Do AI-enabled remote patient monitoring programs reduce Readmissions? A quasi-experimental evaluation of adoption timing and outcomes.

Background

The increasing prevalence of chronic illnesses among Medicare populations requires a shift in care delivery towards a more proactive and sustained surveillance and monitoring approach. Remote patient monitoring paired with technology-enabled, guideline based clinical assistance, so-called remote patient care (RPC), has demonstrated potential benefits in clinical outcome enhancement, but evidence of whether it will produce the sustained declines in acute healthcare usage is mixed.

Objective

This analysis aimed to establish the effectiveness of an AI-supported RPC program in the reduction of hospital readmissions, emergency department (ED) visits, and length of stay (LOS), and to establish which institutional moderators contribute to the success of the program.

Methods

We performed a patient-level retrospective, quasi-experimental analysis based on a 500-patient multi-state cohort of data. It used a Difference-in-Differences (DiD) design to determine the effect of the AI-based iCARE RPC program relative to a propensity-scorematched control group in 12 months of post-activation follow-up.

Results

The RPC program was linked to a statistically significant decrease in the 30-day readmissions (Treatment Effect: β1 = -0.0381; p < 0.0001). Robustness tests proved negative changes in ED visits (-0.0043) and LOS (-0.2044).

Conclusion

High digital maturity (coefficient -0.0242 vs -0.0118) and strong staff capacity (coefficient -0.0085 vs -0.0019) were found to significantly boost the efficacy of the program among high-risk patients. Nevertheless, the digital multiplier of combined EHR systems and the human-in-the-loop requirement of sufficient nursing personnel are the key conditions to achieve the clinical and healthcare outcomes of remote monitoring technologies. Revised manuscript (unmarked, editable source file).