1,720,974 research outputs found
FITBEAT: Towards a cardiopulmonary fitness level index for personalized heart failure management
No full textHeart failure is a chronic disease with increasing prevalence, characterized by high morbidity
and mortality rates and accompanied by frequent (re)hospitalizations. It is a complex
disease that cannot be comprised into a single parameter, making disease follow-up
challenging. Regular follow-up and monitoring of hallmark symptoms are beneficial for HF
patients to follow-up disease progression and to predict disease worsening early. A novel
digital cardiac biomarker, the FitBeat index, recorded by non-invasive remote monitoring
technologies could make this happen.
In this thesis, the initial building blocks towards a cardiopulmonary fitness level index,
comprising the complexity of the disease, were created. First, new insights into the
effectiveness of CR were gained. Following CR after CRT implant proved to be cost-effective.
Moreover, responders and non-responders to CR showed different progression patterns in
functional capacity. These insights show the effectiveness of CR as a secondary prevention
measure but also confirm the need for patient-tailored approaches.
The development of a digital biomarker requires the recording of vital signals from which
physiologically relevant features can be extracted. Initially, signal quality is essential to
ensure reliable and proper clinical implementation. We identified the most optimal electrode
configuration for respiration monitoring, using a wearable bioimpedance device. Moreover,
an algorithm that automatically annotates ECG recordings, validated in a cardiac patient
population, applicable in a wearable low power device was developed.
Physiological features were extracted from the preprocessed signals and their progression
throughout CR was studied. In this thesis, the ECG-derived features played a major role.
Cardiac-related medication intake influenced the progression of HRV parameters throughout
CR, emphasizing the importance of taking confounding factors into account when
interpreting cardiac-related features. The non-responders and responders to CR were
characterized by a different cardiac response throughout the program.
In a final step, analyses techniques were used to investigate the linear and non-linear
relationship in the sensor-derived data. Moreover, interpretable machine learning methods
showed that wearable device-derived parameters alone could be used as a surrogate for
functional capacity in a CR population.
To conclude, this thesis describes the initial steps necessary for a well-examined approach
towards the FitBeat index. The foundations from where future work can start to create and
implement this novel digital biomarker were laid.Hartfalen is een ernstige chronische aandoening, met een stijgende prevalentie en hoge
morbiditeits- en mortaliteitscijfers, en is de oorzaak van frequente (re)hospitalisaties. Het
is een complexe aandoening, die moeilijk te evalueren valt met slechts één enkele
parameter, waardoor de opvolging van patiënten moeilijk en uitdagend is. Regelmatige
controle en het nauwgezet opvolgen van de symptomen dragen bij tot een snellere
vaststelling van ziekteprogressie en tot de preventie van plotse achteruitgang. Een digitale
biomarker, de FitBeat score, opgebouwd uit verschillende parameters gemeten met behulp
van niet-invasieve en draagbare telemonitoring technologie, zou deze opvolging kunnen
vereenvoudigen en verbeteren.
In deze thesis werden de eerste bouwstenen van een cardiopulmonaire fitness level index,
die de complexiteit van HF omvat, gelegd. De weg naar de ontwikkeling van deze nieuwe
digitale biomarker verliep in verschillende stappen. Eerst verkregen we nieuwe inzichten in
de effectiviteit van cardiale revalidatie (CR). Het volgen van CR na CRT-implantatie bleek
kostenefficiënt te zijn. Tenslotte werd vastgesteld dat responders en niet-responders een
andere evolutie van de vooruitgang in functionele capaciteit vertoonden gedurende het CRprogramma. Deze resultaten tonen niet alleen de effectiviteit van CR als een secundaire
preventiemaatregel aan, maar bevestigen ook de nood aan een gepersonaliseerde aanpak.
De ontwikkeling van een digitale biomarker vereist het verzamelen van hoge-kwaliteit vitale
signalen waaruit fysiologische parameters geëxtraheerd kunnen worden om de biomarker
samen te stellen. Eerst moet de kwaliteit van de vitale signalen verzekerd worden zodat de
klinische implementatie correct kan gebeuren. De optimale electrodenconfiguratie voor het
correct meten van ademhalingsparameters werd bepaald, gebruik makend van een
draagbaar bioimpedantietoestel. Ook werd een algoritme voor automatische ecg-annotatie,
gevalideerd op ecg-opnames van een patiëntenpopulatie, ontwikkeld dat eveneens
implementeerbaar is in een draagbaar toestel zodat onmiddellijke verwerking
zonder tussenstappen mogelijk is.
