Social Support in Online Peer Groups for Celiac Disease

Peer support, especially online, is receiving increasing attention as effective and cost-efficient support for people living with chronic health conditions (e.g., Hossain et al., 2021). The present study examined the influence of online peer social support groups on well-being and dietary compliance among people living with celiac disease. Like most people living with a chronic condition, people with celiac disease need to learn and maintain self-management behaviors, consisting primarily of strict compliance with a lifelong gluten-free diet. A lifelong diet, however, requires effort and persistence, which people with celiac disease often perceive as diminishing their quality of life (Hall et al., 2009). Social support can help patients master this challenge (e.g., DiMatteo, 2004). Social support involves the interaction between two or more people in which the goal is to change or facilitate the endurance of a problem state that creates suffering in an individual (Schwarzer, 2004). Importantly, social support has been shown to positively correlate with patient adherence across a variety of disease conditions, including diseases (e.g., diabetes mellitus) that require, similar to celiac disease, complex and long-term health behaviors such as a strict and life-long diet (DiMatteo, 2004).

However, all kinds of social support from all kinds of supporters are not considered equally helpful, especially when it comes to chronic diseases (e.g., Cutrona & Russell, 1990). Specifically, social support seems most likely to have positive effects when it is delivered by people coping with the same stressors (e.g., Ussher et al., 2006), who are often considered to be more empathetic and credible compared to healthy persons (Mehnert et al., 2018). In daily life, it is often hard to identify and meet with people affected in a similar way. Opportunities to connect online with peers are thus an especially valuable resource. However, while there is good evidence for the positive effects of onsite peer social support on well-being and health behavior (e.g., Fisher et al., 2012), there are only a few inconsistent findings on the impact of online self-help groups (Hossain et al., 2021; Rains & Young, 2009) with some suggesting positive and others no effects.

The aim of the present research is twofold. First, it seeks to determine whether reports of received online peer social support are positively related to self-reported dietary compliance and well-being among people living with celiac disease. In doing so, we differentiate between two functions of social support that have been consistently shown to make up the most share of posts in online peer support groups (e.g., Mo & Coulson, 2008), and explore whether they are similarly related to dietary compliance and well-being: Informational online peer support (IS), which refers to experiential advice or information shared online in response to a question, and emotional online peer support (ES), which includes receiving understanding, reinforcement, and affirmation from peers by recognizing and responding to feelings in an encouraging manner (Mo & Coulson, 2008). Second, we aim to explain inconsistent findings in previous research by considering the duration of the chronic condition as a moderator. Specifically, we assume that IS should become less important over the course of the illness, as people’s own knowledge about how to deal with a chronic condition should usually increase over time (e.g., Chen, 2016). The importance of ES, on the other hand, should remain relatively constant over time, as sticking with a strict diet usually remains a challenge and results in recurring problems in social life (Lee & Newman, 2003). These assumptions are in line with previous work. For instance, Cutrona and Russell (1990) suggest that IS facilitates stress management, especially at the beginning of a stressful event, while long-lasting events make ES necessary. A study by Drageset et al. (2012) supports this assumption by showing that women with breast cancer considered IS especially important at the beginning, while ES remained significant at later time points.

Please note that we were primarily interested in the potential role of support actually received in the past in people’s well-being and dietary compliance, and less in the role of expected support in the future (in the sense of a stable personality trait). We thus included a measure of actually received support in our study, but no measure of perceived available support (see Schwarzer & Knoll, 2007, for the distinction between these two forms).

Method

The study’s materials, data files, and analysis scripts are available at https://osf.io/adf6b.

The commissioner for research involving human participants of the FernUniversität in Hagen exempted this study from review.

Results

Relationship Between ES/IS and Well-Being/Dietary Compliance

Table 1 presents intercorrelations, means, and standard deviations of all theoretically relevant variables.

