Abstract Body (Do not enter title and authors here): Background: The PATRIOT trial (2015-2021) tested a personalized behavioral intervention (PBI) to prevent diabetic foot lesions by targeting general and foot-specific diabetes care. From the hospitalizations and deaths in the trial, we evaluated if time to cardiovascular events (CVE) or all-cause events (ACE) differed between PBI and current best practice (CBP).

Methods: Adults with diabetic neuropathy were recruited from two urban Veterans Affairs hospitals and randomly assigned in parallel to PBI or CBP. The PBI included monthly phone calls targeting general diabetes care (medication adherence, physical activity, diet) and foot-specific care. CBP calls focused on non-cardiovascular and non-diabetes care. Participants were followed for 1.5 years post randomization. Hospitalizations and deaths were assessed from medical record review and participant/family report. After examining if CVEs and ACEs differed between arms by logistic regression, we tested if time to CVEs or ACEs differed by arm, using Kaplan-Meier curves and Cox models. Further, to account for repeat events while evaluating PBI, we used negative binomial regression. We report all model estimates with 95% confidence intervals (CI).

Results: We randomized 406 participants (203 per arm). Mean (SD) age was 67.8 (8.4) years in PBI and 68.0 (8.8) years in CBP. Both arms were similar in race, education, social support, and cardiovascular risk factors (all P > .05). There were 151 hospitalizations and 4 deaths. In total, 92 participants had at least one event and 31 had multiple events. The mean (variance) of ACEs was 0.37 (0.78). Most events (44) were CVEs (18 ischemic, 13 arrhythmic, 8 heart failure, and 5 vascular). The odds ratio for CVE in PBI vs. CBP was 0.29 (CI: 0.11, 0.67). There was longer time to CVE for PBI compared to CBP with the 1.5-year cumulative probability of CVE-free survival being 0.95 in PBI and 0.84 in CBP (log-rank P < 0.01). Controlling for age and race, the hazard ratio for CVE comparing PBI to CBP was 0.30 (CI: 0.13, 0.67). Incorporating repeat CVEs, the rate ratio comparing PBI to CBP was 0.30 (CI: 0.13, 0.67). For ACE, PBI was superior to CBP, consistent with all the CVE analyses.

Conclusions: The PBI led to greater CVE-free and ACE-free survival. Such comprehensive interventions that target a specific outcome also holds promise for preventing other serious outcomes, probably through raised health consciousness and more healthful behaviors.

Registration: NCT02356757