Neural Style Transfer as Data Augmentation for Improving COVID-19 Diagnosis Classification.

Coronavirus disease 2019 (COVID-19) has accounted for millions of causalities. While it affects not only individuals but also our collective healthcare and economic systems, testing is insufficient and costly hampering efforts to deal with the pandemic. Chest X-rays are routine radiographic imaging tests that are used for the diagnosis of respiratory conditions such as pneumonia and COVID-19. Convolutional neural networks have shown promise to be effective at classifying X-rays for assisting diagnosis of conditions; however, achieving robust performance demanded in most modern medical applications typically requires a large number of samples. While there exist datasets containing thousands of X-ray images of patients with healthy and pneumonia diagnoses, because COVID-19 is such a recent phenomenon, there are relatively few confirmed COVID-19 positive chest X-rays openly available to the research community. In this paper, we demonstrate the effectiveness of cycle-generative adversarial network, commonly used for neural style transfer, as a way to augment COVID-19 negative X-ray images to look like COVID-19 positive images for increasing the number of COVID-19 positive training samples. The statistical results show an increase in the mean macro f1-score over 21% on a one-tailed t score = 2.68 and p value = 0.01 to accept our alternative hypothesis for an α = 0.05 . We conclude that this approach, when used in conjunction with standard transfer learning techniques, is effective at improving the performance of COVID-19 classifiers for a variety of common convolutional neural networks.

Gait Activity Classification on Unbalanced Data from Inertial Sensors Using Shallow and Deep Learning.

Activity recognition is one of the most active areas of research in ubiquitous computing. In particular, gait activity recognition is useful to identify various risk factors in people's health that are directly related to their physical activity. One of the issues in activity recognition, and gait in particular, is that often datasets are unbalanced (i.e., the distribution of classes is not uniform), and due to this disparity, the models tend to categorize into the class with more instances. In the present study, two methods for classifying gait activities using accelerometer and gyroscope data from a large-scale public dataset were evaluated and compared. The gait activities in this dataset are: (i) going down an incline, (ii) going up an incline, (iii) walking on level ground, (iv) going down stairs, and (v) going up stairs. The proposed methods are based on conventional (shallow) and deep learning techniques. In addition, data were evaluated from three data treatments: original unbalanced data, sampled data, and augmented data. The latter was based on the generation of synthetic data according to segmented gait data. The best results were obtained with classifiers built with augmented data, with F-measure results of 0.812 (σ = 0.078) for the shallow learning approach, and of 0.927 (σ = 0.033) for the deep learning approach. In addition, the data augmentation strategy proposed to deal with the unbalanced problem resulted in increased classification performance using both techniques.

Monitoring behavioral symptoms of dementia using activity trackers.

Tertiary disease prevention for dementia focuses on improving the quality of life of the patient. The quality of life of people with dementia (PwD) and their caregivers is hampered by the presence of behavioral and psychological symptoms of dementia (BPSD), such as anxiety and depression. Non-pharmacological interventions have proved useful in dealing with these symptoms. However, while most PwD exhibit BPSD, their manifestation (in frequency, intensity and type) varies widely among patients, thus the need to personalize the intervention and its assessment. Traditionally, instruments to measure behavioral symptoms of dementia, such as NPI-NH and CMAI, are used to evaluate these interventions. We propose the use of activity trackers as a complement to monitor behavioral symptoms in dementia research. To illustrate this approach we describe a nine week Cognitive Stimulation Therapy conducted with the assistance of a social robot, in which the ten participants wore an activity tracker. We describe how data gathered from these wearables complements the assessment of traditional behavior assessment instruments with the advantage that this assessment can be conducted continuously and thus be used to tailor the intervention to each PwD.

Semi-Automated Data Labeling for Activity Recognition in Pervasive Healthcare.