Table 1 Intercorrelations, means (Ms), and standard deviations (SDs) of all theoretically relevant variables

Variable	1	2	3	4	5
Note. *p < .05; **p < .01 (two-tailed).
1. Emotional support	–	.79**	.43**	.60**	−.16**
2. Informational support		–	.42**	.52**	−.19**
3. Well-being			–	.44**	.05
4. Dietary compliance				–	−.12*
5. Duration of illness					–
M	3.81	4.08	3.39	8.76	9.86
SD	0.95	0.77	0.80	2.51	10.09

Table 1 Intercorrelations, means (Ms), and standard deviations (SDs) of all theoretically relevant variables

View as image HTML

To test how ES and IS relate to the proposed criteria, we computed two regression analyses with ES and IS as proposed predictors, and either well-being or dietary compliance as a proposed criterion (see Table 2). As expected, ES and IS were (marginally) significantly positively related to well-being and dietary compliance.

Table 2 Results from multiple regression analyses with well-being (upper part) and dietary compliance (lower part) as proposed criterion

Proposed predictors	b	CI [LL, UL]	SE	β	t	p	Cohen’s f2
Proposed criterion: Well-being
Emotional support	0.22	95% CI [0.09, 0.34]	0.06	.26	3.44	< .001	0.03
Informational support	0.23	95% CI [0.07, 0.38]	0.08	.22	2.92	.004	0.02
Overall R2 = .20, F(2, 366) = 47.01, p < .001
Proposed criterion: Dietary compliance
Emotional support	1.32	95% CI [0.97, 1.67]	0.18	.50	7.43	< .001	0.15
Informational support	0.42	90% CI [0.06, 0.78]	0.22	.13	1.94	.053	0.01
Overall R2 = .37, F(2, 366) = 106.38, p < .001

Table 2 Results from multiple regression analyses with well-being (upper part) and dietary compliance (lower part) as proposed criterion

View as image HTML

Duration of Illness as a Moderator

To test the assumption that IS becomes less important over the course of the illness while the importance of ES remains constant, we conducted two moderated regression analyses with either well-being or dietary compliance as the proposed criterion, in which IS, ES, and duration of illness (all mean-centered prior to analyses) were added in a first step and the relevant interaction terms (IS × Duration, ES × Duration) in a second step (see Table 3; Figure 1).

Table 3 Results from moderated regression analyses with well-being (upper part) and dietary compliance (lower part) as proposed criterion

Proposed predictors	b	95% CI [LL, UL]		SE	β	t		p	Cohen’s f2
Proposed criterion: Well-being
Step 1
Emotional support (A)	0.22	[0.10, 0.34]		0.06	.26	3.53		< .001	0.03
Informational support (B)	0.25	[0.10, 0.40]		0.08	.24	3.24		.001	0.02
Duration of illness (C)	0.01	[0.004, 0.02]		0.004	.14	2.96		.003	0.02
R2 = .22, F(3, 365) = 34.93, p < .001
Step 2
(A) × (C)	0.01	[−0.004, 0.02]		0.01	.10	1.30		.196	0.007
(B) × (C)	−0.003	[−0.02, 0.01]		0.01	−.04	−0.48		.631	0.003
ΔR2 = .01, F(2, 363) = 1.20, p = .303; overall R2 = .23, F(5, 363) = 21.46, p < .001
Proposed criterion: Dietary compliance
Step 2
Emotional support (A)	1.32	[0.97, 1.67]		0.18	.50	7.42		< .001	0.15
Informational support (B)	0.41	[−0.02, 0.85]		0.22	.13	1.89		.060	0.01
Duration of illness (C)	−0.004	[−0.03, 0.02]		0.01	−.02	−0.41		.683	0.003
R2 = .37, F(3, 365) = 70.82, p < .001
Step 2
(A) × (C)	−0.01	[−0.04, 0.03]		0.02	−.02	−0.29		.772	0.007
(B) × (C)	0.04	[0.01, 0.08]		0.02	.16	2.26		.024	0.02
ΔR2 = .02, F(2, 363) = 5.31, p = .005; overall R2 = .39, F(5, 363) = 45.62, p < .001

