In this national study of US hospitals from 2018 through 2022, there was a 1.4-fold increase in the number of hospitals with RPM services, with nearly half of US hospitals offering these services by 2022. Notably, the majority of this growth occurred between 2018 and 2020 (a 30% relative increase), while the rate of expansion slowed significantly thereafter, with only a 7.9% relative increase from 2020 onward. This increase was observed for both post-discharge and chronic care RPM services. The availability of RPM services varied widely based on hospital and neighborhood characteristics, with lower availability of these services at small, rural, non-teaching hospitals, especially those in the South. Moreover, hospitals in communities with low household incomes and a large proportion of Hispanic individuals had substantially lower RPM availability. Across study years, we observed an increasing proportion of HF and AMI hospitalizations at hospitals offering RPM services, rising from approximately 50% in 2018 to over 70% in 2022, highlighting the expanding reach of RPM services.

Patient-level studies, like the one using insurance claims by Tang et al. (2022), have suggested a substantial recent increase in claims for RPM service in fee-for-service Medicare and Medicaid beneficiaries in the US between 2019 and 2021.^6,26 This increase in the use of these services has occurred despite the lack of robust evidence on the clinical benefits of RPM or clarity on patients or conditions that benefit most from these services. Our findings complement the findings presented in these selected patient populations to provide evidence of the national landscape and describe the rapidly increasing proportion of US hospitals adopting RPM. However, we note that this uptake has varied widely across hospitals according to their size, type, and region of location. We observed that hospitals serving predominantly disadvantaged communities - that also consistently experience worse readmission outcomes¹³ – also have the lowest availability of RPM services. This national hospital-level analyses adds critical insight into the infrastructural capacities that enable or constrain the deployment of digital health technologies. The greater uptake of RPM services among larger, well-resourced hospitals may reflect their stronger institutional capacity to meet complex billing requirements, invest in digital infrastructure, and support the operational changes needed to scale remote monitoring, factors that may limit adoption in smaller or under-resourced settings and contribute to widening disparities in access.

These findings are pertinent given the growing evidence suggesting that emerging digital health technologies can improve patient outcomes.^18–21 Numerous society guidelines emphasize the value of RPM services in enhancing appropriate post-discharge care.^3,22–24 Specifically, many studies have suggested a role for guideline-directed therapy optimization in HF in the post-discharge period, while the critical services in AMI in this setting are less well-defined.^32,33 Emerging randomized clinical trial evidence, however, suggests that RPM specifically designed to target post-acute coronary syndrome, including symptom-triggered monitoring on home wearable ECG, blood pressure, and pulse oximeter, with close coordination with a cardiologist, was associated with 76% lower readmission and emergency department visits over 6 months.³⁴ These observations underscore the need to better characterize the nature of the services provided and their intended application to enable the evaluation of their impact on outcomes at hospitals.

This work also has important policy implications. The considerable investment in RPM services, further accelerated by the COVID-19 pandemic, has been supported by CMS and other payers through coverage in a fee-for-service model over the past five years.^12,35 Besides the high availability and broad insurance coverage of RPM services, there is no standard definition for RPM and no consistent outcome testing.¹² The observed heterogeneity in RPM service availability highlights that not all communities that could benefit from RPM have access to these services covered by national insurance providers. There is also a need to generate the necessary evidence for individual services to improve the efficiency of these investments. These considerations are especially critical as policymakers and payers consider the future of RPM coverage in the post-pandemic era.

Despite the introduction of new billing codes and financial incentives, RPM billing in the US remains highly concentrated among a small number of primary care providers.^28–30 At the same time, health systems have faced considerable challenges in establishing financially sustainable, equitable, and scalable RPM programs.⁹ While current reimbursement models do not provide sufficient constraints on the nature of offered services, they include certain requirements, such as in-person device setup, minimum data transmission, and documentation of interactive communication, which can represent procedural constraints that may affect less resourced hospitals. Specifically, hospitals serving low-income populations, where patients may lack access to Bluetooth-enabled devices or may only be able to submit data manually, thereby failing to meet reimbursement criteria despite meaningful clinical participation. As illustrated by recent implementation studies,^31,32 health systems have had to invest in device loaner programs and technical infrastructure to mitigate these limitations, yet still struggle to meet the criteria necessary for sustained billing under CPT codes. These payer-level constraints likely contribute to the geographic and structural disparities in RPM availability observed in our analysis.

Our study has certain limitations that merit consideration. First, we lack detailed information on the specific components of the RPM services evaluated. The RPM services reimbursed by payers under this category are not well-defined, underscoring the need to more specifically define the components of these services and their alignment with the available evidence for each condition. Second, while we have information on the characteristics of the communities served by these hospitals, we lack information on the number of patients and the nature of the conditions managed with these services at these hospitals. Effective implementation of RPM relies on patients having reliable internet access, suitable devices, digital literacy, and the capacity to engage with monitoring devices, factors often associated with higher socioeconomic status. Consequently, even within hospitals offering RPM, these services may predominantly benefit better-resourced patients who already have a lower baseline risk for adverse outcomes. Further investigations should focus on understanding the nature and quality of RPM services provided across different hospital types in the US, the specific patient populations benefiting from RPM services, the extent of their engagement, and the conditions most commonly managed using these technologies.