Continuous BP monitoring of ICU Taiwanese patients: training-phase evaluation of oCare™ BP100 according to ISO 81060-3: 2022

1 Introduction

Hypertension affects approximately more than 1 billion adults worldwide. Hypertensive patients undergo a healthy lifestyle and hypertensive drug use is based on hypertension device measurements (1). Traditional blood pressure measurement technologies, such as office blood pressure (OBP) and home blood pressure measurement (HBP), are limited by the time of operation and can only provide fragmented information, unable to automatically grasp day and night changes in blood pressure. Since the dynamic change in blood pressure is a challenge, the snapshot measurements provided by the current OBP or HBP are insufficient (2). Although cuff-type 24 h blood pressure monitoring can provide blood pressure data at some time points, its measurement intervals are long and interfere greatly with the user's activities and sleep. It is not suitable for long-term or real-time physiological monitoring and cannot be effectively applied to clinical needs such as pre/postoperative blood pressure monitoring, sudden blood pressure abnormalities during sleep, and detecting of blood pressure changes related to sleep disorders.

A plethysmograph is an instrument that measures changes in volume within an organ or the whole body. A photoplethysmogram (PPG) is an optically obtained plethysmogram that can be used to detect changes in blood volume in the microvascular bed of tissue. This technology had been applied in continuous blood pressure monitoring (CBPM) (3, 4), but the clinical validation of most PPG-related CBPM lack of comparison with an appropriate gold standard, such as a sphygmomanometer or invasive arterial blood pressure (5). Even the validation process was done according the state of the art (6, 7), the accuracy was mainly focused on intermittent blood pressure measurements rather than continuous blood pressure measurements (8, 9). Not to mention the issue of confirming the change test of continuous blood pressure between the CBPM and gold standard.

The investigational medical device under evaluation is oCare^™ BP100, a prototype of non-invasive continuous blood pressure monitoring system developed by Taiwan Biophotonic Co. (tBPC). The main function is to use a finger-clip sensor to receive photoplethysmography (PPG) signals and to collect continuous blood pressure values uninterruptedly within the initialization time. By analyzing the PPG waveform characteristics and combining them with algorithms, we can achieve the ability to output blood pressure values every 30 s to meet the clinical demand for “continuous and accurate” blood pressure monitoring.

Currently, commercially available devices that are tested and validated according to the ISO 81060-3 standard only partially comply with subject recruitment and testing procedures (10, 11). In this study, according to the latest published ISO 81060-3:2022 standard requirements, non-invasive blood pressure measurements by cuff of “Philips” Patient Monitor (Model: IntelliVue MX800 & X2) at start time points were used as oCare^™ BP100 blood pressure calibration. Invasive arterial catheter blood pressure measurements of the A-line were used as the reference invasive BPM equipment (“Philips” Patient Monitor, Model: IntelliVue MX800 & X2). The objectives of the study were to confirm whether the stability of oCare^™ BP100 within the ISO target thresholds of calibration standard deviation ≤6 mmHg per analytical cycle; and whether the performance of oCare^™ BP100 in changes of blood pressure was consistent with reference measurements.

2 Material and methods

2.1 Cuffless blood pressure device

A prototype cuffless device developed by Taiwan Biophotonic Corporation (tBPC), oCare^™ is a continuous blood pressure monitor made by optical detection technology. [Software Version:V3.000; Firmware Version: 3.025 (r0)]. The shell of the oCare^™ BP100 sensor is in contact with the subject's finger, and the material is biomedical resin. The finger bed of the oCare^™ BP100 sensor is in contact with the subject's finger and the material is high-temperature vulcanized silicone. The oCare^™ BP100 watch strap is in contact with the wrist or hand of the subject, and the material is silicon-based thermoplastic vulcanizate. The oCare^™ BP100 case is in contact with the subject's wrist or hand and is made of polycarbonate. Those who are allergic to the above materials should not use this device. To achieve optical cuffless continuous blood pressure measurement technology, the algorithm developed by tBPC implements an integrated approach utilizing multiple advanced techniques. This algorithm not only analyzes blood pressure through detailed photoplethysmography (PPG) waveform features but also incorporates PPG waveform decomposition technology to significantly enhance measurement accuracy. Furthermore, the system employs multi-wavelength detection technology, effectively reducing external interference factors and ensuring measurement data stability and reliability. The synergistic application of these technologies provides a non-invasive continuous blood pressure monitoring solution with considerable potential for clinical applications (Figures 1, 2).

2.2 Reference blood pressure

The reference BP includes 4 h continuous invasive blood pressure (ABP) monitoring provided by the “Philips” Patient Monitor (Model: IntelliVue MX800 & X2). The same cuff model Philips M1574A or M1575A cuff was only used for Time calibration at both the start and end points of the BP measurements (NIBP) (12). The cuff was attached to the arm opposite line A and removed during the 4-hour measurement period. Attach the NIBP tubing to the NBP connector on the X2 module. On the Philips IntelliVue MX800 and X2, ABP and NIBP are handled through the X2 module, which serves as the acquisition device for the MX800 host (13).