Fysiologische parameters werden geëxtraheerd uit deze verwerkte signalen en hun evolutie
doorheen de CR werd bestudeerd. In deze thesis werd voornamelijk gefocust op
parameters geëxtraheerd uit ecg-signalen. Hartmedicatie beïnvloedde de evolutie van HRVparameters tijdens de CR. Deze resultaten benadrukken het belang van de kennis van
omliggende factoren bij de interpretatie van hartgerelateerde parameters. De niet responders en de responders werden gekenmerkt door een verschillende cardiale
respons gedurende het revalidatieprogramma.
In een laatste fase werden analysetechnieken gebruikt om de lineaire en niet-lineaire
relaties tussen de sensordata te onderzoeken. Op basis van interpreteerbare machine
learning technieken werd duidelijk dat door enkel gebruik te maken van parameters,
gemeten door de draagbare sensor, een inschatting gemaakt kan worden van de
functionele capaciteit in een CR-populatie.
Samengevat beschrijft deze thesis de initiële stappen die nodig zijn voor een goed
uitgewerkte aanpak in de ontwikkeling van de FitBeat score. De funderingen waarop
toekomstig werk in de ontwikkeling van de nieuwe biomarker gebouwd kan worden, werden
gelegd.Hélène De Cannière was supported by a doctoral fellowship by the Research Foundation Flanders,
Belgium (FWO, grant number: 1S53616N) and the Limburg Clinical Research Center (LCRC) UHasselt-
ZOL-Jessa, supported by the foundation Limburg Sterk Merk, Hasselt University, Ziekenhuis Oost-
Limburg and Jessa Hospital
FITBEAT: Towards a cardiopulmonary fitness level index for personalized heart failure management
No full textHeart failure is a chronic disease with increasing prevalence, characterized by high morbidity
and mortality rates and accompanied by frequent (re)hospitalizations. It is a complex
disease that cannot be comprised into a single parameter, making disease follow-up
challenging. Regular follow-up and monitoring of hallmark symptoms are beneficial for HF
patients to follow-up disease progression and to predict disease worsening early. A novel
digital cardiac biomarker, the FitBeat index, recorded by non-invasive remote monitoring
technologies could make this happen.
In this thesis, the initial building blocks towards a cardiopulmonary fitness level index,
comprising the complexity of the disease, were created. First, new insights into the
effectiveness of CR were gained. Following CR after CRT implant proved to be cost-effective.
Moreover, responders and non-responders to CR showed different progression patterns in
functional capacity. These insights show the effectiveness of CR as a secondary prevention
measure but also confirm the need for patient-tailored approaches.
The development of a digital biomarker requires the recording of vital signals from which
physiologically relevant features can be extracted. Initially, signal quality is essential to
ensure reliable and proper clinical implementation. We identified the most optimal electrode
configuration for respiration monitoring, using a wearable bioimpedance device. Moreover,
an algorithm that automatically annotates ECG recordings, validated in a cardiac patient
population, applicable in a wearable low power device was developed.
Physiological features were extracted from the preprocessed signals and their progression
throughout CR was studied. In this thesis, the ECG-derived features played a major role.
Cardiac-related medication intake influenced the progression of HRV parameters throughout
CR, emphasizing the importance of taking confounding factors into account when
interpreting cardiac-related features. The non-responders and responders to CR were
characterized by a different cardiac response throughout the program.
In a final step, analyses techniques were used to investigate the linear and non-linear
relationship in the sensor-derived data. Moreover, interpretable machine learning methods
showed that wearable device-derived parameters alone could be used as a surrogate for
functional capacity in a CR population.
To conclude, this thesis describes the initial steps necessary for a well-examined approach
towards the FitBeat index. The foundations from where future work can start to create and
implement this novel digital biomarker were laid.Hartfalen is een ernstige chronische aandoening, met een stijgende prevalentie en hoge
morbiditeits- en mortaliteitscijfers, en is de oorzaak van frequente (re)hospitalisaties. Het
is een complexe aandoening, die moeilijk te evalueren valt met slechts één enkele
parameter, waardoor de opvolging van patiënten moeilijk en uitdagend is. Regelmatige
controle en het nauwgezet opvolgen van de symptomen dragen bij tot een snellere
vaststelling van ziekteprogressie en tot de preventie van plotse achteruitgang. Een digitale
biomarker, de FitBeat score, opgebouwd uit verschillende parameters gemeten met behulp
van niet-invasieve en draagbare telemonitoring technologie, zou deze opvolging kunnen
vereenvoudigen en verbeteren.
