The automated detection of stress is a central problem for ambient assisted living solutions. The paper presents the concepts and results of two studies targeted at stress detection with a low cost heart rate sensor, a chest belt. In the device validation study (), we compared heart rate data and other features from the belt to those measured by a gold standard device to assess the reliability of the sensor. With simple synchronization and data cleaning algorithm, we were able to select highly (>97%) correlated, low average error (2.2%) data segments of considerable length from the chest data for further processing. The protocol for the clinical study () included a relax phase followed by a phase with provoked mental stress, 10 minutes each. We developed a simple method for the detection of the stress using only three time-domain features of the heart rate signal. The method produced accuracy of 74.6%, sensitivity of 75.0%, and specificity of 74.2%, which is impressive compared to the performance of two state-of-the-art methods run on the same data. Since the proposed method uses only time-domain features, it can be efficiently implemented on mobile devices.
Background: Using Ambient Assisted Living sensors to detect acute stress could help people mitigate the harmful effects of everyday stressful situations. This would help both the healthy and those affected more by sudden stressors, e.g., people with diabetes or heart conditions. The study aimed to develop a method for providing reliable stress detection based on heart rate variability features extracted from portable devices. Methods: Features extracted from portable electrocardiogram sensor recordings were used for training various classification algorithms for stress detection purposes. Data were recorded in a clinical trial with 7 participants and two stressors, the Trier Social Stress Test and the Stroop colour word test, both validated by standardised questionnaires. Different heart rate variability feature sets (all, time-domain and non-linear only, frequency-domain only) were tested to investigate how classification performance is affected, in addition to various time window length setups and participant-wise training sessions. The accuracy and F1 score of the trained models were compared and analysed. Results: The best results were achieved with models using time-domain and non-linear heart rate variability features with 5-min-long overlapping time windows, yielding 96.31% accuracy and 96.26% F1 score. Shorter overlapping windows had slightly lower performance, with 91.62–94.55% accuracy and 91.77–94.55% F1 score ranges. Non-overlapping window configurations were less effective, with both accuracy and F1 score below 88%. For participant-wise learning, average F1 scores of 99.47%, 98.93% and 96.1% were achieved for feature sets using all, time-domain and non-linear, and frequency-domain features, respectively. Conclusion: The tested stress detector models based on heart rate variability data recorded by a single electrocardiogram sensor performed just as well as those published in the literature working with multiple sensors, or even better. This suggests that once portable devices such as smartwatches provide reliable hear rate variability recordings, efficient stress detection can be achieved without the need for additional physiological measurements.
1 October 2020 in Computers in Biology and Medicine
Diabetes Mellitus outpatients would benefit from a lifestyle support tool that delivers reliable short term Blood Glucose Level (BGL) predictions. To develop a method for BGL prediction based on the baseline BGL, the insulin dosing and a dietary log. A new training method is proposed for a neural network in which an absorption model is applied that uses the nutrient contents of meals. The numerical characteristics of the computed absorption curve are fed to the neural network as training inputs along with the applied insulin doses and BGL evolution measured by a Continuous Glucose Monitoring System. For comparison, another version of the training in which raw carbohydrate values are used as dietary inputs has also been implemented. The method was validated in a clinical trial with 5 patients using a total of 167 meals. It was found that the proposed method performed significantly better on the 60- and 120-min prediction horizons, with a Root Mean Square Error of 1.12 mmol/l and 1.75 mmol/l, respectively, and more than 96% of the predicted values falling in the 'clinically acceptable' class according to clinical practice. These results surpass those published results to which our method is directly comparable, and also those of the carbohydrate-only version (1.81 mmol/l and 2.53 mmol/l). The integration of the absorption model in the training process has successfully contributed to the success of the model. Future research will focus on a new trial with more patients to verify these promising results.
10 January 2019 in Journal of Healthcare Engineering
According to recent surveys, the current ways of diabetics trying to estimate their insulin need based on experience and conjecture are sometimes inefficient in practice. This paper proposes a prediction algorithm and presents the validation of the model in outpatient care. The algorithm consists of two state-of-the-art models that calculate nutrition absorption and glycaemia including insulin evolution. The combined model is extended with personalized parameter training including genetic algorithm and Nelder–Mead method, and a more realistic, diurnal parameter profile as a representation of the natural biorhythm. This method implemented in a user-friendly application can help diabetics calculate their insulin need. The tests were performed on a data set including a clinical trial involving more than 20 diabetic patients. We experienced 55% improvement in the results due to model training compared to the tests based on literature parameters. In the best case, 92.5% of the predicted blood glucose level values were in the range of clinically acceptable errors, which means around 2.8 mmol/l root mean square error. The results of the validation based on outpatient data are promising compared to others found in the literature. Handling other important factors such as physical activity and stress remains a challenge for future research.
1 January 2022 in Journal of Biomedical Informatics
Public healthcare is a complex domain with many actors and highly variable protocols, which makes traditional process mining tools less effective and calls for specialized methods. The objective of the work was to develop a generally applicable process mining methodology to explore care processes related to diseases. The proposed methodology called Process Mining Methodology for Exploring Disease-specific Care Processes (MEDCP) is based on a systematic, step-wise refinement of the raw event logs by using such a multi-level expert taxonomy of events that encapsulates the professional concepts of the analysis. A treatment process is defined according to domain-specific rules to identify the starting (index) and closing events. Concepts from various levels of the taxonomy support the final process definition for an analysis that can deliver meaningful conclusions for domain experts. The applicability of the methodology was demonstrated on two case studies in the cardiological and oncological care domains, in the public health care system in Hungary over a period of ten years. Thanks to the multi-level taxonomy, these studies successfully identified the most important high-level event sequence patterns and some key anomalies in the national care system, such as the significantly different behavior of low-volume vs. high volume care providers in the oncology study or the geographically connected, homogeneous clusters of providers with similar care spectra in the cardiology study. As the case studies showed, the proposed methodology can improve the efficiency of standard process mining methods, and deliver high level conclusions that are easy to interpret by domain experts. System-level insight into health care processes can serve as a basis for the optimisation and long-term planning of the whole care system.
The study presented in this paper aimed to assess the effect of an Information Technology enabled community gardening program for older adults, developed by an international consortium. We have executed a quantitative, pre- and post-test field trial with older adult volunteers to test the proposed programme in two European countries, Italy and Belgium (n=98). We used standardized and ad hoc questionnaires to measure changes in the volunteers' mental and psychological state during the trial. The statistical data analysis sought for differences in the pre- and post-test values of the key scores related to the perceived quality of life and benefits of gardening via paired-samples t-tests, and also tried to identify the important factors of significant changes via logistic regression. We found significant improvements in the perceived benefits of gardening and also in the scores computed from the WHO Quality of Life instruments, especially in the social sub-domains. The improvements were associated with the country, age, marital state and education of the volunteers. Higher age or being widow, divorced or single increased the odds of a significant improvement in the scores in more than one sub-domains. Though the two trial settings were different in some aspects, the observed significant improvements generally confirmed the positive effects of gardening concerning the perceived quality of life and benefits of gardening.