = 31.09; p < 0.001), problem solving increased the likelihood by 101 percent (χ² = 6.78; p < 0.01), and positive reappraisal increased the likelihood by 103 percent (χ² = 12.7; p < 0.001). In contrast, the inclusion of pleads for audience checkup decreased the likelihood of empathic support by 73 percent (χ² = 10.49; p < 0.001). None of the hypothesized narrative features significantly predicted reassuring support (model 2). Thus, H1 is mostly supported by the empathic support model but not supported by the reassuring support model, which is discussed hereunder.

In answering our research question, the compilation format predicted the greatest likelihood of empathic support. In comparison with teaching vlogs (the baseline category for both models), comments were 185 percent more likely to include empathic support for the compilation format (χ² = 7.98; p < 0.01), 178 percent more likely for self-documentary vlogs (χ² = 39.32; p < 0.001), and 52 percent more likely for personal journal vlogs (χ² = 6.5; p < 0.02). Although the compilation format scored highest in support likelihood, a single instance of this format limits our ability to infer substantively from this result. However, the increased likelihood of social support within comments submitted to personal journal and self-documentary vlogs suggests heightened provision of empathic support in response to these intimate vlog formats.³⁶

Finally, for the reassuring support model, vlog category also emerged as a significant predictor. Comments generated in response to personal journal vlogs were 87 percent more likely to contain reassuring support than those submitted to teaching vlogs (χ² = 4.26; p < 0.04), whereas those submitted in response to compilation vlogs were 203 percent more likely (χ² = 3.7; p < 0.06). However, the scarcity of compilation vlogs in this data set again limits our ability to infer substantively from this particular result. Taken together, these findings suggest that vlog category significantly predicted the likelihood of supportive responses, with intimate vlog formats (personal journal and self-documentary) reporting a greater likelihood of social support than the teaching format.

Discussion

Through a content analysis of cancer vlogs and comments on YouTube, a number of narrative features—diagnosis experience, problem solving, and positive reappraisal—were found to predict social support in comments. Cohering these findings, these features position the patient as the protagonist in their cancer narrative, potentially assisting the formation of an intimate connection with the audience and encouraging supportive response.¹⁸ Because support availability is correlated with positive health outcomes,^11–13 findings have implications for patients seeking support and healthcare providers encouraging successful coping. Furthermore, although this study identifies predictors of support for cancer vlogs, findings may provide insight into coping strategies for other illnesses.

As predicted, vloggers who incorporated stories about their diagnostic history provided evidence of agentive problem solving and suggested that positive reappraisal of their circumstances increased the likelihood of receiving supportive responses. In contrast, vloggers who suggested getting a checkup to the audience were less likely to receive support. In retrospect, although pleads for audience checkup could express concern or care, potentially strengthening the intimate connection with viewers,³⁶ this may instead be experienced as unsolicited “finger-wagging.” Similar effects have been found previously in response to unsolicited advice,^41,42 which may imply a receiver's incompetence.⁴²

In addition to narrative components, vlog format (as outlined by Liu et al.³⁶) also predicted social support in comments. The use of self-documentary increased empathic support compared with teaching vlogs, whereas personal journal increased both empathic and reassuring support. This finding is important as empathic support is considered one of the most valuable support types in interpersonal relationships,^{20,37,43–45} and is linked to improved physical health.^13,46 Although the compilation format elicited the greatest amount of empathic and reassuring support, a single occurrence of this category hinders substantive interpretation of this finding. The scarcity of compilation vlogs may also suggest a preference among cancer vloggers for the teaching, personal journal, and self-documentary formats.

Overall, our findings augment social support research by identifying specific strategies patients can use online to elicit support, which is widely asserted to be a positive agent in psychological and physical recovery.^9–13 These findings potentially have compelling translational value, as healthcare providers, family members, and patients themselves may want to explore techniques that optimize their receipt of support. Furthermore, our findings confirm and expand previous research on the effectiveness of sharing one's cancer narrative as a possible coping mechanism.^15,28

Limitations and future direction

Although our findings provide insight into techniques for eliciting support within YouTube, these findings should be cautiously extended to other channels. Audiences may respond differently within other sites, and findings may not replicate for offline narrative coping. Future research should identify commonalities and differences between narrative-sharing processes online and offline, and across digital channels. Another limitation of this study is the use of binary variables for narrative components in vlogs and social support in comments. Because of the use of dichotomous variables, relative strength of narrative components and social support cannot be ascertained. The use of continuous variables in future research will allow for greater measurement precision. Finally, we have argued that the relationship between narrative components and supportive responses may result from an intimate connection between vloggers and commenters.¹⁸ However, intimacy is difficult to measure through YouTube. Future research should seek to empirically test the link between narrative features and social support receipt through the mediation of intimacy, liking, or other factors stemming from narrative theory. An experimental intervention that encourages narrative coping and measures subsequent support would be especially useful for continued theoretical development and effective medical treatment.

Conclusion

Previous social support research has evidenced the physical and psychological benefits of offline and online social support, but little research has identified theoretical predictors of social support. Our findings reveal that, much like a protagonist in story, vloggers who provided narrative exposition and demonstrated qualities of a strong stereotypical protagonist were most likely to elicit support in comments. These findings synthesize work from social support, coping, and narrative literatures, and provide a framework for understanding how patients might increase social capital and better cope with their cancer experience. Additional work in this area may develop into a robust theoretical framework for predicting support in other online spaces as well.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

References

Bird

, Coleman

, Danson

. Coping with melanoma-related worry: a qualitative study of the experiences and support needs of patients with malignant melanoma. Journal of Clinical Nursing, 2015; 24:937–947.

Davison

, Pennebaker

, Dickerson

. Who talks? The social psychology of illness support groups. American Psychologist, 2000; 55:205–217.

Dunkel-Schetter

, Folkman

, Lazarus

. Correlates of social support receipt. Journal of Personality and Social Psychology, 1987; 53:71–80.

Dunkel-Schetter

, Wortman

. (1982) The interpersonal dynamics of cancer: Problems in social relationships and their impact on the patient. In Friedman

, DiMatteo

, eds. Interpersonal issues in health care. New York: Academic Press, pp. 69–100.

Funch

, Mettlin

. The role of support in relation to recovery from breast surgery. Social Science and Medicine, 1982; 16:91–98.

Rains

, Young

. A meta-analysis of research on formal computer-mediated support groups: examining group characteristics and health outcomes. Human Communication Research, 2009; 35:309–336.

Vaux

. Variations in social support associated with gender, ethnicity, and age. Journal of Social Issues, 1985; 41:89–110.

Wortman

. Social support and the cancer patient. Conceptual and methodologic issues. Cancer, 1984; 53:2339–2362.

Neuling

, Winefield

. Social support and recovery after surgery for breast cancer: frequency and correlates of supportive behavior by family, friends and surgeon. Social Science and Medicine, 1988; 27:385–392.

10.

Schreurs

, de Ridder

. Integration of coping and social support perspectives: implications for the study of adaptation to chronic diseases. Clinical Psychology Review, 1997; 17:89–112.

11.

Berkman

. Assessing the physical health effects of social networks and social support. Annual Review of Public Health, 1984; 5:413–432.

12.

House

, Landis

, Umberson

. Social relationships and health. Science, 1988; 241:540–545.

13.

Uchino

, Cacioppo

, Kiecolt-Glaser

. The relationship between social support and physiological processes: a review with emphasis on underlying mechanisms and implications for health. Psychological Bulletin, 1996; 119:488–531.

14.

Kahneman

. (1999) Objective happiness. In Kahneman

, Diener

, Schwarz

, eds. Well-being: The foundations of hedonic psychology. New York: Russel Sage, pp. 3–25.

15.

Kreuter

, Green

, Cappella

, et al. Narrative communication in cancer prevention and control: a framework to guide research and application. Annals of Behavioral Medicine, 2007; 33:221–235.

16.

Bishop

, Yardley

. Constructing agency in treatment decisions: negotiating responsibility in cancer. Health, 2004; 8:465–482.

17.

Charon

. What to do with stories: the sciences of narrative medicine. Canadian Family Physician, 2007; 53:1265–1267.

18.

Chou

W-Y

, Hunt

, Folkers

, et al. Cancer survivorship in the age of YouTube and social media: a narrative analysis. Journal of Medical Internet Research, 2011; 13:e7.

19.

Midtgaard

, Stelter

, Rørth

, et al. Regaining a sense of agency and shared self-reliance: the experience of advanced disease cancer patients participating in a multidimensional exercise intervention while undergoing chemotherapy—analysis of patient diaries. Scandinavian Journal of Psychology, 2007; 48:181–190.

20.

