Development and clinical application of a computer-aided real-time feedback system for detecting in-bed physical activities
Introduction
Many frail inpatients have the concomitant problem of low mobility, which can result in the morbid phenomenon of being bedridden for long periods of time. This phenomenon occurs for approximately one-third of all inpatients [1], of whom approximately 23%–33% are seniors [2], [3]. Whether for frail bedridden seniors or young inpatients, being bedridden for long periods of time can seriously impact the capacity for physical activities, and this impact becomes progressively exacerbated as the number of days in the hospital increases [4]. In addition to such factors as a patient's age or disease, other contributors to a patient's functional decline in daily life may include being required to lie in bed after special medical treatment [5], [6], [7]. For instance, a patient's movement may be restricted or the patient may be treated with a fixation device to prevent injury after an operation [1], [6], [8]. Studies have shown that even when patients have their disease under control and have resumed some activities, even demonstrating the ability to walk, the duration of confinement to bed still accounts for more than 60% of the typical duration of hospitalization [9], [10]. Early-stage research has shown that being bedridden for long periods of time can not only lead to deterioration in patients’ physiological functions but also affect functional performance in daily life [9], even increasing the number of days in the hospital or in long-term care institutions, and raising the risk of death and the incidence of relevant complications and accidental injuries [8], [11], [12], [13]. Furthermore, bedridden patients are susceptible to cognitive disorders, dizziness, and physical weakness, among other effects on their physiological functions, due to drug side effects [12], [13], resulting in insufficient blood supply to the head following rapid postural changes or dizziness caused by balance dysfunction of the inner ear. These effects may result in accidental injuries if the patient falls from the bed to the floor [14].
An inpatient is most likely to fall and become injured at the bedside, mainly because of the unstable position of the body's center of gravity when the inpatient attempts to rise from the bed and walk or grasp the edge of the bed to stand up [15]. Between approximately 13 and 20% of inpatients suffer at least one fall [14]. Studies indicate that such falls result in death in 13% of patients and physical injuries, such as fractures, joint dislocations, lacerations, and bruising, in 25% of patients [14]. Falls may also cause psychological trepidation and impact functionality in daily life; hence, patients may reduce their participation in daily socializing activities for fear of falls [16], [17], [18], [19], increasing the burden on family, medical personnel, and the medical care service system [20], as well as raising medical expenditures [12], [14], [21], [22]. Therefore, many clinical institutions have equipped their sickbeds with protective handrails or hand grips to prevent patients from accidentally injuring themselves as a result of falling from the bedside [23]. However, research has shown that even equipment with bedside handrails remains incapable of significantly reducing the incidence of falls [23]. For this reason, clinical caregivers need to develop strategies to monitor changes in patients’ in-bed activities and prevent bedside fall accidents.
There are many clinical methods for monitoring inpatients’ in-bed activities and postural changes and for preventing bedside falls, such as case records, anti-fall bracelets, and regular ward rounds [10]. However, studies have shown that these methods have many limitations, including deficiencies in recording patients’ activities whereby transient bedside activities by patients are overlooked or recorded erroneously [10]. As the care resources for the elderly become scarcer because of the aging population trends, relevant studies have also noted that a small number of clinical caregivers who take care of numerous elderly inpatients, especially caregivers who are particularly busy at night, are more likely to overlook dangerous accidents in which elderly patients fall when getting out of bed or slide down to the floor from the edge of the bed [24]. Therefore, relevant studies have recommended the use of a detecting system with real-time sensing of posture changes and providing monitoring information to warn caregivers in advance of patients’ potentially dangerous actions. Such a system should be adjustable to the status in different individual cases [14], [23], [24].
The advantages of integrating sensors to monitor patients’ changes between different postures include continuous monitoring, no need for in-person observation, and objective measurement records [10]. For this reason, early studies have integrated image-sensing components and analytical techniques to determine bedridden patients’ in-bed activities, providing caregivers with monitoring information about dangerous actions or postural changes in individual cases [25]. However, because of concerns over invasion of individual privacy, real-time image monitoring is not recommended for use with clinical patients who are bedridden [26]. In recent years, relevant research has developed lightweight, wearable measurement tools to record and analyze information about individual actions in real time and thereby determine postural changes in or off the bed, such as lying down, sitting up, getting off the bed and standing up, etc. [10], [27], [28], [29], [30], [31], [32]. There have even been studies in which wireless communications technology has been used to transmit data acquired through measurement to a computer, allowing caregivers to carry out follow-up analysis and long-term monitoring [10], [22], [24], [31].
Nonetheless, there may be some limitations to the clinical application of relevant sensing components in posture change determination and detection, including interference with patients’ posture-changing processes, possible compression and injury due to oversized sensing components, and the potential for accidental falls when a patient tries to pick up sensing components that have fallen off [24]. These limitations would affect a patient's daily physical activities [25]. Furthermore, excessive variance in the measurement values of falls by the sensing components [33], errors in the posture change determination formula, oversensitive threshold settings, excessively low sampling frequencies, and other factors may also affect the system's ability to measure postural changes accurately and produce anti-fall warnings effectively [25], [32], [33], [34], [35]. For instance, the accuracy of chest-worn detecting systems in monitoring postural changes is in the range of 72–89.3% [11], [22], [25]. Furthermore, most measuring systems do not base their actual measurements on real hospital sickbeds [36], and errors in operation or placement of the sensing components can also lead to incorrect determination of posture changes [16] and affect the accuracy of postural change detection [22].
