User Experience Design for Intensive Care Alarm Management.

Alarm fatigue is a pressing issue in intensive care units. Based on user experience design, including clinical shadowings and feedback loops, we developed a prototype for a redesigned patient monitor: The prototype moves away from today's threshold-based alarm systems. It combines a sleek design with machine learning driven clinical insights to mitigate alarm fatigue.

Unobtrusive Monitoring of Clinical Deterioration in Smart Homes.

Clinical deterioration (CD) is the physiological decompensation that incurs care escalation, protracted hospital stays, or even death. The early warning score (EWS) calculates the occurrence of CD based on five vital signs. However, there are limited reports regarding EWS monitoring in smart home settings. This study aims to design a CD detection system for health monitoring at home (HM@H) that automatically identifies unstable vital signs and alarms the medical emergency team. We conduct a requirement analysis by interviewing experts. We use unified modeling language (UML) diagrams to define the behavioral and structural aspects of HM@H. We developed a prototype using a SQL-based database and Python to calculate the EWS in the front end. A team of five experts assessed the accuracy and validity of the designed system. The requirement analysis for four main users yielded 30 data elements and 10 functions. Three main components of HM@H are the graphical user interface (GUI), the application programming interface (API), and the server. Results show the possibility of using unobtrusive sensors to collect smart home residents' vital signs and calculate their EWS scores in real-time. However, further implementation with real data, for frail elderly and hospital-discharged patients is required.

Alarm fatigue and sleep quality in medical staff-a Polish-Czech-Slovak study on workplace ergonomics.

Background: Alarms are crucial in informing Healthcare Workers (HCWs) about critical patient needs, but unmanaged frequency and noise of alarms can de-sensitize medical staff and compromise patient safety. Alarm fatigue is identified as the major cause of the clinical alarm management problem. It occurs when the medical staff is overwhelmed by the number of clinical alarms. Methods: The survey was conducted online using Google's form-making tools from June to July 2023. There were three parts to the survey used in the study: a socio-demographic metric, the Alarm Fatigue Assessment Questionnaire (AFAQ), and The Pittsburgh Sleep Quality Index (PSQI). A significance level of 0.05 was used in the analysis. Results: The survey included 756 medical professionals from three European countries (Slovakia, the Czech Republic and Poland). The participants in the study were 42 years old on average, and they had 12 years of work experience. 603 out of 756 survey participants had poor sleep quality, 147 had good sleep quality, and 6 did not provide an answer. This study analyzed the alarm fatigue levels of respondents in every country. In the Czech Republic, Poland and Slovakia, a statistically significant association (p = 0.039, p = 0.001, p < 0.001) was found between alarm fatigue and sleep quality in medical staff. Conclusion: Based on our study, alarm fatigue and sleep quality of HCWs are correlated. Therefore, alarm fatigue and sleep hygiene should be monitored.

Evaluating the Construct Validity of the Charité Alarm Fatigue Questionnaire using Confirmatory Factor Analysis.

Background: The Charité Alarm Fatigue Questionnaire (CAFQa) is a 9-item questionnaire that aims to standardize how alarm fatigue in nurses and physicians is measured. We previously hypothesized that it has 2 correlated scales, one on the psychosomatic effects of alarm fatigue and the other on staff's coping strategies in working with alarms. Objective: We aimed to validate the hypothesized structure of the CAFQa and thus underpin the instrument's construct validity. Methods: We conducted 2 independent studies with nurses and physicians from intensive care units in Germany (study 1: n=265; study 2: n=1212). Responses to the questionnaire were analyzed using confirmatory factor analysis with the unweighted least-squares algorithm based on polychoric covariances. Convergent validity was assessed by participants' estimation of their own alarm fatigue and exposure to false alarms as a percentage. Results: In both studies, the χ2 test reached statistical significance (study 1: χ226=44.9; P=.01; study 2: χ226=92.4; P<.001). Other fit indices suggested a good model fit (in both studies: root mean square error of approximation <0.05, standardized root mean squared residual <0.08, relative noncentrality index >0.95, Tucker-Lewis index >0.95, and comparative fit index >0.995). Participants' mean scores correlated moderately with self-reported alarm fatigue (study 1: r=0.45; study 2: r=0.53) and weakly with self-perceived exposure to false alarms (study 1: r=0.3; study 2: r=0.33). Conclusions: The questionnaire measures the construct of alarm fatigue as proposed in our previous study. Researchers and clinicians can rely on the CAFQa to measure the alarm fatigue of nurses and physicians.

Variations in Alarm Burden, Source, and Cause Across Inpatient Units at a Children's Hospital.

BACKGROUND AND OBJECTIVES: Alarms at hospitals are frequent and can lead to alarm fatigue posing patient safety risks. We aimed to describe alarm burden over a 1-year period and explored variations in alarm rates stratified by unit type, alarm source, and cause. METHODS: A retrospective study of inpatient alarm and patient census data at 1 children's hospital from January 1, 2019, to December 31, 2019, including 8 inpatient units: 6 medical/surgical unit (MSU), 1 PICU, and 1 NICU. Rates of alarms per patient day (appd) were calculated overall and by unit type, alarm source, and cause. Poisson regression was used for comparisons. RESULTS: There were 7 934 997 alarms over 84 077 patient days (94.4 appd). Significant differences in alarm rates existed across inpatient unit types (MSU 81.3 appd, PICU 90.7, NICU 117.5). Pulse oximetry (POx) probes were the alarm source with highest rate, followed by cardiorespiratory leads (54.4 appd versus 31). PICU had lowest rate of POx alarms (33.3 appd, MSU 37.6, NICU 92.6), whereas NICU had lowest rate of cardiorespiratory lead alarms (16.2 appd, MSU 40.1, PICU 31.4). Alarms stratified by cause displayed variation across unit types where "low oxygen saturation" had the highest overall rate, followed by "technical" alarms (43.4 appp versus 16.3). ICUs had higher rates of low oxygenation saturation alarms, but lower rates of technical alarms than MSUs. CONCLUSIONS: Clinical alarms are frequent and vary across unit types, sources, and causes. Unit-level alarm rates and frequent alarm sources (eg, POx) should be considered when implementing alarm reduction strategies.

Continuous prediction and clinical alarm management of late-onset sepsis in preterm infants using vital signs from a patient monitor.

BACKGROUND AND OBJECTIVE: Continuous prediction of late-onset sepsis (LOS) could be helpful for improving clinical outcomes in neonatal intensive care units (NICU). This study aimed to develop an artificial intelligence (AI) model for assisting the bedside clinicians in successfully identifying infants at risk for LOS using non-invasive vital signs monitoring. METHODS: In a retrospective study from the NICU of the Máxima Medical Center in Veldhoven, the Netherlands, a total of 492 preterm infants less than 32 weeks gestation were included between July 2016 and December 2018. Data on heart rate (HR), respiratory rate (RR), and oxygen saturation (SpO2) at 1 Hz were extracted from the patient monitor. We developed multiple AI models using 102 extracted features or raw time series to provide hourly LOS risk prediction. Shapley values were used to explain the model. For the best performing model, the effect of different vital signs and also the input type of signals on model performance was tested. To further assess the performance of applying the best performing model in a real-world clinical setting, we performed a simulation using four different alarm policies on continuous real-time predictions starting from three days after birth. RESULTS: A total of 51 LOS patients and 68 controls were finally included according to the patient inclusion and exclusion criteria. When tested by seven-fold cross-validations, the mean (standard deviation) area under the receiver operating characteristic curve (AUC) six hours before CRASH was 0.875 (0.072) for the best performing model, compared to the other six models with AUC ranging from 0.782 (0.089) to 0.846 (0.083). The best performing model performed only slightly worse than the model learning from raw physiological waveforms (0.886 [0.068]), successfully detecting 96.1 % of LOS patients before CRASH. When setting the expected alarm window to 24 h and using a multi-threshold alarm policy, the sensitivity metric was 71.6 %, while the positive predictive value was 9.9 %, resulting in an average of 1.15 alarms per day per patient. CONCLUSIONS: The proposed AI model, which learns from routinely collected vital signs, has the potential to assist clinicians in the early detection of LOS. Combined with interpretability and clinical alarm management, this model could be better translated into medical practice for future clinical implementation.

Applying Socio-Technical Models to Alarm Management: Tailoring Bed Exit Alerts in Medical-Surgical Units.