In deze thesis werden de eerste bouwstenen van een cardiopulmonaire fitness level index,
die de complexiteit van HF omvat, gelegd. De weg naar de ontwikkeling van deze nieuwe
digitale biomarker verliep in verschillende stappen. Eerst verkregen we nieuwe inzichten in
de effectiviteit van cardiale revalidatie (CR). Het volgen van CR na CRT-implantatie bleek
kostenefficiënt te zijn. Tenslotte werd vastgesteld dat responders en niet-responders een
andere evolutie van de vooruitgang in functionele capaciteit vertoonden gedurende het CRprogramma. Deze resultaten tonen niet alleen de effectiviteit van CR als een secundaire
preventiemaatregel aan, maar bevestigen ook de nood aan een gepersonaliseerde aanpak.
De ontwikkeling van een digitale biomarker vereist het verzamelen van hoge-kwaliteit vitale
signalen waaruit fysiologische parameters geëxtraheerd kunnen worden om de biomarker
samen te stellen. Eerst moet de kwaliteit van de vitale signalen verzekerd worden zodat de
klinische implementatie correct kan gebeuren. De optimale electrodenconfiguratie voor het
correct meten van ademhalingsparameters werd bepaald, gebruik makend van een
draagbaar bioimpedantietoestel. Ook werd een algoritme voor automatische ecg-annotatie,
gevalideerd op ecg-opnames van een patiëntenpopulatie, ontwikkeld dat eveneens
implementeerbaar is in een draagbaar toestel zodat onmiddellijke verwerking
zonder tussenstappen mogelijk is.
Fysiologische parameters werden geëxtraheerd uit deze verwerkte signalen en hun evolutie
doorheen de CR werd bestudeerd. In deze thesis werd voornamelijk gefocust op
parameters geëxtraheerd uit ecg-signalen. Hartmedicatie beïnvloedde de evolutie van HRVparameters tijdens de CR. Deze resultaten benadrukken het belang van de kennis van
omliggende factoren bij de interpretatie van hartgerelateerde parameters. De niet responders en de responders werden gekenmerkt door een verschillende cardiale
respons gedurende het revalidatieprogramma.
In een laatste fase werden analysetechnieken gebruikt om de lineaire en niet-lineaire
relaties tussen de sensordata te onderzoeken. Op basis van interpreteerbare machine
learning technieken werd duidelijk dat door enkel gebruik te maken van parameters,
gemeten door de draagbare sensor, een inschatting gemaakt kan worden van de
functionele capaciteit in een CR-populatie.
Samengevat beschrijft deze thesis de initiële stappen die nodig zijn voor een goed
uitgewerkte aanpak in de ontwikkeling van de FitBeat score. De funderingen waarop
toekomstig werk in de ontwikkeling van de nieuwe biomarker gebouwd kan worden, werden
gelegd.Hélène De Cannière was supported by a doctoral fellowship by the Research Foundation Flanders,
Belgium (FWO, grant number: 1S53616N) and the Limburg Clinical Research Center (LCRC) UHasselt-
ZOL-Jessa, supported by the foundation Limburg Sterk Merk, Hasselt University, Ziekenhuis Oost-
Limburg and Jessa Hospital
Remote Monitoring of COVID-19 Patients Following Discharge from a Tertiary Care Center
No full textThe COVID-19 pandemic has affected people, healthcare systems and caregivers on a global scale causing bottlenecks in hospital resources and overload of healthcare systems. The presence of disease sequelae in
patients hospitalized due to COVID-19 warrants additional care and monitoring of these patients. Remote monitoring techniques have been implemented in several domains of healthcare such as cardiology, cardiac rehabilitation and nephrology. Monitoring of vital signs using these technologies has allowed the tracking of patients with more granularity, resulting in better clinical outcomes such as reduction in hospitalizations. Therefore, we hypothesize that remote monitoring is beneficial in managing COVID-19 patients post-hospitalization, enabling home-based patient follow-up. In this study, we investigated the use of remote