Activity recognition, a key component in pervasive healthcare monitoring, relies on classification algorithms that require labeled data of individuals performing the activity of interest to train accurate models. Labeling data can be performed in a lab setting where an individual enacts the activity under controlled conditions. The ubiquity of mobile and wearable sensors allows the collection of large datasets from individuals performing activities in naturalistic conditions. Gathering accurate data labels for activity recognition is typically an expensive and time-consuming process. In this paper we present two novel approaches for semi-automated online data labeling performed by the individual executing the activity of interest. The approaches have been designed to address two of the limitations of self-annotation: (i) The burden on the user performing and annotating the activity, and (ii) the lack of accuracy due to the user labeling the data minutes or hours after the completion of an activity. The first approach is based on the recognition of subtle finger gestures performed in response to a data-labeling query. The second approach focuses on labeling activities that have an auditory manifestation and uses a classifier to have an initial estimation of the activity, and a conversational agent to ask the participant for clarification or for additional data. Both approaches are described, evaluated in controlled experiments to assess their feasibility and their advantages and limitations are discussed. Results show that while both studies have limitations, they achieve 80% to 90% precision.

Incorporating Conversational Strategies in a Social Robot to Interact with People with Dementia.

BACKGROUND: Socially assistive robots (SARs) have the potential to assist nonpharmacological interventions based on verbal communication to support the care of persons with dementia (PwDs). However, establishing verbal communication with a PwD is challenging. Thus, several authors have proposed strategies to converse with PwDs. While these strategies have proved effective at enhancing communication between PwDs and their caregivers, they have not been used or tested in the domain of human-robot interaction. OBJECTIVES: This study aimed to assess the effectiveness of incorporating conversational strategies proposed in the literature for caregivers, during PwD-robot interactions. METHODS: We conducted a total of 23 group sessions based on music and conversation therapy, where a SAR interacted with 12 PwDs (mean = 80.25 years) diagnosed with mild to moderate-stage dementia. Using a single subject research approach, we designed an AB study to assess the effectiveness of the conversational strategies in the PwD-robot interaction. Our analysis focuses on the direct communication between the PwDs and the robot, and the perceived enjoyment of PwDs. RESULTS: The number of utterances made from a PwD to the robot increased significantly when the conversational strategies were included in the robot. In addition, PwDs engaged in more sustained conversations. Additionally, PwDs enjoyed conversing with the robot Eva, as much as listening to music. These results indicate that the use of these conversational strategies is -effective at increasing the interaction between PwD and a SAR. CONCLUSIONS: PwDs who participated in the study engaged and enjoyed the interaction with the SAR. The results provide evidence of the importance of incorporating appropriate conversational strategies in SARs that support interventions for the care and social stimulation of PwDs.

Assessing empathy and managing emotions through interactions with an affective avatar.

Assistive technologies can improve the quality of life of people diagnosed with different forms of social communication disorders. We report on the design and evaluation of an affective avatar aimed at engaging the user in a social interaction with the purpose of assisting in communication therapies. A human-avatar taxonomy is proposed to assist the design of affective avatars aimed at addressing social communication disorder. The avatar was evaluated with 30 subjects to assess how effectively it conveys the desired emotion and elicits empathy from the user. Results provide evidence that users become used to the avatar after a number of interactions, and they perceive the defined behavior as being logical. The users' interactions with the avatar entail affective reactions, including the mimic emotions that users felt, and establish a preliminary ground truth about prototypic empathic interactions with avatars that is being used to train learning algorithms to support social communication disorder evaluation.

Effect of technology on aging perception.

Technology can assist older adults to maintain an active lifestyle. To better understand the effect that technology has on aging perception, we conducted two studies. In the first study, through supraliminal priming, we analyzed the effects of aging- and technology-related stimuli on age estimation. In the second study, we conducted a technological intervention with a group of elders who used four interactive devices and analyzed effects on perceived aging. Results showed that technology-related stimuli did not affect estimated age. From the second study, we generated a sociotechnical model that explains the processes connecting technology use with successful aging. We concluded that the use of technology affects aging perception, although it depends on whether the elder people have a proactive attitude toward their aging process a priori.

Detecting Anxiety States when Caring for People with Dementia.