Therefore, some studies have employed a metal bed surface with pressure-sensing components for detecting postural changes. However, this type of equipment has shortcomings, including patient discomfort, reduced sensing component lifespan, and interference with data measurement resulting from frequent compression of sensing components due to postural changes [16], [24]. Research has also shown that wearable sensing components cannot be used for very long durations to monitor in-bed posture changes, as battery replacement is required approximately every 7–10 days [22], [36], whereas only monitoring systems or instruments such as computer-aided monitoring systems that can measure patients’ physical postural changes over long periods of time can effectively prevent bedridden patients from suffering accidental injuries [12].
The objectives of this research were to develop a computer-aided real-time system for detecting in-bed physical activities and to establish tests of the validity and reliability of systematic measurement of in-bed postural changes, to highlight the value of clinical application of the detection of in-bed postural changes and warnings against falls for frail bedridden patients.
Section snippets
Hardware architecture
The hardware used in this research included one standard sickbed equipped with four load cells (TEDEA 615 Huntleigh Co.), a signal amplifier, an NI-6008 data acquisition card (DAQ, 12-bit analog-to-digital converter, National Instruments Corp., Austin, TX, USA), and a laptop computer (Fig. 1). Each load cell was mounted under the foot of the standard sickbed and was able to measure 980 N to evaluate the pressure of the body weight. The voltage output from the circuitry for each load cell was
Results
The results showed that the physical activity detecting system demonstrated high validity (r = 0.928, p < 0.05) with an accuracy reaching 87.9% for monitoring real individuals’ postural changes during eight in-bed postures. These results suggest that the system can be used to accurately measure and determine subjects’ in-bed postural changes and states in the range of 70–100% (Table 2). The results also find that incorrect determinations may occur during specific postural changes, especially in
System development
A physical activity detecting system that can effectively determine subjects’ in-bed postural changes and states and provide accurate validity and test–retest reliability for posture change determination was developed in this study.
Previous studies also described the design of a force sensing-based mattress system to detect different postures and movement onset times for caregivers or investigators, which is helpful in repositioning the sleeping posture for bedridden patients in order to reduce
Conclusions
A computer-aided real-time bed activities monitoring system, consisting of a clinical sickbed with pressure-sensing components and programmable software, was developed in this study. This study indicated that the system can be used to detect postural changes with good accuracy and excellent test-retest reliability.
Acknowledgments
This study was supported by the Tri-Service General Hospital (TSGH-C105-105), Ministry of National Defense-Medical Affairs Bureau (MAB-105-097) and Taipei Medical University (TMU103- AE1-B24).
References (45)
- et al.
Adverse consequences of hospitalization in the elderly
Soc. Sci. Med.
(1982) - et al.
Serious falls in hospitalized patients: correlates and resource utilization
Am. J. Med.
(1995) - et al.
Predictors of falling in elderly hospital patients
Arch. Gerontol. Geriatrics
(2004) - et al.
Sensitivity and specificity of fall detection in people aged 40 years and over
Gait Posture
(2009) - et al.
The sensitivity and specificity of an activity monitor in detecting functional activities in young people with cerebral palsy
Arch. Phys. Med. Rehab.
(2009) - et al.
Functional roles of abdominal and back muscles during isometric axial rotation of the trunk
J. Orthopaed. Res.
(2001) - et al.
A predictive index for functional decline in hospitalized elderly medical patients
J. Gener. Internal Med.
(1993) - et al.
The provision of physical activity to hospitalized elderly patients
Arch. Internal Med.
(1991) - et al.
Functional disability in the hospitalized elderly
JAMA
(1982) - et al.
The natural history of functional morbidity in hospitalized older patients
J. Am. Geriatrics Soc.
(1990)
Physiology and complications of bed rest
J. Am. Geriatrics Soc.
Hazards of hospitalization of the elderly
Ann. Internal Med.
Hospital-associated deconditioning and dysfunction
J. Am. Geriatrics Soc.
Prevalence and outcomes of low mobility in hospitalized older patients
J. Am. Geriatrics Soc.
The underrecognized epidemic of low mobility during hospitalization of older adults
J. Am. Geriatrics Soc.
Predictors of immediate and 6-month outcomes in hospitalized elderly patients. The importance of functional status
J. Am. Geriatrics Soc.
The value of assessing falls in an elderly population. A randomized clinical trial
Ann Internal Med
In-bed patients behaviour monitoring system, biocomputation, bioinformatics, and biomedical technologies
Interventions for preventing falls in acute- and chronic-care hospitals: a systematic review and meta-analysis
J Am Geriatrics Soc
Fall prevention in acute care hospitals: a randomized trial
JAMA
Medicare nonpayment, hospital falls, and unintended consequences
New Engl. J. Med.
Cited by (5)
A mobile computer aided system for optic nerve head detection
2018, Computer Methods and Programs in BiomedicineCitation Excerpt :It has now become possible to run computationally intensive image processing algorithms, such as face detection and recognition, object identification and tracking, on smartphones. In the medical area, smartphones are now being used for Computer-Aided-System (CAS) of physical activity detection [26], neurology performance evaluation [45], etc.; and for Computer-Aided-Diagnosis (CAD) systems including electroencephalogram EEG [46] and electrocardiograph (ECG) [2,52]. In this paper, we are focusing on a smartphone-based CAD system for retinopathy.
The novel use of an Extreme learning machines for clinical decision support systems
2017, Computer Methods and Programs in BiomedicineContributions of the Kinematics Analysis with EMG for Rotator Cuff Tests—A Case Report
2023, Lecture Notes in BioengineeringOptimization Process of an Innovative Rehabilitation Device based on Pre-Clinical Results
2023, WSEAS Transactions on Information Science and ApplicationsRisk-Taking Behaviors of Adult Bedridden Patients in Neurosurgery: What Could/Should We Do?
2021, Frontiers in Medicine
- 1
The author contributed equally to the first author.