Integration of smartphone technology with the patient call-bell system provides the opportunity to enhance patient safety by supporting nurses' ability to communicate and prioritize care delivery directly. However, challenges are associated with achieving a balance between alarm support and alarm fatigue, including distracting nurses from patient care or desensitizing the nurse to other alarms and calls [1]. Our hospitals have quantitative and anecdotal reports of seriously high volumes of wireless alerts on the nurses' smartphones. Nurses have complained that the phones are generating too much noise to consume or timely prioritize. Preliminary alarm inventory revealed the Bed Exit wireless alert as a leading contributor of signal volume across many units and hospitals. The lack of standard policies and workflow improvement processes has increased nuisance alarms, making these Health Information Technologies less useful and safe. Using system data, workflow observations, and nursing interviews, Singh and Sittig's HIT Safety framework [2] was applied to identify and prioritize sociotechnical factors and interventions that impact the end-to-end Bed Exit alarm workflow. This study reviews the application of sociotechnical models and frameworks to reduce wireless calls without introducing risk and impacting patient care.

Impact in real-world of intermittent-scanned continuous glucose monitoring with alarms on hypoglycemia and its recognition in type 1 diabetes.

BACKGROUND: Recent studies have demonstrated that real-time CGM use reduce the incidence severe hypoglycemic events and impaired awareness of hypoglycemia (IAH) However, there are few real-world studies evaluating the effect of intermittently scanned continuous glucose monitoring (isCGM) on hypoglycemic episodes and hypoglycemia unawareness (IAH). The present study was designed to cover this research-practice gap. METHODS: This is a real-world, observational, prospective cohort study with 2 years of follow-up in which 60 subjects with T1D who experienced frequent hypoglycemic events were included. All the patients were invited to use isCGM type Abbott FreeStyle Libre 2® on a continuous basis for 2 years. Glucometric parameters were obtained during the initial 2 weeks using isCGM and compared with data collected for the same period at 1 year and at the end of follow-up. The IAH was evaluated using the Clarke questionnaire, and to assess psychological aspects related to hypoglycemia the Hypoglycemia Fear Survey (HFS) was used. RESULTS: After 2-years of follow-up using isCGM, we observed a decrease in glucose variability (40.3 ± 0.8 % vs. 37.1 ± 0.9 %, p = 0.003), time in low glucose range (54-69 mg/dL) (5.2 ± 0.4 % vs. 3.6 ± 0.3 %, p = 0.001), time in very low glucose range (<54 mg/dL) (3.2 ± 0.5 % vs. 0.8 ± 0.2 %, p < 0.001), less events related to low glucose levels (10.6 ± 1.1 vs 8.0 ± 1.0, p = 0.042) and a short duration of hypoglycemia episodes (106.1 ± 5.9 min vs. 85.7 ± 5.7 min, p = 0.008). In addition, participants presented a reduction of 47 % in the frequency of IAH, assessed by the Clarke questionnaire scores (24.6 % vs. 11.6 %, p = 0.034), as well as hypoglycemia fear (77.8 ± 2.4 vs 68.2 ± 2.1, p < 0.001). Furthermore, a reduction in total insulin dose was also observed (0.64 ± 0.30 UI/Kg/day vs 0.56 ± 0.11 UI/Kg/day, p = 0.018). CONCLUSIONS: In the real-world, long-term use of isCGM could reduce both hypoglycemic episodes and IAH in people with T1D.

Utilizing Data and Alarm Champions to Enhance Alarm Management: A Pediatric Quality Improvement Initiative.

BACKGROUND: Nuisance and false alarms distract clinicians from urgent alerts, raising patient safety risks. LOCAL PROBLEM: High alarm rates in a pediatric progressive care unit resulted in experiencing 180-250 alarms per day or 1 alarm every 3 to 4 minutes per clinician. METHODS: Through Plan-Do-Study-Act cycles, environmental, policy, and technology changes were implemented to decrease the average alarms/day/bed and percentage of time in alarm. INTERVENTIONS: Alarm settings tailored to patient needs using features embedded within the patient monitoring system were implemented and monitored with the assistance of alarm champions. RESULTS: The average number of alarms/day/bed decreased from 177.69 to 96.94 over the course of 10 years, a 45.45% reduction. The percentage of time in alarm decreased from 7.52% to 2.83%, a 62.37% reduction. CONCLUSIONS: Arming clinicians with technology to analyze real-time clinical data made alarms meaningful and actionable, decreasing false alarms without compromising patient safety.

