Continuous Vital Signs Monitoring in Patients Hospitalized at Home: Burden or Benefit? (2025)

To the Editor

Hospital-at-Home (HaH) programs are increasingly popular alternatives to inpatient hospitalization,1, 2, 3 proving to be a tenable model of care especially during the COVID-19 pandemic.4^,5 They demonstrated benefits of shorter lengths of stay,2 improved patient experiences,2 and reduced costs.6 Wearable monitors allow continuous wireless monitoring of acutely ill patients at home, but whether this increases the safety of HaH care is uncertain. Remote vital sign monitoring has been used in HaH programs,6 but this study is the first to describe its use and impact on a HaH service.

NUHS@Home, a HaH program of a tertiary acute hospital in Western Singapore, was piloted with 34 patients (Table1) from September 2020 to January 2021. Physicians conducted home visits in person at least once daily, with nurses visiting up to 3 times a day for intravenous therapy. Patients were discharged from the program when they met conventional inpatient discharge criteria.

Remote vitals monitoring platform (ISANSYS Patient Status Engine,7^,8) consisting of a 4G-enabled Samsung tablet connected by Bluetooth to a blood pressure, pulse oximeter, and a 2-lead electrocardiogram patch (Lifetouch). The patches allowed continuous heart rate and respiratory rate monitoring. Blood pressure and pulse oximetry was self-measured at a prescribed frequency.

When pre-set vital signs thresholds were breached, a SMS alert was sent to the doctor on duty. Subsequent action was left to the clinician's discretion. Vital signs and SMS alert data were subsequently collected. A patient experience survey was administered.

Heart and respiratory rate readings were successfully transmitted for 86.4%±15.9% of the duration of monitoring. Nontransmission may be due to patients removing the patch, interference from physical activity, or poor signal. Compliance of patients to blood pressure and oxygen monitoring was 97.5% and 81%, respectively.

In total, 248 alerts were received in 34 patients over 163 patient days (Table1), an average of 1.52 alerts per patient day. Note that 65.3% of alerts were from the wearable patch. Of all alerts, only 8 readings across 4 patient episodes were clinically significant; others were erroneous readings or self-resolved. Among these 4 episodes, 1 was associated with anaphylaxis in attendance of a nurse, 2 episodes were picked up by self-measured blood pressure or oximetry monitoring; only 1 episode was picked up by the continuous patch (for high respiratory rate) in absence of other signs and symptoms–the patient was transferred back to the hospital for worsening cellulitis.

All patients felt reassured by continuous heart rate monitoring, and no deteriorations were missed by this system. Most (95%) agreed that the patch for continuous heat rate monitoring was comfortable, and 76.1% agreed it was important. A sensitivity analysis showed that increasing heart rate threshold to 110 would result in a 11.1% reduction in alerts and no significant alerts missed; increasing respiratory rate threshold to 30 would reduce alerts by 24% but miss 1 significant alert; if SpO₂ threshold was reduced from 95% to 94%, alerts would reduce by 37.5% with no significant alerts missed.

Our findings suggest that remote vitals sign monitoring systems can be an integral component of HaH care, with several key considerations. First, the use of remote vital signs monitoring systems is feasible for use in the HaH setting as all our patients were able to engage with it. Second, continuous monitoring creates a high alert burden with low rates of deterioration detection. All deterioration episodes in our study would have been picked up by a staff member or by means of patient symptoms. Third, intermittent self-measured vital signs have good compliance rates. Finally, threshold-based automated alerts, instead of having a dedicated staff member to supervise a dashboard round-the-clock, can be used to maximize efficiency of HaH on-calls.

Excessive alerts were also reported in a trial of continuous monitoring in a surgical ward, and threshold adjustment was eventually required to cope.9 In home-based care where patients are more active than in a hospital ward, the rate of alerts from a continuous monitoring patch is even higher. We recommend that instead of being routinely deployed, continuous monitoring may be considered only for higher risk patients with a higher pre-test probability for deterioration. Liberalizing alert thresholds for continuous monitoring of ambulant patients or application of individualized thresholds per patient,10 may be helpful. Finally, remote monitoring should be considered as an adjunct but not replace patient education on warning signs and symptoms and regular reviews by the clinical care team.

Lessons learned in this study may guide vital signs monitoring for similar HaH programs. Further studies are warranted to risk-stratify which patients may benefit from continuous monitoring and validate early warning scores in the ambulatory setting that corresponds with clinical deterioration.