Abstract Body (Do not enter title and authors here): Background: Home health aides (HHAs) frequently care for heart failure (HF) patients, but many lack training, confidence, and cannot reach their nurse supervisors when they need help. We examined the feasibility, acceptability, and effectiveness of a novel education and communication-based intervention among HHAs in HF.

Methods: In partnership with a large home care agency in NY, we conducted a pilot randomized control trial (RCT) with HHAs caring for a HF patient. HHAs were randomized to: a) HF training (control) or b) HF training plus mHealth app that allowed HHAs to text-message nurses in real-time (intervention) and followed for 90-days. Using the RE-AIM implemention framework, feasibility and acceptability were assessed with surveys and interviews. Effectiveness (primary) outcomes: HF knowledge (DHFK; 0-15) and HF Caregiver Self-Efficacy (SE; 0-100), secondary outcome: preventable 911 calls. Patient outcomes (exploratory): ED visits and hospitalization. Mixed effects models compared trajectory of outcomes between and within arms. Qualitative data were analyzed thematically.

Results: Of 386 screened HHAs, 105 consented, and 102 completed baseline surveys (n=50 control; n=52 intervention), comprising the analytic sample. Mean age of 54 years (10.5 SD), 96% women, 62% Black. Baseline median DHFK score was 7.0 (IQR 5.0, 8.0) and SE score was 75.0 (62.5, 82). Overall, 80.4% completed HF training and 73% were retained at 90-days. At 90-days, DHFK and SE scores improved within both arms, but there was not a significant difference between arms. The proportion of HHAs with preventable 911 calls decreased in the intervention arm (0.51 vs. 0.34, p=0.06) and increased in the control arm (0.42 vs. 0.54, p=0.21); this difference was significant (p=0.04). Patients cared for by intervention HHAs were 56% less likely to have an ED visit vs. control (IRR 0.56 [0.25-1.28], p=0.17). Overall 65% of intervention HHAs texted their nurse; of these 42% texted regularly about clinical care. HHAs reported high levels of satisfaction, although some noted technology barriers.

Conclusion: We recruited, retained, and engaged HHAs, a workforce that has rarely been the focus of HF interventions. Training improved HHAs HF knowledge and caregiving SE. The intervention reduced preventable 911 calls. Patients cared for by intervention HHAs had fewer ED visits. Findings can inform the design of a future large-scale trial, and how to better support and integrate HHAs providing care HF care.