Alarm Management in Intensive Care: Qualitative Triangulation Study.

BACKGROUND: The high number of unnecessary alarms in intensive care settings leads to alarm fatigue among staff and threatens patient safety. To develop and implement effective and sustainable solutions for alarm management in intensive care units (ICUs), an understanding of staff interactions with the patient monitoring system and alarm management practices is essential. OBJECTIVE: This study investigated the interaction of nurses and physicians with the patient monitoring system, their perceptions of alarm management, and smart alarm management solutions. METHODS: This explorative qualitative study with an ethnographic, multimethods approach was conducted in an ICU of a German university hospital. Using triangulation in data collection, 102 hours of field observations, 12 semistructured interviews with ICU staff members, and the results of a participatory task were analyzed. The data analysis followed an inductive, grounded theory approach. RESULTS: Nurses and physicians reported interacting with the continuous vital sign monitoring system for most of their work time and tasks. There were no established standards for alarm management; instead, nurses and physicians stated that alarms were addressed through ad hoc reactions, a practice they viewed as problematic. Staff members' perceptions of intelligent alarm management varied, but they highlighted the importance of understandable and traceable suggestions to increase trust and cognitive ease. CONCLUSIONS: Staff members' interactions with the omnipresent patient monitoring system and its alarms are essential parts of ICU workflows and clinical decision-making. Alarm management standards and workflows have been shown to be deficient. Our observations, as well as staff feedback, suggest that changes are warranted. Solutions for alarm management should be designed and implemented with users, workflows, and real-world data at the core.

Short communication: A pilot study to determine safety and feasibility of a commercial enuresis alarm to detect voiding in infants.

Accurate measurement of post-void residual (PVR) volumes requires accurate determination of the timing of voiding, which is challenging in non-verbal patients. As a proof of principle, we sought to test the feasibility, safety and efficacy of using an enuresis alarm to indicate voiding in ten infants. Each infant was observed for 4 h with alarm in the diaper, and diapers checked every 15-30 min to confirm voiding. The alarm activated in 31 of 33 voids (93.9%). No adverse events occurred. Further work will investigate whether this approach may improve accuracy of PVR measurement.

Fewer emergency department alarms is associated with reduced use of medications for acute agitation.

BACKGROUND AND OBJECTIVES: Patient monitoring systems provide critical information but often produce loud, frequent alarms that worsen patient agitation and stress. This may increase the use of physical and chemical restraints with implications for patient morbidity and autonomy. This study analyzes how augmenting alarm thresholds affects the proportion of alarm-free time and the frequency of medications administered to treat acute agitation. METHODS: Our emergency department's patient monitoring system was modified on June 28, 2022 to increase the tachycardia alarm threshold from 130 to 150 and to remove alarm sounds for several arrhythmias, including bigeminy and premature ventricular beats. A pre-post study was performed lasting 55 days before and 55 days after this intervention. The primary outcome was change in number of daily patient alarms. The secondary outcomes were alarm-free time per day and median number of antipsychotic and benzodiazepine medications administered per day. The safety outcome was the median number of patients transferred daily to the resuscitation area. We used quantile regression to compare outcomes between the pre- and post-intervention period and linear regression to correlate alarm-free time with the number of sedating medications administered. RESULTS: Between the pre- and post-intervention period, the median number of alarms per day decreased from 1332 to 845 (-37%). This was primarily driven by reduced low-priority arrhythmia alarms from 262 to 21 (-92%), while the median daily census was unchanged (33 vs 32). Median hours per day free from alarms increased from 1.0 to 2.4 (difference 1.4, 95% CI 0.8-2.1). The median number of sedating medications administered per day decreased from 14 to 10 (difference - 4, 95% CI -1 to -7) while the number of escalations in level of care to our resuscitation care area did not change significantly. Multivariable linear regression showed a 60-min increase of alarm-free time per day was associated with 0.8 (95% CI 0.1-1.4) fewer administrations of sedating medication while an additional patient on the behavioral health census was associated with 0.5 (95% CI 0.0-1.1) more administrations of sedating medication. CONCLUSION: A reasonable change in alarm parameter settings may increase the time patients and healthcare workers spend in the emergency department without alarm noise, which in this study was associated with fewer doses of sedating medications administered.

Machine Alarm Fatigue among Hemodialysis Nurses in 29 Tertiary Hospitals.

