Background and Objectives Smart healthcare systems use technologies such as wearable devices, Internet of Medical Things and mobile internet technologies to dynamically access health information, connect patients to health professionals and health institutions, and to actively manage and respond intelligently to the medical ecosystem's needs. However, smart healthcare systems are affected by many challenges in their implementation and maintenance. Key among these are ensuring the security and privacy of patient health information. To address this challenge, several mitigation measures have been proposed and some have been implemented. Techniques that have been used include data encryption and biometric access. In addition, blockchain is an emerging security technology that is expected to address the security issues due to its distributed and decentralized architecture which is similar to that of smart healthcare systems. This study reviewed articles that identified security requirements and risks, proposed potential solutions, and explained the effectiveness of these solutions in addressing security problems in smart healthcare systems. Methods This review adhered to the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guidelines and was framed using the Problem, Intervention, Comparator, and Outcome (PICO) approach to investigate and analyse the concepts of interest. However, the comparator is not applicable because this review focuses on the security measures available and in this case no comparable solutions were considered since the concept of smart healthcare systems is an emerging one and there are therefore, no existing security solutions that have been used before. The search strategy involved the identification of studies from several databases including the Cumulative Index of Nursing and Allied Health Literature (CINAL), Scopus, PubMed, Web of Science, Medline, Excerpta Medical database (EMBASE), Ebscohost and the Cochrane Library for articles that focused on the security for smart healthcare systems. The selection process involved removing duplicate studies, and excluding studies after reading the titles, abstracts, and full texts. Studies whose records could not be retrieved using a predefined selection criterion for inclusion and exclusion were excluded. The remaining articles were then screened for eligibility. A data extraction form was used to capture details of the screened studies after reading the full text. Of the searched databases, only three yielded results when the search strategy was applied, i.e., Scopus, Web of science and Medline, giving a total of 1742 articles. 436 duplicate studies were removed. Of the remaining articles, 801 were excluded after reading the title, after which 342 after were excluded after reading the abstract, leaving 163, of which 4 studies could not be retrieved. 159 articles were therefore screened for eligibility after reading the full text. Of these, 14 studies were included for detailed review using the formulated research questions and the PICO framework. Each of the 14 included articles presented a description of a smart healthcare system and identified the security requirements, risks and solutions to mitigate the risks. Each article also summarized the effectiveness of the proposed security solution. Results The key security requirements reported were data confidentiality, integrity and availability of data within the system, with authorisation and authentication used to support these key security requirements. The identified security risks include loss of data confidentiality due to eavesdropping in wireless communication mediums, authentication vulnerabilities in user devices and storage servers, data fabrication and message modification attacks during transmission as well as while the data is at rest in databases and other storage devices. The proposed mitigation measures included the use of biometric accessing devices; data encryption for protecting the confidentiality and integrity of data; blockchain technology to address confidentiality, integrity, and availability of data; network slicing techniques to provide isolation of patient health data in 5G mobile systems; and multi-factor authentication when accessing IoT devices, servers, and other components of the smart healthcare systems. The effectiveness of the proposed solutions was demonstrated through their ability to provide a high level of data security in smart healthcare systems. For example, proposed encryption algorithms demonstrated better energy efficiency, and improved operational speed; reduced computational overhead, better scalability, efficiency in data processing, and better ease of deployment. Conclusion This systematic review has shown that the use of blockchain technology, biometrics (fingerprints), data encryption techniques, multifactor authentication and network slicing in the case of 5G smart healthcare systems has the potential to alleviate possible security risks in smart healthcare systems. The benefits of these solutions include a high level of security and privacy for Electronic Health Records (EHRs) systems; improved speed of data transaction without the need for a decentralized third party, enabled by the use of blockchain. However, the proposed solutions do not address data protection in cases where an intruder has already accessed the system. This may be potential avenues for further research and inquiry.