2.3 Recruitment

This study was conducted in the National Taiwan University Hospital (NTUH) Hsin-Chu Branch. A total of 40 intensive care unit (ICU) patients were recruited with the inclusion criteria as follows:

Inclusion Criteria. Exclusion Criteria. Subjects who met the inclusion criteria and signed the informed consent form were formally enrolled in the study protocol. If the measurements were incomplete or the waveform did not meet the normal pattern, the data were invalid for downstream analysis. Common types of poor-quality PPG signals observed in this study include: (1) high-frequency noise, primarily resulting from improper attachment of the fingertip sensor to the patient's finger or from motion-induced vibrations at the detection site during monitoring, leading to significant oscillatory artifacts in the PPG waveform; (2) waveform distortion, often caused by external compression of the sensor, which in turn applies pressure to the measurement site, resulting in abnormal deformation of the waveform morphology; and (3) marked attenuation or near absence of the AC component in the PPG signal, where the pulsatile waveform is significantly reduced or flat, compromising the identification of systolic peaks and other morphological features. Subjects underwent follow-up within 7 days post-trial to evaluate adverse events (AEs). The study concluded thereafter.

• Age ≥ 22 years at the time of informed consent
• Already receiving radial artery cannulation (A-line) for blood pressure monitoring based on clinical indication
• Ability to understand the written informed consent form and willingness to participate, as indicated by signed informed consent from the patient or legal representative

• Pregnancy
• Use of intra-aortic balloon pump
• Presence of arm or finger injuries
• History of vascular surgery involving the upper extremities or contraindications to local treatment in the upper extremities
• Arterial line placement in the brachial artery
• Determined unsuitable for study participation based on investigator judgment (e.g., poor compliance, impaired mental status, unstable health condition)
• Refusal by the participant or legal representative to sign the informed consent form

2.4 Procedure

• The researcher confirms that the subject has the A-line embedding needle for radial artery blood drawing, and the “Philips” patient monitor (Model: IntelliVue MX800 & X2) is on (named “A” measurements). The side with A-line referred to as side A.
• Turn on the notebook computer that can receive two units of oCare™ BP100.
• For Side A, the researcher assists the subject in wearing one unit of oCare™ BP100 (named “T1” measurements), turns it on, and confirms that there are signals on the platform.
• Attach the blood pressure cuff of “Philips” Patient Monitor to the arm on the other side B, and connect to the NIBP slot of “Philips” Patient Monitor (Model: IntelliVue MX800 & X2) (named “C” measurements).
• For Side B, the researcher assists the subject in wearing another unit of oCare™ BP100 (named “T2” measurement), turns on it and confirms that there are measurement values on the platform.
• Start the cuff-type blood pressure measurements on the arm of side B. This is the starting point of the test.
• Start the 4 h measurement on the reference medical device and investigational medical device.
• Start the cuff-type blood pressure measurements on the arm of side B. This is the end point of the test.
• Stop measurements, and shut down the two investigational medical devices. And remove the two investigational medical devices and the cuff of the “Philips” Patient Monitor.
• The researchers confirm the storage and recording of measured values on the platform.
• End of the test.
• Physicians continued patient care in accordance with the pre-established treatment plan.
• Researchers conducted follow-up visits with subjects within seven days following the completion of monitoring.

2.5 Ethics

The protocol has been approved by the 175th meeting of Research Ethics Committee B of the National Taiwan University Hospital on Oct 20, 2023. The committee is organized under, and operates in accordance with, the Good Clinical Practice guidelines and governmental laws and regulations. The duration of this approval is one year (from Oct 20, 2023 to Oct 19, 2024). NTUH-REC No.: 202309072DSB.

2.6 Trial registration

The project has been registered in ClinicalTrials.gov (ID: NCT06815614) https://clinicaltrials.gov/study/NCT06815614.

2.7 Algorithm

The device utilizes green, red, and infrared wavelengths at a 200 Hz sampling rate. This multiwavelength approach enhances signal quality against variations in skin tone, motion, and perfusion by capturing surface features (green) and deep vascular data (red/IR); specifically, certain IR bands assist in calibration due to stable hemoglobin absorption. The algorithm applies band-pass filtering and leverages accelerometer (ACC) data to eliminate motion-induced noise from PPG signals. Calibration is a single-point procedure that maps 30 s of extracted PPG features to NIBP measurements. Artifact detection relies on ACC-derived motion thresholds and a 90% similarity requirement between adjacent PPG pulses. If valid pulses comprise less than 60% of the 30 s window, the blood pressure calculation is suspended. For re-initialization, the system continuously scans and resumes calculation once PPG features meet these quality standards within a subsequent 30-second period. This framework ensures technical transparency and adherence to ISO 81060-3 standards in complex environments.