OBJECTIVES: To understand the status quo and related influencing factors of machine alarm fatigue of hemodialysis nurses in tertiary hospitals in Liaoning Province. METHODS: This cross-sectional study employed convenience sampling to select 460 nurses from 29 tertiary hospitals in Liaoning Province, who are involved in hemodialysis care. Surveys were conducted using the General Information Questionnaire, Alarm Fatigue Scale, National Aeronautics and Space Administration Task Load Index, and Maslach Burnout Inventory Scale. RESULTS: The overall machine alarm fatigue score for 460 hemodialysis nurses from 29 tertiary hospitals in Liaoning Province was 17.04 ± 3.21, indicating a moderate level. The multiple linear regression analysis shows that years of experience in hemodialysis nursing, the number of patients managed per shift, whether specialized nursing training has been received, self-reported health status, emotional exhaustion, and workload have statistically significant associations with alarm fatigue among hemodialysis nurses (p < 0.05). Among them, the years of experience in hemodialysis nursing are negatively correlated with alarm fatigue among hemodialysis nurses, whereas the number of patients managed per shift and workload are positively correlated with alarm fatigue among hemodialysis nurses. CONCLUSION: This study indicates that certain demographic factors, workload, and occupational burnout are associated with machine alarm fatigue among hemodialysis nurses. Therefore, hemodialysis-related managers should establish a Machine Alarm Management System, implement Personalized Thresholds and Delayed Alarms, ensure reasonable staffing arrangements, improve compassion fatigue, and enhance anticipatory care. Our findings have implications for improving the health and well-being of hemodialysis nurses, providing a conducive environment for professional training in hemodialysis, and ultimately addressing the current situation of machine alarm fatigue among hemodialysis nurses.

Effect of switch from flash glucose monitoring to flash glucose monitoring with real-time alarms on hypoglycaemia in people with type 1 diabetes mellitus.

We aimed to evaluate the utility of the FreeStyle Libre 2 device for reducing time below range level 1 and level 2 compared with the Freestyle Libre device (without alarms) in people with type 1 diabetes mellitus. We conducted longitudinal observational follow-up study of a cohort of 100 people with type 1 diabetes mellitus who had switched from FreeStyle Libre to FreeStyle Libre 2 as part of routine clinical practice. Three months after switching to FreeStyle Libre 2, compared with results with FreeStyle Libre, there were a significant improvements in time below range level 1 (p = 0.02) and level 2 (p <0.001), time in range (p <0.001), time above range level 1 (p = 0.002), glucose management indicator (p= 0.04) and mean glucose (p= 0.04) during follow-up. Furthermore there was a significant direct association between age and change in TIR with a coefficient of 0.23, and a significant inverse association between age and change in TAR-1 with a coefficient of 0.11. Switching to a flash glucose monitoring system with alarms improves time below range, time in range and coefficient of variation in people with type 1 diabetes mellitus.

The introduction of a sound reduction bundle in the intensive care unit and its impact on sound levels and patients.

BACKGROUND: ICU outcomes are continuing to improve. However, this has not been matched by similar improvements of the ICU bedspace environment, which can detrimentally impact on patient outcomes. Excessive sound and noise, especially, has been linked with adverse and potentially preventable patient outcomes and staff errors. There are many sources of sound in the ICU, with alarms from bedside equipment frequently listed as a main source. The number of alarms is increasing in parallel with the introduction of new and more sophisticated technologies to monitor and support patients. However, most alarms are not accurate or critical and are commonly ignored by staff. OBJECTIVE: The objective of this study was to evaluate the impact of a sound reduction bundle on sound levels, number of alarms, and patients' experience and perceived quality of sleep in the ICU. METHODS: This was a pre-post, quasi-experimental study investigating the impact of three study interventions implemented sequentially (staff education, visual warnings when sound levels exceeded the preset levels, and monitor alarm reconfigurations). Effects of staff education were evaluated using pre-education and post-education questionnaires, and the impact on patients was evaluated via self-report questionnaires. A sound-level monitor was used to evaluate changes in sound levels between interventions. Alarm audits were completed before and after alarm reconfiguration. RESULTS: Staff knowledge improved; however, sound levels did not change across interventions. The number of monthly monitor alarms reduced from 600,452 to 115,927. No significant differences were found in patients' subjective rating of their experience and sleep. CONCLUSION: The interventions did not lead to a sound-level reduction; however, there was a large reduction in ICU monitor alarms without any alarm-related adverse events. As the sources of sound are diverse, multidimensional interventions, including staff education, alarm management solutions, and environmental redesign, are likely to be required to achieve a relevant, lasting, and significant sound reduction.