monitoring on a COVID-19 patient cohort discharged from a tertiary care center. A post-hoc division of patients into two groups (alert-generating patients and non-alert generating patients) was performed. The longitudinal progression of sensor and questionnaire data was studied using linear mixed-effect models. The measured heart rate values were statistically significant in terms of the intercept (p<0.001), indicating a difference between the two patient groups at baseline immediately post-discharge.The authors would like to acknowledge Dr. David Ruttens for his help in data collection for this project. The authors would like to thank the assistants and nurses at the Department of Pneumology, Ziekenhuis Oost-Limburg as well as the scientific researchers of Future Health for their help in the data collection procedure
Identifying Changes in Functional Capacity of Cardiorespiratory Patients Undergoing Exercise Rehabilitation
No full textHeart failure (HF) and chronic obstructive pulmonary disease (COPD) are two common complex multimorbid cardiorespiratory diseases. Due to their complex nature, it is challenging to identify changes in health status of affected patients. In this study, we analyse the progression of functional capacity as a first step in identifying changes in disease status of HF and COPD patients. 60 patients (NHF=35, NCOPD=25) undergoing cardiopulmonary rehabilitation were included in this study. A six-minute walk test (6MWT) assessing the sixminute walking distance (6MWD) was used to monitor the functional capacity of these patients. Patients performed five 6MWTs in total (1 baseline, 4 follow-up) with spotcheck HR and SpO2 values also being measured before and after each 6MWT. The progression of the 6MWDs was analysed using a two-way mixed ANOVA. To predict changes in functional capacity, patients were divided into two groups (“improved” vs “not improved”) based on a minimal clinically significant distance change. A decision tree classifier was trained on 6MWD, HR and SpO2 data features and evaluated using balanced accuracy. The mixed ANOVA showed a significant interaction effect as well as significant between-subjects and withinsubject effects. The classifier showed good performance in predicting improvement of functional capacity.The authors would like to thank Daimy Roebroek, Frauke Somers and Julie Deckers for their help in collecting data from patients for this study
Real Time Electrocardiogram Annotation with a Long Short Term Memory Neural Network
No full textContinuous monitoring of electrocardiogram from wearable devices can enable early detection of heart diseases. Ubiquitous monitoring on wearable electronics requires a novel class of algorithms that are low-power and have low-memory requirements. This work proposes a wearable compatible, and automatic solution for annotating Electrocardiogram (ECG) recordings while maintaining high accuracy of detection when users are carrying daily activities such as sitting, walking, and resting. We validate our solution with two Physionet datasets: the MITDB [1] (Boston's Beth Israel Hospital and MIT Arrhythmia Database), and the EDB [2] (European ST-T Database). In addition, we validate our method on a newly recorded dataset in collaboration with the 'Ziekenhuis Oost-Limburg' Hospital(1) that has been collected using a prototype wearable device [3]. Our solution exploits a recurrent neural network that achieves an average F1 score of 94.8% over all three datasets. Our solution achieves better generalization performance than the gold standard method Pan Tompkins which achieves an average F1 score of 93%. In addition, our method can be extended to full ECG annotation. We used the QTDB dataset [4] and we report an accuracy of 91.6% while annotating all 5 waves (P-Q-R-S-T) of the ECG complex.ITEA3 PARTNER projectThis work is supported by the ITEA3 PARTNER project (Patient-care Advance-ment with Responsive Technologies aNd Engagement togetheR).Corradi, F (reprint author), Stichting IMEC Nederland, Ultra Low Power Syst IoT, Eindhoven, Netherlands.