3 Standard requirements and methods

All statistical analyses were performed in accordance with the ISO 81060-3:2022 standard for clinical investigation of continuous automated non-invasive sphygmomanometers. And, all analyses were implemented in SAS 9.4+. This analysis included both stability [Clause 5.2 in (10)] and blood pressure changes [Clause 5.3 in (10)], and complied with the detailed procedures defined in Annex D in (10) (13) (Tables 1A, 1B).

• Stability (Clause 5.2 in (10)):
• Blood Pressure Changes [Clause 5.3 in (10)]:

Table 1A. Notations for stability test.

Table 1B. Notations for blood pressure change test.

then calculate the absolute error, Epercent, according to Formula [3]. ΔP_ref: the change in BP for the reference sphygmomanometer.

Epercent=|ΔPsut−ΔPref|max(|ΔPsut|,|ΔPref|)×100%(3)

ΔP_sut: the change in BP for the oCare™ BP100.

• 2.2.7. After pooling all Epercent values of each subject, calculate the 50th percentile of all the Epercent values per subject, E50th.
• 2.2.8. After pooling all Epercent values of each subject, calculate the 85th percentile of all the Epercent values per subject, E85th.
• 2.2.9. Averaging the calculated 50th overall subjects, obtain E¯50th. Averaging the calculated 85th overall subjects, obtain E¯85th.

• 2.3 Confirm the following endpoints.

The ISO target thresholds for the Stability Test is defined as (A): the corrected standard deviation s_corr ≤ 6 mmHg. For the Blood Pressure Changes, the criteria are (B1): the average 50th percentile of relative errors ≤25%, and (B2): the average 85th percentile of relative errors ≤50%.

4 Results

The results were evaluated according to the ISO target thresholds defined in the Methods section: (A) for stability, and (B1) and (B2) for blood pressure changes.

4.1 Participant selection, general characteristics and blood pressure distribution

During 2023/10/01∼2024/09/30, this study was conducted in the NTUH Hsin-Chu Branch. A total of 40 ICU participants were initially selected and included in the training phase according to ISO 81060-3:2022, including 29 men and 11 women participants, who had a mean age of 67.2 ± 12.6 years, with a range of 45–96 years (Table 2).

Table 2. Demographic characteristic of subjects.

4.2 Stability test

Since the initialization period was set 1.5–5 h, there were 6 analysis periods were required for each subject. There must be at least 15 min between two analysis periods. Ideal (pair) readings are required 29 readings in each analysis period, subjects without ≥29 readings were invalid in the calculation for the selected segment. Table 3 lists the valid pair readings and Subject Number used for each analysis period.

Table 3. Number of readings and sample size of subjects used in stability test.

Table 4 shows that for the corrected standard deviations of SBP, DBP, and MAP parameter within the ISO target thresholds (A) s_corr ≤ 6 mmHg for each analysis period. The corrected standard deviation listed for Period A–F:

• SBP: 3.15, 3.83, 3.67, 3.99, 4.13, 4.29;
• DBP: 1.84, 2.35, 2.25, 2.23, 2.85, 2.65;
• MAP: 2.17, 2.54, 2.48, 2.69, 2,81, 3.12.

Table 4. Stability test analysis.

Bold values indicate results that comply with the ISO target thresholds (A) (S_corr ≤ 6 mmHg).

The above values represent that repeated measurement of SBP, DBP, and MAP of the investigational medical device are stable, even the gender ratio variation and the number drop-off from Period A to Period F. The corrected experimental standard deviation S_corr, are related to the total number of participants and the number of paired measurements. In Table 4, the fewer the participants, the higher the S_corr.

4.3 Blood pressure changes

For each blood pressure parameter to be investigated, the criteria (B1) and (B2) are analyzed. For example, the valid reading of blood pressure in SBP, DBP, and MAP was demonstrated in the circle symbols for Subject (ID: BP15), (Figure 3). For each subject, the valid periods varied due to SBP, DBP, and MAP. Invalid BP changes were excluded from the BP Change test analysis. Table 5 lists the analysis period and the invalid number of readings used in the blood pressure change test. Table 6 presents the E¯50th of SBP, DBP, and MAP are 11.40%, 9.93%, and 10.08%; and E¯85th of SBP, DBP, and MAP are 27.81%, 24.84%, and 26.31%. The ISO target thresholds for the monitoring of Blood Pressure Change Test (B1) E¯50th≤25% and (B2) E¯85th≤50% are fulfilled, the secondary hypothesis is established: the efficacy of the investigational medical device and the reference medical device in blood pressure change tests are consistent. The number of participants in the blood pressure change test was 22, 15, and 16 for SBP, DBP, and MAP, respectively. In generally, DBP changes are relatively mild and usually remain within a smaller range. In this test, its fluctuation range is inherently smaller than SBP and MAP. Therefore, the percentile of absolute errors E_50th and E_85th were directly related to the absolute error of change difference between oCare™ BP100 and the reference device in hemodynamic measurements of systolic and diastolic blood pressure. Gender distribution is not a direct influencing factor for Stability Test nor BP Change Test.