IoT-based incubator monitoring and machine learning powered alarm predictions.

BACKGROUND: Incubators, especially the ones for babies, require continuous monitoring for anomaly detection and taking action when necessary. OBJECTIVE: This study aims to introduce a system in which important information such as temperature, humidity and gas values being tracked from incubator environment continuously in real-time. METHOD: Multiple sensors, a microcontroller, a transmission module, a cloud server, a mobile application, and a Web application were integrated Data were made accessible to the duty personnel both remotely via Wi-Fi and in the range of the sensors via Bluetooth Low Energy technologies. In addition, potential emergencies were detected and alarm notifications were created utilising a machine learning algorithm. The mobile application receiving the data from the sensors via Bluetooth was designed such a way that it stores the data internally in case of Internet disruption, and transfers the data when the connection is restored. RESULTS: The obtained results reveal that a neural network structure with sensor measurements from the last hour gives the best prediction for the next hour measurement. CONCLUSION: The affordable hardware and software used in this system make it beneficial, especially in the health sector, in which the close monitoring of baby incubators is vitally important.

Critical Care Alarm Fatigue and Monitor Customization: Alarm Frequencies and Context Factors.

BACKGROUND: Alarm fatigue among nurses working in the intensive care unit has garnered considerable attention as a national patient safety priority. A viable solution for reducing the frequency of alarms and unnecessary noise is intensive care unit alarm monitor customization. LOCAL PROBLEM: A 24-bed cardiovascular and thoracic surgery intensive care unit in a large academic medical center identified a high rate of alarms and associated noise as a problem contributing to nurse alarm fatigue. METHODS: An alarm monitor quality improvement project used both alarm frequency and nurse surveys before and after implementation to determine the effectiveness of interventions. Multimodal interventions included nurse training sessions, informational flyers, organizational policies, and an alarm monitor training video. Unexpected results inspired an extensive investigation and secondary analysis, which included examining the data-capturing capabilities of the alarm monitors and the impact of context factors. RESULTS: Alarm frequencies unexpectedly increased after the intervention. The software data-capturing features of the alarm monitors for determining frequency did not accurately measure nurse interactions with monitors. Measured increases in patient census, nurse staffing, and data input from medical devices from before to after the intervention substantially affected project results. CONCLUSIONS: Alarm frequencies proved an unreliable measure of nurse skills and practices in alarm customization. Documented changes in context factors provided strong anecdotal evidence of changed circumstances that clarified project results and underscored the critical importance of contemporaneous collection of context data. Designs and methods used in quality improvement projects must include reliable outcome measures to achieve meaningful results.

Learning From Alarms: A Robust Learning Approach for Accurate Photoplethysmography-Based Atrial Fibrillation Detection Using Eight Million Samples Labeled With Imprecise Arrhythmia Alarms.

Atrial fibrillation (AF) is a common cardiac arrhythmia with serious health consequences if not detected and treated early. Detecting AF using wearable devices with photoplethysmography (PPG) sensors and deep neural networks has demonstrated some success using proprietary algorithms in commercial solutions. However, to improve continuous AF detection in ambulatory settings towards a population-wide screening use case, we face several challenges, one of which is the lack of large-scale labeled training data. To address this challenge, we propose to leverage AF alarms from bedside patient monitors to label concurrent PPG signals, resulting in the largest PPG-AF dataset so far (8.5 M 30-second records from 24,100 patients) and demonstrating a practical approach to build large labeled PPG datasets. Furthermore, we recognize that the AF labels thus obtained contain errors because of false AF alarms generated from imperfect built-in algorithms from bedside monitors. Dealing with label noise with unknown distribution characteristics in this case requires advanced algorithms. We, therefore, introduce and open-source a novel loss design, the cluster membership consistency (CMC) loss, to mitigate label errors. By comparing CMC with state-of-the-art methods selected from a noisy label competition, we demonstrate its superiority in handling label noise in PPG data, resilience to poor-quality signals, and computational efficiency.