Clinical validation of a point-of-care device versus standard monitoring of vital signs in patients undergoing hematopoietic cell transplantation

Abstract

Abstract Introduction: Alterations in vital signs often precede clinical deterioration, allowing for timely intervention. This is particularly crucial in hematopoietic cell transplantation (HCT) care, where early and appropriate action can improve patient outcomes. Commercially available wearable and point-of-care sensors are increasingly being explored to enhance patient safety. Objective: To clinically validate point-of-care monitoring in patients undergoing autologous and allogeneic HCT. Methods: This prospective, single-center pilot study enrolled 20 inpatient HCT recipients (including both autologous and allogeneic HCT), aged 18 years or older, with reliable reading capacity. The BioMonitor device (BioZ4Life) intermittently measured heart rate (via ECG), blood pressure (mean arterial pressure—MAP), pulse rate (PR), respiratory rate (RR), oxygen saturation (SpO₂), and temperature, wirelessly transmitting data to a smartphone. Measurements were taken two to three times daily over 48–72 h per week from pre-transplant until discharge or day 30 post-HCT, whichever occurred first. Vital signs were sequentially measured by an investigator using both the BioMonitor and validated standard equipment (gold standard). The primary outcome was agreement between the BioMonitor and the gold standard measurements using the Bland and Altman method to calculate mean bias with 95% limits of agreement (LoA). Prior studies indicated that LoA of less than ±0.5°C for temperature, ±3% for SpO₂, ±5 bpm for HR/PR, ±3 rpm for RR, and ±10 mmHg for blood pressure are considered acceptable. Results: Of the 20 patients, 4 (20%) underwent allogeneic HCT and 16 (80%) autologous HCT, with a median age of 34 years. Diagnoses included myelodysplasia, acute leukemia, Hodgkin lymphoma, non-Hodgkin lymphoma, and gammopathies. Febrile neutropenia occurred in 94% of patients; sepsis was observed in one allogeneic HCT recipient. One allogeneic HCT recipient died due to graft failure; no deaths or relapses were recorded in the autologous group. A total of 1,244 pairs of vital sign measurements were obtained. The mean bias and LoA between the two devices were as follows: temperature -0.37°C (LoA +0.83 to -1.58); PR +0.49 bpm (+46.2 to -45.22); SpO₂ +1.8% (+8.75 to -5.15); MAP -2.16 mmHg (+14.46 to -18.78); RR -0.23 rpm (+4.34 to -4.81); and HR -3.26 bpm (+7.83 to -14.34). Conclusion: Point-of-care sensors have the potential to anticipate clinical deterioration and facilitate home monitoring. Despite their advantages, the LoA for temperature and PR rate were not reproducible, whereas for SpO₂, HR, RR, and MAP, the LoA were borderline acceptable. The fact that all measurement pairs were taken sequentially rather than simultaneously may have affected reproducibility. The measurements of SpO₂, HR, RR, and MAP by the BioMonitor warrant further investigation in the outpatient HCT setting.