A theory-based video intervention to enhance communication and engagement in online health communities: two experiments

Figures & data

Figure 1. Screenshot of the intervention video with the communication strategies based on Self-Determination Theory (content translated from German to English).

Figure 2. Participant flowchart for Study 2.

Table 1. Baseline characteristics and participant differences in the control and intervention conditions (Study 1).

Variable	Control group (n = 36)	Intervention group (n = 40)	p
Demographic characteristics
Age (years), M (SD)	26.00 (13.30)	24.52 (9.53)	.584
Female, n (%)	29 (81)	30 (75)	.594
Educational attainment, n (%)
Low (ISCED 0–2)	0 (0)	0 (0)	.881
Medium (ISCED 3 & 4)	28 (78)	32 (80)
High (ISCED 5–8)	7 (19)	8 (20)
Other	1 (3)	0 (0)
Outcome (preintervention)
Need-supportive communication strategy use (number of need-supportive communication strategies), M (SD)	5.47 (3.1)	4.65 (2.51)	.211

Note: The control and intervention groups were compared with Welch’s two-sample t tests (means) or Fisher’s exact tests (proportions) whereby p shows the significance level of the comparisons. ISCED = International Standard Classification of Education.

Figure 3. Mean number of need-supportive communication strategies used (Study 1).

Note: Means and standard errors (error bars) for the number of need-supportive communication strategies used in the responses to fictive online postings, by measurement time and experimental condition.

Table 2. Baseline characteristics and participant differences in the control and intervention group (Study 2).

Variable	Control group (n = 301)	Intervention group (n = 236)	p
Demographic characteristics
Age (years), M (SD)	42.02 (11.64)	43.46 (11.30)	.150
Body mass index (kg/m^2), M (SD)	28.52 (6.52)	28.57 (6.69)	.939
Female, n (%)	294 (97.7)	231 (97.9)	.999
Educational attainment, n (%)			.999
Low (ISCED 0–2)	3 (1.0)	2 (0.9)
Medium (ISCED 3 & 4)	138 (45.9)	108 (45.8)
High (ISCED 5–8)	159 (52.8)	126 (53.4)
NA	1 (0.3)	0 (0.0)
Occupational skill level, n (%)			.942
Low (ISCO skill level 1, e.g. unskilled worker)	24 (8.0)	16 (6.8)
Medium (ISCO skill level 2, e.g. skilled worker)	117 (38.9)	92 (39.0)
High (ISCO skill levels 3 & 4, e.g. higher skilled worker/academic job)	159 (52.8)	127 (53.8)
NA	1 (0.3)	1 (0.4)
Professional position, n (%)			.300
Full-time employees	152 (50.5)	108 (45.8)
Part-time employees	74 (24.6)	73 (30.9)
Students (higher education)	25 (8.3)	23 (9.8)
Other	50 (16.6)	32 (13.6)
Goal type, n (%)			.695
Fruit intake	20 (6.6)	15 (6.4)
Vegetable intake	84 (27.9)	56 (23.7)
Moderate physical activity	84 (27.9)	67 (28.4)
Vigorous physical activity	113 (37.5)	98 (41.5)
Outcomes (baseline)
Fruit intake (no. portions), M (SD)	1.51 (1.01)	1.50 (0.94)	.957
Vegetable intake (no. portions), M (SD)	2.05 (1.41)	1.83 (1.25)	.048
Moderate physical activity (min/week), M (SD)	214.70 (193.09)	203.23 (192.77)	.494
Vigorous physical activity (min/week), M (SD)	33.19 (35.72)	32.56 (35.51)	.837
Autonomous motivation, M (SD)	3.94 (1.00)	3.85 (0.92)	.311
Controlled motivation, M (SD)	4.12 (1.52)	4.17 (1.51)	.720
Instrumental attitude, M (SD)	7.00 (0.00)	7.00 (0.00)	N/A^a
Experiential attitude, M (SD)	5.50 (1.14)	5.55 (1.14)	.593
Self-efficacy, M (SD)	3.82 (0.67)	3.78 (0.68)	.582
Perceived descriptive norms, M (SD)	3.88 (0.68)	3.95 (0.68)	.219
Perceived injunctive norms, M (SD)	3.94 (0.78)	4.01 (0.80)	.271
Other
Number of other active forum members, M (SD)	165.92 (93.77)	167.57 (95.09)	.841
Omnivore diet, n (%)	185 (61.5)	153 (64.8)	.472
Weight loss diet, n (%)	70 (23.3)	53 (22.5)	.837
Fructose intolerance, n (%)	288 (95.7)	226 (95.8)	.999