Table 3 Results from moderated regression analyses with well-being (upper part) and dietary compliance (lower part) as proposed criterion

View as image HTML

Well-Being as Proposed Criterion

The ES × Duration of illness interaction was non-significant (see Table 3, upper part). Subsequent simple slope analyses (Hayes, 2018) indicated that ES was (marginally) significantly and positively related to well-being independently of the duration of illness, that is, among participants scoring in the 16th percentile (shorter duration of illness) and participants scoring in the 84th percentile (longer duration of illness) (see Figure 1A; see Electronic Supplementary Material, ESM 1 for a more detailed description of the results of all simple slopes analyses). The IS × Duration interaction was non-significant as well (see Table 3, upper part), indicating that IS was significantly positively related to well-being independent of the duration of illness (see Figure 1B).

Dietary Compliance as Proposed Criterion

The ES × Duration interaction was again non-significant (see Table 3, lower part), with ES being significantly and positively related to dietary compliance independent from the duration of illness (see Figure 1C). The IS × Duration interaction term was significant (see Table 3, lower part), indicating that, contrary to expectations, IS was significantly positively related to dietary compliance among participants with a longer duration of illness, but unrelated among participants with a shorter duration (see Figure 1D).

Discussion

Our results confirm that people with celiac disease can benefit from both functions of online peer social support, though effect sizes were larger for received ES compared to received IS. This finding is in line with other studies emphasizing the crucial role of ES in online contexts (e.g., Oh et al., 2013). Moreover, the observed effect sizes are comparable to those found in offline settings (see DiMatteo, 2004), suggesting that online peer ES can affect health behavior to a similar extent as face-to-face ES. Future research is needed to address the underlying mechanisms of the relationship between online peer social support and health outcomes. Numerous studies found that such adaptive outcomes are not affected directly by social support, but indirectly by increasing self-efficacy expectations (e.g., Schwarzer & Knoll, 2007). The present study included a measure of self-efficacy that was positively related to both functions of social support as well as both health outcomes (see ESM 1). However, due to the cross-sectional design, we cannot address the predictive direction of this relationship.

Moderation analyses confirmed that the strength of the positive relationships between ES and both health outcomes were independent of the duration of illness, suggesting that ES is an important resource throughout the course of celiac disease. Contrary to expectations, the importance of IS for health outcomes did not diminish over time: It seems that IS is positively related to well-being throughout the course of celiac disease and becomes even stronger related to dietary compliance with increasing duration of illness. One reason for this finding could be that knowing about the latest diet recommendation plays a particularly important role among people with celiac disease, thus making IS a valuable resource even later in the course of the disease.

One limitation of the study’s design is that one cannot infer the causal role of IS and ES. Future work should thus attempt to replicate the present findings using a longitudinal or experimental design. Moreover, given that both functions of online social support were highly correlated, one might question whether a distinction between IS and ES can be justified in our study. Still, ES was more strongly related to dietary compliance than IS, suggesting the former to be more useful for people living with celiac disease (see optimal matching theory, e.g., Cutrona & Russell, 1990). Future research could test this assumption by explicitly measuring the perceived match between different functions of online social support and stressor demands in the context of chronic diseases. In addition, to gain a comprehensive understanding of the role of online peer social support in chronic conditions, further research examining additional potential moderators or contextual factors is needed (e.g., type of chronic condition). Finally, as in other research on online social support, the effects in our study are small to moderate. They are still of practical relevance, though: Given that online peer social support is a very cost-effective method that can reach large numbers of people – perhaps even especially those most in need of support – it can have a considerable impact on a societal level. It should thus be proactively advertised in addition to established therapies.

Electronic Supplementary Materials