BACKGROUND: Caregiving is a complex, stressful activity, which frequently leads to anxiety and the development of depressive disorders. Recent advances in wearable sensing allows to monitor relevant physiological data of the caregiver for detecting anxiety spans and for enacting coping strategies to reduce their anxiety when needed. OBJECTIVES: This work proposes a method to infer anxiety states of caregivers when caring for people with dementia, by using physiological data. METHODS: A model using Markov chains for detecting internal anxiety states is proposed. The model is tested with a physiological dataset gathered from a naturalistic enactment experiment with 10 participants. A visual analysis for observing anxiety states is employed. The Markov chain model is evaluated by using Inter-beat Interval (IBI) data to detect 4 internal states: "Relaxed", "Arousing", "Anxiety", and "Relaxing". RESULTS: From the visual inspection of inter-beat interval data, self-report and observation labels a total of 823 state segments were identified which contained the following states: 137 "relaxed", 91 "arousing", 410 "anxious", and 185 "relaxing". By using the average IBI value of 60 seconds segments as classification feature, the model was evaluated with a "leave one-out" cross validation with an average accuracy of 73.03 %. CONCLUSIONS: We proposed a Markov chain model for internal anxiety state detection of caregivers that care for people with dementia. The model was evaluated in a naturalistic enactment experiment with 10 participants. The resulting accuracy is comparable to previous results on stress classification.

Use and Adoption of an Assisted Cognition System to Support Therapies for People with Dementia.

The cognitive deficits in persons with dementia (PwD) can produce significant functional impairment from early stages. Although memory decline is most prominent, impairments in attention, orientation, language, reasoning, and executive functioning are also common. Dementia is also characterized by changes in personality and behavioral functioning that can be very challenging for caregivers and patients. This paper presents results on the use and adoption of an assisted cognition system to support occupational therapy to address psychological and behavioral symptoms of dementia. During 16 weeks, we conducted an in situ evaluation with two caregiver-PwD dyads to assess the adoption and effectiveness of the system to ameliorate challenging behaviors and reducing caregiver burden. Evaluation results indicate that intervention personalization and a touch-based interface encouraged the adoption of the system, helping reduce challenging behaviors in PwD and caregiver burden.

Naturalistic Enactment to Elicit and Recognize Caregiver State Anxiety.

Caring for people with dementia imposes significant stress on family members and caregivers. Often, these informal caregivers have no coping strategy to deal with these behaviors. Anxiety and stress episodes are often triggered by problematic behaviors exhibited by the person who suffers from dementia. Detecting these behaviors could help them in dealing with them and reduce caregiver burden. However, work on anxiety detection using physiological signals has mostly been done under controlled conditions. In this paper we describe an experiment aimed at inducing anxiety among caregivers of people with dementia under naturalistic conditions. We report an experiment, using the naturalistic enactment technique, in which 10 subjects were asked to care for an older adult who acts as if she experiences dementia. We record physiological signals from the participants (GSR, HR, EEG) during the sessions that lasted for approximately 30 min. We explain how we obtained ground truth from self-report and observation data. We conducted two different tests using the Support Vector Machine technique. We obtained an average precision of 77.8 % and 38.1 % recall when classifying two different possible states: "Anxious" and "Not anxious". Analysis of the data provides evidence that the experiment elicits state anxiety and that it can be detected using wearable sensors. Furthermore, if episodes of problematic behaviors can also be detected, the recognition of anxiety in the caregiver can be improved, leading to the enactment of appropriate interventions to help caregivers cope with anxiety episodes.

Using Mixed Methods in Health Information Technology Evaluation.

With the increasing adoption of interactive systems in healthcare, there is a need to ensure that the benefits of such systems are formally evaluated. Traditionally quantitative research approaches have been used to gather evidence on measurable outcomes of health technology. Qualitative approaches have also been used to analyze how or why particular interventions did or did not work in specific healthcare contexts. Mixed methods research provides a framework for carrying out both quantitative and qualitative approaches within a single research study. In this paper an international group of four informatics scholars illustrate some of the benefits and challenges of using mixed methods in evaluation. The diversity of the research experience provides a broad overview of approaches in combining robust analysis of outcome data with qualitative methods that provide an understanding of the processes through which, and the contexts in which, those outcomes are achieved. This paper discussed the benefits that mixed methods brought to each study.