Note: The control and intervention groups were compared with Welch’s two-sample t tests (means) or Fisher’s exact tests (proportions), whereby p shows the significance of the comparisons. The outcomes goal attainment, perceived need support, perceived social support, number of logins, number of postings, and subjective forum use frequency are only available at follow-up. ISCED = International Standard Classification of Education. ISCO = International Standard Classification of Occupations. NA = Missing values. N/A = Not applicable.

^a Not applicable because the distribution of the variable was highly skewed and there was little variance; after winsorization, all participants scored the highest value on the scale (7).

Figure 4. Percentage of postings with a specific number of need-supportive communication strategies based on Self-Determination Theory by experimental condition (Study 2).

Note: The maximum possible number of strategies per posting is 6. Error bars represent 95% confidence intervals.

Table 3. Intervention effects and estimated marginal means for intervention and control condition from mixed models (Study 2).

Variable	Estimated marginal mean (SE)		Intervention effect estimate B (SE)^a	p
	Control group (n = 301)	Intervention group (n = 236)
Proximal outcomes
Number of need-supportive communication strategies^b	1.48 (0.03)	1.44 (0.03)	−0.01 (0.05)	.778
Primary outcomes
Perceived need support	3.03 (0.16)	3.08 (0.17)	0.05 (0.22)	.840
Goal attainment	1.12 (0.07)	1.22 (0.07)	0.11 (0.09)	.260
Number of postings^b	1.81 (0.20)	2.42 (0.28)	0.31 (0.15)	.046
Number of logins^b	3.51 (0.33)	3.59 (0.25)	0.10 (0.08)	.250
Subjective forum visit frequency	2.47 (0.10)	2.71 (0.10)	0.24 (0.13)	.101
Secondary outcomes
Autonomous motivation	3.99 (0.06)	3.96 (0.09)	−0.03 (0.10)	.795
Controlled motivation	4.20 (0.10)	4.09 (0.11)	–0.11 (0.13)	.384
Self-efficacy	3.39 (0.04)	3.39 (0.05)	–0.01 (0.06)	.915
Experiential attitude	5.13 (0.11)	5.20 (0.12)	0.07 (0.16)	.669
Instrumental attitude	6.58 (0.03)	6.64 (0.04)	0.06 (0.05)	.319
Perceived descriptive norms	3.77 (0.09)	3.77 (0.05)	0.00 (0.10)	.977
Perceived injunctive norms	3.38 (0.08)	3.47 (0.08)	0.09 (0.10)	.389
Perceived social support	3.23 (0.18)	3.28 (0.15)	0.05 (0.22)	.813

Note: Analyses were conducted with winsorized values and the effect of the intervention condition (dummy coded) is controlled for baseline values of the outcome, variables with baseline differences between completers and noncompleters, and between the 4 self-selected goal types, that is age, fruit intake, vegetable intake, moderate physical activity, vigorous physical activity, body mass index, perceived descriptive norms, perceived injunctive norms, and the mean number of active forum users. The raw values were used for the poisson-distributed count variables number of need-supportive communication strategies, number of logins and number of postings.

^a Differences between estimated marginal means and estimates originate from rounding.

^b For the number of postings and the number of logins, the intervention’s effect is multiplicative (e^estimate) rather than additive since the models use a log-link-function for the count data. M (SE) represents the raw values because estimated marginal means could not be derived for nonlinear mixed models.

Data availability statement

The data that support the findings of this study are openly available in the Open Science Framework at https://osf.io/xsb2e/.