[email protected]
Effective Orifice Area during Exercise in Bileaflet Mechanical Valve Prostheses
No full textBackground: The aims of this study were to investigate the evolution of the transprosthetic pressure gradient and effective orifice area (EOA) during dynamic bicycle exercise in bileaflet mechanical heart valves and to explore the relationship with exercise capacity. Methods: Patients with bileaflet aortic valve replacement (n = 23) and mitral valve replacement (MVR; n = 16) prospectively underwent symptom-limited supine bicycle exercise testing with Doppler echocardiography and respiratory gas analysis. Transprosthetic flow rate, peak and mean transprosthetic gradient, EOA, and systolic pulmonary artery pressure were assessed at different stages of exercise. Results: EOA at rest, midexercise, and peak exercise was 1.66 +/- 0.23, 1.56 +/- 0.30, and 1.61 +/- 0.28 cm(2), respectively (P = .004), in aortic valve replacement patients and 1.40 +/- 0.21, 1.46 +/- 0.27, and 1.48 +/- 0.25 cm(2), respectively (P = .160), in MVR patients. During exercise, the mean transprosthetic gradient and the square of transprosthetic flow rate were strongly correlated (r = 0.65 [P < .001] and r = 0.84 [P < .001] for aortic valve replacement and MVR, respectively), conforming to fundamental hydraulic principles for fixed orifices. Indexed EOA at rest was correlated with exercise capacity in MVR patients only (Spearman r = 0.68, P = .004). In the latter group, systolic pulmonary artery pressures during exercise were strongly correlated with the peak transmitral gradient (r = 0.72, P < .001). Conclusions: In bileaflet mechanical valve prostheses, there is no clinically relevant increase in EOA during dynamic exercise. Transprosthetic gradients during exercise closely adhere to the fundamental pressure-flow relationship. Indexed EOA at rest is a strong predictor of exercise capacity in MVR patients. This should be taken into account in therapeutic decision making and prosthesis selection in young and dynamic patients.Dr. Bertrand is supported by a grant from the Research Foundation-Flanders (11N7214N). Drs. Bertrand and Vandervoort are researchers for the Limburg Clinical Research Program UHasselt-ZOL-Jessa, supported by the foundation Limburg Sterk Merk, Hasselt University, Ziekenhuis Oost-Limburg, and Jessa Hospital. Dr. Dion has received consulting fees from Edwards Lifesciences, Johnson & Johnson, Sorin, Medtronic, and St. Jude Medical
Is multidisciplinary cardiac rehabilitation following cardiac resynchronization cost-effective? (#)
No full textIs multidisciplinary cardiac rehabilitation following cardiac resynchronization cost-effective? (#)
No full textMonitoring hemodynamics in dialysis: on the feasibility of semi-continuous and multifrequency thoracic bioimpedance measurements by a wearable device
No full textTowards personalized fluid monitoring in haemodialysis patients: thoracic bioimpedance signal shows strong correlation with fluid changes, a cohort study
No full textBACKGROUND: Haemodialysis (HD) patients are burdened by frequent fluid shifts which amplify their comorbidities. Bioimpedance (bioZ) is a promising technique to monitor changes in fluid status. The aim of this study is to investigate if the thoracic bioZ signal can track fluid changes during a HD session. METHODS: Prevalent patients from a single centre HD unit were monitored during one to six consecutive HD sessions using a wearable multi-frequency thoracic bioZ device. Ultrafiltration volume (UFV) was determined based on the interdialytic weight gain and target dry weight set by clinicians. The correlation between the bioZ signal and UFV was analysed on population level. Additionally regression models were built and validated per dialysis session. RESULTS: 66 patients were included, resulting in a total of 133 HD sessions. Spearman correlation between the thoracic bioZ and UFV showed a significant strong correlation of 0.755 (p < 0.01) on population level. Regression analysis per session revealed a strong relation between the bioZ value and the UFV (R2 = 0.982). The fluid extraction prediction error of the leave-one-out cross validation was very small (56.2 ml [- 121.1-194.1 ml]) across all sessions at all frequencies. CONCLUSIONS: This study demonstrated that thoracic bioZ is strongly correlated with fluid shifts during HD over a large range of UFVs. Furthermore, leave-one-out cross validation is a step towards personalized fluid monitoring during HD and could contribute to the creation of autonomous dialysis.sponsorship: We would like to thank the clinical and technical staff at the participating units for their help and support. We would like to thank the engineers from imec the Netherlands for their technical support. This research is part of the Limburg Clinical Research Center (LCRC) UHasselt-ZOL-Jessa, supported by the foundation Limburg Sterk Merk (LSM), province of Limburg, Flemish government, UHasselt, Ziekenhuis Oost-Limburg and Jessa Hospital, Belgium. (foundation Limburg Sterk Merk (LSM), province of Limburg, Flemish government, UHasselt, Ziekenhuis Oost-Limburg, Jessa Hospital, Belgium)status: Publishe
- …