<p>Wearable devices and their ubiquitous use and deployment across multiple areas of health provide key insights in patient and individual status via big data through sensor capture at key parts of the individual’s body. While small and low cost, their limitations rest in their computational and battery capacity. One key use of wearables has been in individual activity capture. For accelerometer and gyroscope data, oscillatory patterns exist between daily activities that users may perform. By leveraging spatial and temporal learning via CNN and LSTM layers to capture both the intra and inter-oscillatory patterns that appear during these activities, we deployed data sparsification via autoencoders to extract the key topological properties from the data and transmit via BLE that compressed data to a central device for later decoding and analysis. Several autoencoder designs were developed to determine the principles of system design that compared encoding overhead on the sensor device with signal reconstruction accuracy. By leveraging asymmetric autoencoder design, we were able to offshore much of the computational and power cost of signal reconstruction from the wearable to the central devices, while still providing robust reconstruction accuracy at several compression efficiencies. Via our high-precision Bluetooth voltmeter, the integrated sparsified data transmission configuration was tested for all quantization and compression efficiencies, generating lower power consumption to the setup without data sparsification for all autoencoder configurations. </p> <p><br></p> <p>Human activity recognition (HAR) is a key facet of lifestyle and health monitoring. Effective HAR classification mechanisms and tools can provide healthcare professionals, patients, and individuals key insights into activity levels and behaviors without the intrusive use of human or camera observation. We leverage both spatial and temporal learning mechanisms via CNN and LSTM integrated architectures to derive an optimal classification architecture that provides robust classification performance for raw activity inputs and determine that a LSTMCNN utilizing a stacked-bidirectional LSTM layer provides superior classification performance to the CNNLSTM (also utilizing a stacked-bidirectional LSTM) at all input widths. All inertial data classification frameworks are based off sensor data drawn from wearable devices placed at key sections of the body. With the limitation of wearable devices being a lack of computational and battery power, data compression techniques to limit the quantity of transmitted data and reduce the on-board power consumption have been employed. While this compression methodology has been shown to reduce overall device power consumption, this comes at a cost of more-or-less information loss in the reconstructed signals. By employing an asymmetric autoencoder design and training the LSTMCNN classifier with the reconstructed inputs, we minimized the classification performance degradation due to the wearable signal reconstruction error The classifier is further trained on the autoencoder for several input widths and with quantized and unquantized models. The performance for the classifier trained on reconstructed data ranged between 93.0\% and 86.5\% accuracy dependent on input width and autoencoder quantization, showing promising potential of deep learning with wearable sparsification. </p>

Submitted by Silvana Teresinha Dornelles Studzinski (sstudzinski) on 2016-12-21T15:44:57Z No. of bitstreams: 1 Rodrigo Silva Pereira_.pdf: 951778 bytes, checksum: 90c6af826318df7c8204565678dff935 (MD5) / Made available in DSpace on 2016-12-21T15:44:57Z (GMT). No. of bitstreams: 1 Rodrigo Silva Pereira_.pdf: 951778 bytes, checksum: 90c6af826318df7c8204565678dff935 (MD5) Previous issue date: 2016-08-28 / Nenhuma / Com a emergência do crowdsourcing junto a difusão mundial de smartphones esforços recentes e pesquisas importantes sobre o uso de crowdsourcing na área da saúde ou ainda smarthealth visam auxiliar na melhoria hábitos de saúde, construção de históricos médicos pessoais de longo prazo, análise e revisão de dados médica, controle de dietas alimentares, gerenciamento do estresse, analise e comparação de informações e assistência em tempo real para catástrofes. Porém, nenhum deles usou de crowdsourcing para recomendação de centros clínicos de saúde. Segundo Chatzimilioudis crowdsourcing refere-se "a um modelo distribuído de solução de problemas em que uma multidão de tamanho indefinido é contratada para resolver um problema complexo através de um convite aberto". Neste âmbito, este trabalho apresenta um modelo de sistema de recomendação de centros clínicos de saúde, chamado CrowdHealth. A principal contribuição do modelo de sistema de recomendação de centros clínicos é possibilitar a criação de uma relação ganha-ganha entre seus usuários que podem ser cidadãos, médicos ou ainda entidades ligadas ao governo. Na literatura encontramos alguns trabalhos que carecem a abordagem do uso de crowdsourcing como fonte de dados para recomendação de centros clínicos de saúde. Nós desenvolvemos um protótipo de aplicação baseada no modelo de sistema de recomendação de centros clínicos de saúde para proporcionar uma visão do que seria uma aplicação baseada no modelo de sistema de recomendação de centros clínicos de saúde. Para avaliar o nosso modelo, apresentamos um cenário hipotético baseado numa possível aplicação para mensurar a percepção dos usuários quanto a utilidade dos centros clínicos de saúde. Os cenários descritos levavam em consideração os seguintes critérios: (1) a distância entre do usuário ao centro clinico, (2) a avaliação dos usuários em relação ao atendimento recebido nos centros clínicos e (3) o tempo de atendimento informado pelos usuários. Desta forma realizamos uma simulação de requisições de recomendações de usuários usando um dataset real contendo informações do Foursquare. O arquivo do dataset