Table 4.
Trial outcomes: Intent to treat and adjusted intent to treat
| Intent to treata | Adjusted intention to treatb | |||
|---|---|---|---|---|
| Difference (95% CI) | Hedges’ g SMD (95% CI) | Difference (95% CI) | Hedges’ g SMD (95% CI) | |
| Primary outcome (post-intervention) | ||||
| Leader self-efficacy (SSGLSS) scorec | 16.68 (11.04, 22.32) | 0.84 (0.58, 1.09) | 17.05 (11.81, 22.29) | 0.85 (0.59, 1.11) |
| Secondary outcomes (post-intervention) | ||||
| Emotional distress (PHQ-8) score | − 0.93 (− 1.84, − 0.03) | − 0.26 (− 0.50, − 0.02) | − 1.15 (− 2.03, − 0.26) | − 0.32 (− 0.56, − 0.08) |
| Burnout (OLBI) score (among experienced leaders) | − 0.53 (− 1.66, 0.59) | − 0.13 (− 0.38, 0.11) | − 0.64 (− 1.81, 0.54) | − 0.16 (− 0.41, 0.09) |
| Volunteer satisfaction (VSI) score (among experienced leaders) | 5.08 (3.25, 6.91) | 0.70 (0.46, 0.94) | 4.94 (3.05, 6.82) | 0.68 (0.45, 0.92) |
| Secondary outcomes (3 months post-intervention) | ||||
| Leader self-efficacy (SSGLSS) score | 15.61 (10.24, 20.97)d | 0.73 (0.49, 0.98)d | 16.20 (10.63, 21.78)d | 0.76 (0.51, 1.01)d |
| Emotional distress (PHQ-8) score | − 0.81 (− 1.86, 0.25) | − 0.19 (− 0.41, 0.02) | − 1.01 (− 2.03, 0.00) | − 0.24 (− 0.46, − 0.03)e |
| Burnout (OLBI) score (among experienced leaders) | − 0.91 (− 1.94, 0.13) | − 0.25 (− 0.51, 0.00) | − 0.98 (− 2.05, 0.08) | − 0.28 (− 0.53, − 0.02)e |
| Volunteer satisfaction (VSI) score (among experienced leaders) | 3.91 (2.10, 5.72) | 0.53 (0.31, 0.75) | 3.91 (2.09, 5.72) | 0.53 (0.32, 0.75) |
All models are presented with multiply imputed data. For the Scleroderma Support Group Leader Self-Efficacy Scale and Volunteer Satisfaction Index, positive numbers favour the intervention. For the Patient Health Questionnaire-8 and Oldenburg Burnout Inventory, negative numbers favour the intervention. Analyses of SSGLSS and PHQ-8 include all 148 participants, whereas analyses of OLBI and VSI include only the 114 experienced participants, since these are not relevant to candidate leaders
CI confidence interval, PHQ-8 Patient Health Questionnaire-8, OLBI Oldenburg Burnout Inventory, SMD standardized mean difference, SSGLSS Scleroderma Support Group Leader Self-efficacy Scale, VSI Volunteer Satisfaction Index
aAdjusted for baseline outcome score only
bAdjusted for baseline score plus age (continuous); sex (male vs. female); whether the leader has scleroderma (no vs. yes); and, for outcomes that included both experienced and new leaders, whether the leader was a prospective vs. experienced leader
cModels were simplified to remove random slopes and to only include a single random intercept for each separate intervention group and for all waitlist participants combined to facilitate convergence
dOne intervention arm participant scored 118 on the Scleroderma Support Group Leader Self-Efficacy Scale (possible scores 32–192) at baseline, 187 post-intervention, and 32 (all items “strongly disagree”) at 3 months post-intervention. We inquired with the participant about the unusual variability in scores, and she indicated that she had intended to score all items as “strongly agree” at 3 months post-intervention (score = 192) but had mistakenly responded backwards. Thus, her score was counted as missing at 3 months post-intervention
eThe Borenstein and Hedges’ formula for converting regression coefficients to standardized mean differences was adapted for use with the PN-RCT design and, thus, closely approximates the SMD and 95% CI. In two cases, there were discrepancies between statistical significance based on raw scale scores and Hedges’ g; the raw score results should be interpreted for statistical significance. See Additional file 6: S6