Table 4.

Randomized trials with neutral results based on external‐device remote patient monitoring (RPM)

Study name	Sample size	Study design and tested modality	Potential explanantion for lack of benefit
TIM‐HF (Koehler Circulation 2011)	N = 710 (355 on RPM)	Randomized trial of a Bluetooth‐enabled device designed to follow 3‐lead electrocardiography, BP, and weight	Participants had stable HF, so it may be that remote monitoring is not as effective in lower‐risk patients
Tele‐HF ( Chaudhry N Engl J Med 2010)	N= 1653 (826 on RPM)	Telephone‐based interactive voice response system with a higher risk population than in the TIM‐HF study	Patient adherence was poor, with <55% of the study subjects using the device 3 days per week by the end of the study. Interestingly, a smaller previous trial had shown benefit; this difference in results implies that how a technology is implemented might determine benefit
BEAT‐HF (Ong JAMA Intern Med 2016)	N = 1437 (715 on RPM)	Health‐coaching telephone calls with monitoring of weight, BP, HR, and symptoms in a high‐risk population with 50% rehospitalization rate	Nonadherence was the primary limitation, with only 61% of patients more than half‐adherent in the first 30 days
Mayo Clinic Study (Takahashi Arch Intern Med 2012)	N = 205 (102 on RPM)	Telemonitoring in a PC panel (various health conditions and not only HF) in the top 10% of Elder Risk Assessment Index managed with biometrics (BP, HR, weight, pulse oximetry, etc) plus daily symptom assessment. Video conference capability was present.	Abnormal telehealth data were directed to PC providers. It is unclear what action this drove. It might have caused the PC provider to direct the patient to an emergency department or a hospital. Could increased symptom surveillance actually increase healthcare utilization?
TEHAF (Boyne Eur J Heart Fail 2012)	N = 382 (197 on RPM)	Electronic device to assess symptoms and educate patients with HF. Abnormal symptoms directed to a monitoring nurse. Device tailored itself to patient’s knowledge.	Excellent adherence with use of the device. Planned and unplanned face‐to‐face HF nurse visits were higher in the control group. Event rates for both groups were lower than expected. Primary limitation appeared to be the excellent outcomes in the control group.
LINK‐HF (Stehlik, CIrc HF 2020)	N=100	Disposable multisensor chest patch for 3 months linked via smartphone to cloud analytics. Apply machine‐learning algorithm.	Pilot study, compliance eroded. However, this detected precursors of hospitalization for HF exacerbation with 76% to 88% sensitivity and 85% specificity.

Abbreviations: BP, blood pressure; HF, heart failure; HR, heart rate; PC, primary care; RPM, remote patient monitoring.