Get started
Work

Mensgericht onderzoek en ontwikkeling van een omgevingsmonitoringsysteem om incontinentiezorg in woonzorgcentra te ondersteunen

Year: 2024
Type: dissertation
Abstract: The integration of technology in continence care management within nursing homes offers a promising avenue for enhancing the quality of life for residents while reducing the workload for care personnel. Advanced technological solutions, such as sensor-based systems, enable monitoring and timely interventions, potentially leading to personalised incontinence assessments. However, current technologies primarily focus on managing incontinence, rather than promoting continence. In other words, these technologies are designed to maintain dryness through various continence aids, rather than actively supporting nursing home residents to achieve continence. As such, there is a need for further innovation in this area. This manuscript outlines a human-centred research approach to the development of an unobtrusive ambient monitoring sensor system for continence care, with the primary objective of promoting continence. Initially, 27 nursing home stakeholders (i.e., residents and their families, care personnel, and management) were interviewed and observed to gather their wishes and needs for such technology. In our research, we noted that the current technology supporting continence care in nursing homes is mainly driven by engineering efforts, with limited attention given to human-centred perspectives. Consequently, we identified design considerations for such technology aimed at fostering trust and dignity as central concerns. We proposed that future technologies should re-shift and broaden the focus, develop sensitive methods around intimacy, stigma and taboo, and employ approaches that embrace a socio-technical perspective. These design considerations guided the development of the initial prototype of our unobtrusive ambient monitoring sensor system to detect toileting activities, including voiding events and leakage. Off-the-shelf ammonia gas sensors were selected for odour detection based on their detection range within the concentration of ammonia found in urine. Initial laboratory experiments demonstrated the ability of semiconductor gas sensors to detect relevant concentrations of ammonia in solutions through air. However, some limitations were identified, leading to refinements such as employing a study setup covered with bed sheets and eliminating electrochemical gas sensors for the application. Subsequently, an adjusted sensor system connected to the care bed was subjected to an exploratory study conducted with seven residents of two Flemish nursing homes. This study revealed that the dynamic nature of the nursing home environment, characterised by activities including opening doors and windows, presented challenges to reliable odour detection. Another challenge encountered was finding an appropriate annotation method to serve as a ground truth labelling for validation purposes. Obtaining data through manual data entry by care personnel proved to be difficult. Hence, a wearable ultrasound measurement device was evaluated; however, the data obtained was not conclusive. Finally, an exploration of inertial measurement units applied on a viscoelastic care mattress was carried out to monitor nighttime movements. Six participants followed a seven-step protocol, resulting in the identification of four classes: in/out of bed, turning, and agitation. This involved employing an XGBoost model combined with a leave-one-subject-out cross-validation for analysis, resulting in a promising overall F1-score of 79.56% for the selected machine learning algorithm. In conclusion, this doctoral research contributes to the field of healthcare technology, particularly in the context of nursing home care. It offers detailed insights into human-centred research, development, and evaluation of an unobtrusive ambient monitoring sensor system aimed at promoting continence care. In addition, a critical reflection on conducting human-centred research in nursing homes is shared, translated into actionable guidelines focusing on participant selection, care personnel dynamics, and environmental complexity. Although the findings presented in this manuscript offer a nuanced understanding of ambient intelligence for continence care solutions, it is evident that future sensor systems and machine learning algorithms may advance to detect various toileting activities. These activities may include monitoring saturation levels of absorption material, detecting leakage, observing nighttime behaviour, and identifying agitation. The integration of this information into nursing home information technology has the potential to establish a decision support tool for continence care management. Such a tool could provide actionable insights, ultimately enhancing the quality of care.
Source: Lirias
Language: English
Open Access status: green
OpenAlex ID: W7039463554
Cites:
Cited by:
Related to: