Supporting rehabilitation training of COPD patients through multivariate
sensor-based monitoring and autonomous control using a Bayesian network:
prototype and results of a feasibility study
Repeated endurance training - supervised by an expert - is one of
the most effective rehabilitation methods for patients with chronic
obstructive pulmonary disease (COPD) to improve physical function.
Monitoring of vital signs in combination with an automatic intelligent
training control and emergency detection facilitates supervised training
without the physical presence of an expert as well as training optimisation
through individualisation. The aim of this study is the development
of a suitable analysis and control method for this purpose. Healthy
volunteers and patients with COPD were equipped with body sensors
during ergometer training to enable measuring their vital signs continuously.
Depending on these values, the exercise load of the ergometer was
controlled automatically using a Bayesian network. The network, trained
with expert knowledge and training data, is embedded in our system
by using Java application programming interface. Extensive tests
in a laboratory setting have proved safe usage of our prototype.
In a case study, evaluation during training sessions with patients
with COPD took place. Due to the automatic control the patients'
vital signs ranged inside the predefined optimal thresholds for at
least 95\% of the time. Furthermore, our results suggest an increase
of the training efficiency compared with the conventional method
(constant exercise load).
Peter L. Reichertz Institute for Medical Informatics, University
of Braunschweig-Institute of Technology and Hannover Medical School,
Mühlenpfordtstrasse 23, Braunschweig, Germany. mareike.schulze@plri.de
journal
Inform Health Soc Care
number
3-4
pages
144--156
volume
35
medline-pst
ppublish
pmid
21133769
file
Schulze2010_IHSC_GAL_COPD Training Bayes.pdf:0_Finale Versionen\\Schulze2010_IHSC_GAL_COPD Training Bayes.pdf:PDF
%0 Journal Article
%1 Schulze2010
%A Schulze, Mareike
%A Song, Bianying
%A Gietzelt, Matthias
%A Wolf, Klaus-Hendrik
%A Kayser, Riana
%A Tegtbur, Uwe
%A Marschollek, Michael
%D 2010
%J Inform Health Soc Care
%K Aged; Bayes Chronic Disease, Ergometry; Exercise Female; Humans; Male; Middle Obstructive, Pulmonary Remote Sensing Technology, Theorem; Therapy, instrumentation instrumentation; rehabilitation;
%N 3-4
%P 144--156
%R 10.3109/17538157.2010.528659
%T Supporting rehabilitation training of COPD patients through multivariate
sensor-based monitoring and autonomous control using a Bayesian network:
prototype and results of a feasibility study
%V 35
%X Repeated endurance training - supervised by an expert - is one of
the most effective rehabilitation methods for patients with chronic
obstructive pulmonary disease (COPD) to improve physical function.
Monitoring of vital signs in combination with an automatic intelligent
training control and emergency detection facilitates supervised training
without the physical presence of an expert as well as training optimisation
through individualisation. The aim of this study is the development
of a suitable analysis and control method for this purpose. Healthy
volunteers and patients with COPD were equipped with body sensors
during ergometer training to enable measuring their vital signs continuously.
Depending on these values, the exercise load of the ergometer was
controlled automatically using a Bayesian network. The network, trained
with expert knowledge and training data, is embedded in our system
by using Java application programming interface. Extensive tests
in a laboratory setting have proved safe usage of our prototype.
In a case study, evaluation during training sessions with patients
with COPD took place. Due to the automatic control the patients'
vital signs ranged inside the predefined optimal thresholds for at
least 95\% of the time. Furthermore, our results suggest an increase
of the training efficiency compared with the conventional method
(constant exercise load).
@article{Schulze2010,
abstract = {Repeated endurance training - supervised by an expert - is one of
the most effective rehabilitation methods for patients with chronic
obstructive pulmonary disease (COPD) to improve physical function.
Monitoring of vital signs in combination with an automatic intelligent
training control and emergency detection facilitates supervised training
without the physical presence of an expert as well as training optimisation
through individualisation. The aim of this study is the development
of a suitable analysis and control method for this purpose. Healthy
volunteers and patients with COPD were equipped with body sensors
during ergometer training to enable measuring their vital signs continuously.
Depending on these values, the exercise load of the ergometer was
controlled automatically using a Bayesian network. The network, trained
with expert knowledge and training data, is embedded in our system
by using Java application programming interface. Extensive tests
in a laboratory setting have proved safe usage of our prototype.
In a case study, evaluation during training sessions with patients
with COPD took place. Due to the automatic control the patients'
vital signs ranged inside the predefined optimal thresholds for at
least 95\% of the time. Furthermore, our results suggest an increase
of the training efficiency compared with the conventional method
(constant exercise load).},
added-at = {2013-07-30T15:13:48.000+0200},
author = {Schulze, Mareike and Song, Bianying and Gietzelt, Matthias and Wolf, Klaus-Hendrik and Kayser, Riana and Tegtbur, Uwe and Marschollek, Michael},
biburl = {https://www.bibsonomy.org/bibtex/288aeea2127a847376762c0f50faac958/khwolf},
doi = {10.3109/17538157.2010.528659},
file = {Schulze2010_IHSC_GAL_COPD Training Bayes.pdf:0_Finale Versionen\\Schulze2010_IHSC_GAL_COPD Training Bayes.pdf:PDF},
institution = {Peter L. Reichertz Institute for Medical Informatics, University
of Braunschweig-Institute of Technology and Hannover Medical School,
Mühlenpfordtstrasse 23, Braunschweig, Germany. mareike.schulze@plri.de},
interhash = {b720af3f48c20d152ad0c9872396eb8c},
intrahash = {88aeea2127a847376762c0f50faac958},
journal = {Inform Health Soc Care},
keywords = {Aged; Bayes Chronic Disease, Ergometry; Exercise Female; Humans; Male; Middle Obstructive, Pulmonary Remote Sensing Technology, Theorem; Therapy, instrumentation instrumentation; rehabilitation;},
language = {english},
medline-pst = {ppublish},
number = {3-4},
owner = {Klaus-Hendrik Wolf},
pages = {144--156},
pmid = {21133769},
timestamp = {2013-07-30T15:13:54.000+0200},
title = {Supporting rehabilitation training of COPD patients through multivariate
sensor-based monitoring and autonomous control using a Bayesian network:
prototype and results of a feasibility study},
volume = 35,
year = 2010
}