possuia 227428 check-in’s na cidade de Nova Iorque, EUA. O arquivo, foi dividido em duas partes, onde a primeira representava os check-in’s realizados pelos usuários nos centros clínicos, e a segunda representava usuários requisitando por recomendações de centros clínicos em outros locais. Assim, foram criadas funções para simular os processos de cálculo do tempo de atendimento e avaliação dos centros clínicos por parte dos usuários. Também simulou-se usuários requisitando por recomendações de centros clínicos em outros locais. Então, medimos precisão e recuperação dos centros clínicos de saúde sugeridos para cada usuário. Obtivemos valores médios de 57,5% e 61,33% para precisão e recuperação, respectivamente. Com isso, nossa avaliação retrata que centros clínicos de saúde recomendados por uma aplicação baseada no CrowdHealth poderiam aumentar beneficamente a utilidade de centros clínicos de saúde recomendados para os usuários. / With the emergence of crowdsourcing with the worldwide spread of smartphones recent efforts and important research on the use of crowdsourcing in health or smart-health are intended to assist in improving health habits, construction of historical long-term medical personnel, medical analysis and data review, control diets, stress management, analysis and comparison of information and real-time assistance for disasters. However, none of them used the crowdsourcing for recommendation clinical health centers. In this context, this paper presents a model of clinical health centers recommendation system called CrowdHealth. The main contribution of clinical health centers recommendation system model is possible to create a win-win relationship between its users that can be citizens, doctors or entities linked to the government. In the literature we find some papers that require the use of crowdsourcing as a data source for recommendation clinical health centers approach. We have developed a prototype application based on clinical health centers recommendation system model to provide a vision of what would be an application based on the clinical health centers recommendation system model. To evaluate our model, we present a hypothetical scenario based on a possible application to measure the perception of users and the utility of clinical health centers. The scenarios described took into consideration the following criteria: (1) the distance from the user to the clinical center, (2) the evaluation of other users on the service received in the clinical centers and (3) the time of service reported by users. Thus we performed a simulation of user requests recommendations using a real dataset containing information of Foursquare. The file dataset haved 227428 check in’s in New York City, USA. The file was divided into two parts, where the first represented the textit check in ’s performed by users in clinical centers, and the second represented by requesting users polyclinics recommendations elsewhere. Thus, functions were created to simulate service time calculation and evaluation processes of polyclinics by users. Also users was simulated by ordering polyclinics recommendations elsewhere. So we measure precision and recall of health clinical centers suggested for each user. Average values obtained from 57.5 % and 61.33 % for precision and recall, respectively. Thus, our assessment that portrays clinical health centers recommended by an application based on CrowdHealth could increase beneficially the usefulness of clinical health centers recommended for users.

Health monitoring devices, such as Fitbit, are gaining popularity both as wellness tools and as a source of information for healthcare decisions. Predicting such wellness goals accurately is critical for the users to make informed lifestyle choices. The core objective of this thesis is to design and implement such a system that takes energy consumption and privacy into account. This research is modelled as a time-series forecasting problem that makes use of Spiking Neural Networks (SNNs) due to their proven energy-saving capabilities. Thanks to their design that closely mimics natural neural networks (such as the brain), SNNs have the potential to significantly outperform classic Artificial Neural Networks in terms of energy consumption and robustness. In order to prove our hypotheses, a previous research by Sonia et al. [1] in the same domain and with the same dataset is used as our starting point, where a private forecasting system using Long short-term memory (LSTM) is designed and implemented. Their study also implements and evaluates a clustering federated learning approach, which fits well the highly distributed data. The results obtained in their research act as a baseline to compare our results in terms of accuracy, training time, model size and estimated energy consumed. Our experiments show that Spiking Neural Networks trades off accuracy (2.19x, 1.19x, 4.13x, 1.16x greater Root Mean Square Error (RMSE) for macronutrients, calories burned, resting heart rate, and active minutes respectively), to grant a smaller model (19% less parameters an 77% lighter in memory) and a 43% faster training. Our model is estimated to consume 3.36μJ per inference, which is much lighter than traditional Artificial Neural Networks (ANNs) [2]. The data recorded by health monitoring devices is vastly distributed in the real-world. Moreover, with such sensitive recorded information, there are many possible implications to consider. For these reasons, we apply the clustering federated learning implementation [1] to our use-case. However, it can be challenging to adopt such techniques since it can be difficult to learn from data sequences that are non-regular. We use a two-step streaming clustering approach to classify customers based on their eating and exercise habits. It has been shown that training different models for each group of users is useful, particularly in terms of training time; however this is strongly dependent on