DeLongis

, Holtzman

. Coping in context: the role of stress, social support, and personality in coping. Journal of Personality, 2005; 73:1633–1656.

21.

Hale

, Tighe

, Mongreau

. Effects of event type and sex on comforting messages. Communication Research Reports, 1997; 14:214–220.

22.

Billings

, Moos

. The role of coping responses and social resources in attenuating the stress of life events. Journal of Behavioral Medicine, 1981; 4:139–157.

23.

Crowley

. (2016) Marshaling social support. In Berger

, Roloff

, eds. The international encyclopedia of interpersonal communication. West Sussex, United Kingdom: John Wiley and Sons, pp. 1046–1051.

24.

Holland

, Holahan

. The relation of social support and coping to positive adaptation to breast cancer. Psychology and Health, 2003; 18:15–29.

25.

Klemm

, Bunnell

, Cullen

, et al. Online cancer support groups: a review of the research literature. Computers, Informatics, Nursing, 2003; 21:136–142.

26.

Sposito

, Silva-Rodrigues

, Sparapani

, et al. Coping strategies used by hospitalized children with cancer undergoing chemotherapy. Journal of Nursing Scholarship, 2015; 47:143–151.

27.

Braithwaite

, Waldron

, Finn

. Communication of social support in computer-mediated groups for people with disabilities. Health Communication, 1999; 11:123–151.

28.

Walther

, Boyd

. (2002) Attraction to computer-mediated social support. In Lin

, Atkin

, eds. Communication technology and society: audience adoption and uses. Cresskill, NJ: Hampton, pp. 153–188.

29.

Wright

, Bell

. Health-related support groups on the Internet: linking empirical findings to social support and computer-mediated communication theory. Journal of Health Psychology, 2003; 8:39–54.

30.

Kellas

. (2016) Narratives and social interaction. In Berger

, Roloff

, eds. The international encyclopedia of interpersonal communication. West Sussex, United Kingdom: John Wiley and Sons, pp. 1162–1166.

31.

Kellas

, Manusov

. What's in a story? The relationship between narrative completeness and adjustment to relationship dissolution. Journal of Social and Personal Relationships, 2003; 20:285–307.

32.

McAdams

. The problem of narrative coherence. Journal of Constructivist Psychology, 2006; 19:109–125.

33.

Coleman

. Social capital in the creation of human capital. American Journal of Sociology, 1988; 94:S95–S120.

34.

Lin

. Building a network theory of social capital. Connections, 1999; 22:28–51.

35.

Reis

, Franks

. The role of intimacy and social support in health outcomes: two processes or one?. Personal Relationships, 1994; 1:185–197.

36.

Liu

, Huh

, Neogi

, et al. (2013) Health vlogger-viewer interaction in chronic illness management. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems. Paris. pp. 49–58.

37.

High

, Dillard

. A review and meta-analysis of person-centered messages and social support outcomes. Communication Studies, 2012; 63:99–118.

38.

High

, Scharp

. Examining family communication patterns and seeking social support direct and indirect effects through ability and motivation. Human Communication Research, 2015; 41:459–479.

39.

Lynch

. Who supports whom? How age and gender affect the perceived quality of support from family and friends. The Gerontologist, 1998; 38:231–238.

40.

Moon

S-S

, Park

S-M

, Cho

S-I

. The association of social support and activities with health in South Korea: differences in age and gender. Journal of Biosocial Science, 2010; 42:409–424.

41.

Goldsmith

, Fitch

. The normative context of advice as social support. Human Communication Research, 1997; 23:454–476.

42.

Smith

, Goodnow

. Unasked-for support and unsolicited advice: age and the quality of social experience. Psychology and Aging, 1999; 14:108–121.

43.

Burleson

, MacGeorge

. (2002) Supportive communication. In Knapp

, Daly

, eds. Handbook of interpersonal communication. Thousand Oaks, CA: Sage Publications, pp. 374–424.

44.

Cunningham

, Barbee

. (2000) Social support. In Hendrick

, Hendrick

, eds. Close relationships: a sourcebook. Thousand Oaks, CA: Sage, pp. 272–285.

45.

Gottlieb

, Wagner

. (1991) Stress and support processes in close relationships. In Eckenrode

, ed. The social context of coping. New York: Plenum, pp. 165–188.

46.

Iso-Ahola

, Park

. Leisure-related social support and self-determination as buffers of stress-illness relationship. Journal of Leisure Research, 1996; 28:169–187.