Biomedical Engineering Reference
In-Depth Information
1
Introduction
1.1
Background
Patients with chronic obstructive pulmonary disease (COPD) are suffering from the
consequences of a chronic inflammation of their pulmonary system. This leads to an
obstruction of the bronchi that causes airflow limitation and shortness of breath. Of-
ten, immobility and social isolation are the consequences, which in turn reinforce the
degeneration of muscle mass and aggravate the symptoms. The Global Initiative for
Chronic Obstructive Lung Disease (GOLD) summarizes: “COPD is the fourth leading
cause of death in the world and further increases in its prevalence and mortality can
be predicted in the coming decades” [1]. Just the direct medical costs attributable to
COPD were estimated at $49.5 billion in the US [2].
Beside the pharmacological treatment, an important part of therapy is regular endur-
ance training. Pulmonary rehabilitation training improves physical capacity, reduces
breathlessness, reduces the number of hospitalizations and increases the quality of life
[1]. The cost of continuous monitoring of these training sessions in clinics is high and
additionally requires the patient to travel to a clinic for each single session. Performing
the rehabilitation training at home can raise the patients' compliance and reduce costs.
Another unsolved problem in today's healthcare systems is the connection and data
exchange between these different actors over the borders of institutions and health
care sectors. Especially for COPD it is recommended to involve lounge specialists,
nutritionists, psychologists, and family doctors to ensure an optimal treatment. In
addition the patient should be involved in his own treatment and is although the only
one who could provide data about the own sleeping behavior, nutrition, tobacco con-
sume, and sport activities.
1.2
Related Work
To ensure a safe telerehabilitation at home, the detection of abnormal events during
the training session and an autonomous training control are critical prerequisites.
Nearly all the existing detection and control algorithms compare the patient's “is”
state with the “should” state. Therefore different sets of vital signs are used as an
indicator to derive the health state of the patient, which reflects the training intensity.
Achten and Jeukendrup summarized current research achievements in the field of
heart rate monitoring in 2003 and state: “…the most important application of HR
monitoring is to evaluate the intensity of the exercise performed” [3]. They conclude
that the important influence factors on HR are age, gender, environmental tempera-
ture, hydration and altitude. They estimated the day-to-day variance under controlled
conditions to be 2-4 beats per minute (bpm).
Different techniques have been proposed for controlling the training performance
and to raise alarms on basis of HR. Velikic et al. used data from an accelerometer for
a comparison of different models (linear, non-linear, Kalman filter) for HR prediction
of healthy subjects and such with congestive heart failure [4]. The two linear models
delivered the best results for a short term prediction of 20 minutes. Su et al. intro-
duced a model to control HR during treadmill exercise [5]. Further approaches for the
same application were provided by Cheng et al. [6] and Mazenc et al. [7]. Neither
Search WWH ::
Custom Search