the cluster size. Our experiments conclude that there is a decrease in error and training time if the clusters contain enough data to train the models. Finally, this study addresses the issue of data privacy by using state of-the-art differential privacy. We apply e-differential privacy to both our baseline model (trained on the whole dataset) and our federated learning based approach. With a differential privacy of ∈= 0.1 our experiments report an increase in the measured average error (RMSE) of only 25%. Specifically, +23.13%, 25.71%, +29.87%, 21.57% for macronutrients (grams), calories burned (kCal), resting heart rate (beats per minute (bpm), and minutes (minutes) respectively. / Hälsoövervakningsenheter, som Fitbit, blir allt populärare både som friskvårdsverktyg och som informationskälla för vårdbeslut. Att förutsäga sådana välbefinnandemål korrekt är avgörande för att användarna ska kunna göra välgrundade livsstilsval. Kärnmålet med denna avhandling är att designa och implementera ett sådant system som tar hänsyn till energiförbrukning och integritet. Denna forskning är modellerad som ett tidsserieprognosproblem som använder sig av SNNs på grund av deras bevisade energibesparingsförmåga. Tack vare deras design som nära efterliknar naturliga neurala nätverk (som hjärnan) har SNNs potentialen att avsevärt överträffa klassiska artificiella neurala nätverk när det gäller energiförbrukning och robusthet. För att bevisa våra hypoteser har en tidigare forskning av Sonia et al. [1] i samma domän och med samma dataset används som utgångspunkt, där ett privat prognossystem som använder LSTM designas och implementeras. Deras studie implementerar och utvärderar också en klustringsstrategi för federerad inlärning, som passar väl in på den mycket distribuerade data. Resultaten som erhållits i deras forskning fungerar som en baslinje för att jämföra våra resultat vad gäller noggrannhet, träningstid, modellstorlek och uppskattad energiförbrukning. Våra experiment visar att Spiking Neural Networks byter ut precision (2,19x, 1,19x, 4,13x, 1,16x större RMSE för makronäringsämnen, förbrända kalorier, vilopuls respektive aktiva minuter), för att ge en mindre modell ( 19% mindre parametrar, 77% lättare i minnet) och 43% snabbare träning. Vår modell beräknas förbruka 3, 36μJ, vilket är mycket lättare än traditionella ANNs [2]. Data som registreras av hälsoövervakningsenheter är enormt spridda i den verkliga världen. Dessutom, med sådan känslig registrerad information finns det många möjliga konsekvenser att överväga. Av dessa skäl tillämpar vi klustringsimplementeringen för federerad inlärning [1] på vårt användningsfall. Det kan dock vara utmanande att använda sådana tekniker eftersom det kan vara svårt att lära sig av datasekvenser som är oregelbundna. Vi använder en tvåstegs streaming-klustringsmetod för att klassificera kunder baserat på deras mat- och träningsvanor. Det har visat sig att det är användbart att träna olika modeller för varje grupp av användare, särskilt när det gäller utbildningstid; detta är dock starkt beroende av klustrets storlek. Våra experiment drar slutsatsen att det finns en minskning av fel och träningstid om klustren innehåller tillräckligt med data för att träna modellerna. Slutligen tar denna studie upp frågan om datasekretess genom att använda den senaste differentiell integritet. Vi tillämpar e-differentiell integritet på både vår baslinjemodell (utbildad på hela datasetet) och vår federerade inlärningsbaserade metod. Med en differentiell integritet på ∈= 0.1 rapporterar våra experiment en ökning av det uppmätta medelfelet (RMSE) på endast 25%. Specifikt +23,13%, 25,71%, +29,87%, 21,57% för makronäringsämnen (gram), förbrända kalorier (kCal), vilopuls (bpm och minuter (minuter).

The development of internet-based technologies, the design and adoption of wireless wearable and smart devices have been a growing study spot in all domains. The healthcare sector as many others is making technological progress to improve healthcare services and patients wellbeing and avoid or minimize the use of manual and traditional practices such as the use of paper notes to record the vital signs parameters data. The vital signs parameters are the most monitored physiology features, they produce a big amount of data and request a close follow up to define the health condition of a patient. Continuous vital signs monitoring involves the usage of different devices and systems, which if appropriate positively impact the activities involved, by enabling the continuous generation of data and information about the overall health status of patients and contribute to the wellbeing of individuals, in terms of preventing and reducing fatal risks. To investigate this situation, this research’s focus was in three parts; first, investigate recent research about patient’s health predictions based on vital signs parameters and the impacts of continuous monitoring on the care given. Second, explore the availability in terms of i.e. sensors used in devices that can continuously track vital signs parameters. Last, to provide a possible design recommendation to improve and/or replace the existing devices for vital signs parameters measuring and monitoring in emergency and post-operative care. A qualitative approach and participatory design approach were used to collect data. The qualitative part was achieved through interviews and the participatory design part was accomplished by the future workshop and two prototyping techniques, paper and digital prototypes. The findings of this research were analysed using conceptual analysis, and also discussed using those concepts. Together with the participants, this research resulted in three design suggestions which if implemented shall improve the vital signs continuous monitoring activities, by facilitating the healthcare professionals in their clinical responsibilities and improving the patients wellbeing while admitted in Emergency and Post-operative wards.