This study examined differential responses among partners who participated in a RCT designed to compare two social cognitive theory interventions, one designed for patients only (P-only) and one for patients and their intimate partners (P + P). The interventions were delivered following the patient receiving an initial ICD implant. Partner health outcomes were examined longitudinally from baseline at hospital discharge to 3, 6, and 12 months. Outcomes included 6 measures: partner physical and mental health status (Short-Form-36 PCS and MCS), depression (Patient Health Questionnaire-9), anxiety (State-Trait Anxiety Inventory), caregiver burden (Oberst Caregiver Burden Scale), and self-efficacy in ICD management (Sudden Cardiac Arrest Self-efficacy scale). Growth mixture and mixed effect modeling were used to identify and compare trajectories of 6 health outcomes within the P-only and P + P arms of the study. Partners (n = 301) were on average 62 years old, female (74.1%) and Caucasian (83.4%), with few co-morbidities (mean Charlson Co-morbidity index, 0.72 ± 1.1). Two types of profiles were observed for P-only and P + P, one profile where patterns of health outcomes were generally better across 12 months and one with outcome patterns that were generally worse across time. For PCS, no significant partner differences were observed between P-only or P + P in either the better (p = 0.067) or the worse (p = 0.129) profile types. Compared to P-only, partners in the worse profile improved significantly over 12 months in MCS (p = 0.006), caregiver burden P + P (p = 0.004) and self-efficacy P + P (p = 0.041). Compared to P-only, P + P partners in the low anxiety profile improved significantly (p = 0.001) at 3 months. Partners with more psychosocial distress at hospital discharge benefited most from the P + P intervention. Among partners with generally low levels of anxiety, those in the P + P intervention compared to P-only showed greater improvement in anxiety over 12 months.