Testing the mediating effects of self-efficacy on physical and psychological outcomes following an Implantable Cardioverter Defibrillator (ICD)

Abstract

Purpose/Aims: Approximately 100,000 persons with heart failure or cardiac arrhythmias receive an implantable cardioverter defibrillator (ICD) each year in the U.S. Post-ICD implant recovery requires multiple lifestyle changes that can markedly influence the recovery trajectory. Few intervention programs have been based on social-cognitive theory, and none have included the patient’s significant other in intervention programs. The purpose of this study was (I) to test the mediating effects of four dimensions of self-efficacy (self-efficacy expectations, outcome expectations, self-management behaviors, and knowledge assessment), and (II) to differentiate the mediating effects of eight content-specific elements of self-efficacy (SE-ICD function, SE-physical changes, SE- lifestyle, SE-emotions, SE-ICD shock, SE-safety, SE-cognition, SE-relationship) on the relationship between the P+P intervention conditions and patient physical function and psychological adjustment following completion of the P+P trial at 12 months. This research was also designed to (III) compare if ICD indication (primary vs. secondary prevention) moderated the strength of the mediation pathways from the intervention through self-efficacy to the physical function and psychological adjustment at 12 months post ICD-implant. Methods: This was a secondary analysis based on data from a randomized clinical trial that compared two interventions: Patient+Partner (P+P) vs. Patient-Only intervention (P-Only). Data used on this study were collected at hospital discharge, 3 and 12-months post-ICD implant. Structural equation modeling (SEM) with multiple indicators was used to test for the mediating effects of the dimensions of self-efficacy and content-specific elements of self-efficacy on physical function and psychological adjustment. ANOVA was used to compare study groups/ICD indication at baseline, 3 and 12-months and across time. Multiple group SEM was used to examine for differences in models for individuals receiving an ICD for primary vs. secondary prevention of sudden cardiac arrest. Results: Participants (N = 301) were primarily male and white with a mean age of 64.14 (11.90). Mean ejection fraction was 34.08(14.33) with ICD implanted for primary (59.8%) and secondary prevention (40.2%). The models had satisfactory to good fit indices. There was an indirect effect of self-efficacy between the paths of interventions, physical function: β = 0.035 p = 0.07 and psychological adjustment β = 0.023, p = 0.07. The interventions also showed effects on physical function (β = 0.056 p = 0.08) and (β = 0.028 p = 0.07) psychological adjustment through outcome expectation. For the content specific elements of self-efficacy, a mediation effect between intervention condition and both psychological adjustment and physical function were observed for SE-Physical changes (β = 0.053 p = <0.05; β = 0.028 p = 0.07, respectively). The indirect effect for the mediation of SE-Relationship between intervention condition and psychological adjustment was β = 0.049 p = 0.08. Comparing primary vs. secondary prevention, no statistically significant differences were shown for the dimensions of self-efficacy at baseline and three-months. The dimensions of self-efficacy improved for both groups across time (all p<0.001) independent of ICD indication. For physical function, bodily pain scores were higher for those with ICD implanted for secondary prevention. Similarly, psychological adjustment for depression, role emotional and social functioning had higher scores for those who had a secondary prevention ICD. Across time, physical functioning, role physical and depression had the greatest improvement in secondary prevention participants. Five separate models, one for each dimension of self-efficacy, and a full model were tested. ANOVA comparisons between models yielded no differences in the structural model groups for primary vs. secondary ICD indication: (I) Self-Efficacy Model. ANOVA χ2diff = 32.4, dfdiff = 33, p = 0.50); (II) Outcome Expectations Model. ANOVA χ2diff = 42.22, dfdiff = 33, p = 0.13; (III) Self-Management Behavior Model. ANOVA χ2diff = 40.79, dfdiff = 33, p = 0.16; (IV) ICD Knowledge Model. ANOVA χ2diff = 40.60, dfdiff = 33, p = 0.17; (V) Full Model (SE + OE + SMB + KSA). ANOVA χ2diff = 45.15, dfdiff = 48, p = 0.59. Conclusion: For Aim 1, the effect of the P+P intervention vs. the P-only intervention on physical function and psychological adjustment was mediated by self-efficacy and by outcome expectations. For Aim 2, we observed that in the content-specific elements of self-efficacy analyses, the mediation pathways for psychological adjustment were with SE-physical changes and SE-relationship. For physical function, mediation was through SE-physical changes. For Aim 3, the hypothesized model describing the dimensions of self-efficacy mediation pathways between the interventions with physical function and psychological adjustment was similar for patients who received an ICD for either primary or secondary prevention of cardiac arrest. However, outcome trajectories differed by ICD indication, patients with an ICD implanted for secondary prevention had better physical (bodily pain) and psychological outcomes (depression, role emotional and social function). In addition, over 12-months period secondary prevention participants had sharpest improvement on physical functioning, role physical and depression.