Healthcare Engineering Defined: A White Paper.

Engineering has been playing an important role in serving and advancing healthcare. The term "Healthcare Engineering" has been used by professional societies, universities, scientific authors, and the healthcare industry for decades. However, the definition of "Healthcare Engineering" remains ambiguous. The purpose of this position paper is to present a definition of Healthcare Engineering as an academic discipline, an area of research, a field of specialty, and a profession. Healthcare Engineering is defined in terms of what it is, who performs it, where it is performed, and how it is performed, including its purpose, scope, topics, synergy, education/training, contributions, and prospects.

Dashboards for improving patient care: review of the literature.

AIM: This review aimed to provide a comprehensive overview of the current state of evidence for the use of clinical and quality dashboards in health care environments. METHODS: A literature search was performed for the dates 1996-2012 on CINAHL, Medline, Embase, Cochrane Library, PsychInfo, Science Direct and ACM Digital Library. A citation search and a hand search of relevant papers were also conducted. RESULTS: One hundred and twenty two full text papers were retrieved of which 11 were included in the review. There was considerable heterogeneity in implementation setting, dashboard users and indicators used. There was evidence that in contexts where dashboards were easily accessible to clinicians (such as in the form of a screen saver) their use was associated with improved care processes and patient outcomes. CONCLUSION: There is some evidence that implementing clinical and/or quality dashboards that provide immediate access to information for clinicians can improve adherence to quality guidelines and may help improve patient outcomes. However, further high quality detailed research studies need to be conducted to obtain evidence of their efficacy and establish guidelines for their design.

Effects of communication media choice on the quality and efficacy of emergency calls assisted by a mobile nursing protocol tool.

The transition from paper to electronic-based records in the healthcare industry has posed several challenges to conventional medical practices. The introduction of technology in day-to-day medical and nursing practices deserves careful consideration. In this work, we report the results of a controlled experiment to compare nurses' consultation in emergency calls in six different conditions. We studied the effect that the type of communication media (face-to-face, telephone, videoconference) and type of nursing protocol media (paper-based, electronic-based) can have on consultation time, mistakes made, pauses during consultation, eye contact, and efficacy of the consultation. We found that the type of communication media has an effect on consultation time; on average, fewer mistakes were made during telephone-based consultations; for eye contact, there were significantly fewer eye contacts during face-to-face than during videoconference consultations; finally, the type of communication media or protocol media did not have any effect in the efficacy of the consultation.

Intervention tailoring in augmented cognition systems for elders with dementia.

We present an approach for personalizing nonpharmacological interventions for people with dementia (PwD) using ontologies. We conducted two case studies to derive an ontological model to personalize the planning and execution of interventions to address problematic behaviors. The paper describes how the ontology was derived, and illustrates how it is used to tailor an ambient-assisted intervention system (AAIS) at two stages: first, to decide on the services that the AAIS will offer the PwD, and then to adapt these services at runtime using contextual information. The results of a deployment of an AAIS during 2 months in the home of a PwD, indicate that the AAIS successfully addressed some of the problematic behaviors exhibited by the PwD, helping to reduce the burden of the caregiver.

International priorities for research in nursing informatics for patient care.

The Nursing Informatics International Research Network (NIIRN) is a group of experts who are collaborating on the development of internationally relevant research programs for nursing informatics. In this paper we outline key findings of a survey exploring international research priorities for nursing informatics. The survey was available online during May-August 2012. Respondents were asked to rate each of 20 listed research topics in terms of respondent's views of its priority for nursing informatics research. 468 completed surveys were received representing respondents from six World Health Organization regions. The two most highly ranked areas of importance for research were development of systems to provide real time feedback to nurses and assessment of the impact of HIT on nursing care and patient outcomes. The lowest ranked research topics were theory development and integrating genomic data into clinical information systems. The identification of these priorities provides a basis for future international collaborative research in the field of nursing informatics.