Table 5. Invalid BP change reading pairs for all subjects (%).

Table 6. BP change analysis for E₅₀ and E₈₅ of all subjects (%).

4.4 AE/serious adverse event (SAE)

All subjects were tracked within 7 days, none of AE/SAE was reported during the follow-up within 7 days. Two subjects died from pneumonia following study completion. These SAEs were reported and assessed as unrelated to the investigational device or the trial.

5 Discussion

In current published studies, rare commercial devices claim to meet the criteria of ISO 81060-3 (10). The comparisons of the proposed prototype oCare™ BP100 and related devices were detailed in Table 7. Hove et al. (10) presented that the DBP of healthy participants achieving the criteria of BP change test but not in SBP which were higher than the limit for systolic blood pressure (56% vs. ≤50%) and for all parameters for the 50th percentile (32%–39% vs. ≤25%). Furthermore, the importance of the study design, as specified in ISO 81060-3, is that an invasive reference is required, i.e., an invasive blood pressure monitoring device is inserted in all subjects, which was not present in their recruited healthy participants (10). Boretsky et al. (11) conducted a pediatric study to validate Vitalstream™, which is a medical device approved by the United States Food and Drug Administration for use in adults. Since the pediatric trial were addressed in children ages 2–17 undergoing major surgeries during hemodynamically stable periods, the validation of Blood Pressure Changes was lacking and the Vitalstream™ in Type A or Type T mode blood pressure monitoring was unclear. Khayat et al. (14) briefed the ambulatory BP monitoring (ABPM) passing the BP change test, even only 27 participants used which not compliance the requirement for ≥30 subjects. Detailed comparison in Table 7.

Table 7. Checklist for ISO 81060-3:2022 standard requirements.

This study is the first PPG training-phase evaluation in the Type T protocol of ISO 81060-3:2022 using invasive reference measurements in a complex ICU environment. The investigational device, oCare™ BP 100, showed promising ability to meet the stability and blood pressure change criteria in a high-acuity setting. The high-frequency, every-30-second BP output meets the need for near-continuous monitoring and may contribute to improved patient safety and outcome detecting. Besides, its use of non-invasive finger-clip PPG sensing may enhance patient comfort and facilitate continuous data acquisition with minimal disruption to clinical procedures.

However, the study has some limitations. While the age distribution strictly complies with the requirements of ISO 81060-3:2022, the gender distribution did not fully comply (female participants accounted for only 27.5%, slightly below the 30% requirement). In addition, due to the prototype stage of the oCare™ BP100 sensor, readings with fewer than 29 measurements per analysis period were excluded from the Stability Test. The number of participants in the blood pressure change test was 22, 15, and 16 for SBP, DBP, and MAP, respectively. This exploratory result suggests that the change-tracking cohort falls below typical participant requirements. This selection bias or survivorship bias might limit generalizability to the most unstable ICU patients.

In future studies, further validation in accordance with the full ISO 81060-3:2022 protocol—including expanded subject demographics and testing across varied clinical scenarios—is necessary. Future evaluations involving patients with cardiovascular instability, during surgery, or in ambulatory settings will be essential to confirm the device's broader applicability and to support regulatory submission and clinical adoption.

6 Conclusion

This training-phase study demonstrated that the oCare™ BP100, a PPG-based non-invasive continuous blood pressure monitor, within the required ISO thresholds for stability and blood pressure changes under ISO 81060-3:2022 Type T mode in an ICU population. Stability Test were demonstrated with the correlated standard deviation less than 6 mmHg in SBP, DBP, and MAP for each analysis period. And, the Blood Pressure Changes Test was shown to be consistent with oCare™ BP100 and the reference device (“Philips” Patient Monitor, Model: IntelliVue MX800 & X2). These results support the potential of oCare™ BP100 to provide continuous and accurate hemodynamic monitoring in critical care environments. Furthermore, oCare™ BP100 needs to be validated in the downstream testing-phase per ISO 81060-3:2022 Type T mode before it can be launched on the market.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Ahsan H. Khandoker, Khalifa University, United Arab Emirates

Reviewed by: Mahmudah Salwa Gianti, Indonesia University of Education, Indonesia

Huijin Lee, Seoul National University Hospital, Republic of Korea