Nurse home visits with or without alert buttons versus usual care in the frail elderly: a randomized controlled trial.

OBJECTIVE: To assess whether an intervention based on nurse home visits including alert buttons (NV+AB) is effective in reducing frailty compared to nurse home visits alone (NV-only) and usual care (control group) for older adults. DESIGN: Unblinded, randomized, controlled trial. SETTING: Insured population covered by the Mexican Social Security Institute living in the city of Ensenada, Baja California, Mexico. PARTICIPANTS: Patients were aged over 60 years with a frailty index score higher than 0.14. INTERVENTION: After screening and informed consent, participants were allocated randomly to the control, NV+AB, or NV-only groups. MEASUREMENTS: The primary outcome was the frailty score 9 months later. Quality of life, depression, comorbidities, health status, and health service utilization were also considered. RESULTS: The framing sample included 819 patients. Of those, 591 were not located because they did not have a landline/telephone (341 patients), they had died (107), they were ill (50), or they were not currently living in the city (28). A screening interview was applied to 228 participants, and 57 had a score ≤0.14, 171 had ≥0.14, and 16 refused to complete the baseline questionnaire. A home visit was scheduled for 155 patients. However, 22 did not complete the baseline questionnaire. The final 133 subjects were randomized into the NV+AB (n = 45), NV-only (n = 44), and control (n = 44) groups. There were no statistically significant differences in the baseline characteristics of the groups. The mean age overall was 76.3 years (standard deviation 4.7) and 45% were men. At the baseline, 61.65% were classified as frail. At end of follow-up the adjusted prevalence of frailty in NV+AB group was 23.3% versus 58.3% in the control group. CONCLUSION: An intervention based on NV+AB seems to have a positive effect on frailty scores.

Understanding and supporting lightweight communication in hospital work.

Informal communication is an essential resource in hospital work; it is used as a means to collaborate and coordinate the way in which work is performed, as well as to locate and gather the artifacts and human resources required for patient care. The need of physical proximity to establish and hold informal communications has motivated the development of tools that support remote informal interaction. However, this kind of technology has not been widely adopted in hospitals, where workers experience intense mobility. In this paper, we present the results of an observational study in a hospital aimed at understanding how local mobility changes the rules in the provision of support for informal communication, and how technology could improve this form of communication. We found that local mobility fosters opportunities for colocated communication; however, it faces some inconveniences related to the affordances of the physical environment. We address these issues through the design of SOLAR, a collaborative application that supports colocated interactions in hospital work through the implementation and integration of five services that form a ubicomp infrastructure. Additionally, we present a discussion about the perception of users related to the usefulness of this tool.

Mobility in hospital work: towards a pervasive computing hospital environment.

Handheld computers are increasingly being used by hospital workers. With the integration of wireless networks into hospital information systems, handheld computers can provide the basis for a pervasive computing hospital environment; to develop this designers need empirical information to understand how hospital workers interact with information while moving around. To characterise the medical phenomena we report the results of a workplace study conducted in a hospital. We found that individuals spend about half of their time at their base location, where most of their interactions occur. On average, our informants spent 23% of their time performing information management tasks, followed by coordination (17.08%), clinical case assessment (15.35%) and direct patient care (12.6%). We discuss how our results offer insights for the design of pervasive computing technology, and directions for further research and development in this field such as transferring information between heterogeneous devices and integration of the physical and digital domains.

Location-aware access to hospital information and services.

Hospital workers are highly mobile; they are constantly changing location to perform their daily work, which includes visiting patients, locating resources, such as medical records, or consulting with other specialists. The information required by these specialists is highly dependent on their location. Access to a patient's laboratory results might be more relevant when the physician is near the patient's bed and not elsewhere. We describe a location-aware medical information system that was developed to provide access to resources such as patient's records or the location of a medical specialist, based on the user's location. The system is based on a handheld computer which includes a trained backpropagation neural-network used to estimate the user's location and a client to access information from the hospital information system that is relevant to the user's current location.