Matching IoT Devices to the Fog Service Providers: A Mechanism Design Perspective.

In the Internet of Things (IoT) + Fog + Cloud architecture, with the unprecedented growth of IoT devices, one of the challenging issues that needs to be tackled is to allocate Fog service providers (FSPs) to IoT devices, especially in a game-theoretic environment. Here, the issue of allocation of FSPs to the IoT devices is sifted with game-theoretic idea so that utility maximizing agents may be benign. In this scenario, we have multiple IoT devices and multiple FSPs, and the IoT devices give preference ordering over the subset of FSPs. Given such a scenario, the goal is to allocate at most one FSP to each of the IoT devices. We propose mechanisms based on the theory of mechanism design without money to allocate FSPs to the IoT devices. The proposed mechanisms have been designed in a flexible manner to address the long and short duration access of the FSPs to the IoT devices. For analytical results, we have proved the economic robustness, and probabilistic analyses have been carried out for allocation of IoT devices to the FSPs. In simulation, mechanism efficiency is laid out under different scenarios with an implementation in Python.

Developing a survey instrument to assess the readiness of primary care data, genetic and disease registries to conduct linked research: TRANSFoRm International Research Readiness (TIRRE) survey instrument.

BACKGROUND: Clinical data are collected for routine care in family practice; there are also a growing number of genetic and cancer registry data repositories. The Translational Research and Patient Safety in Europe (TRANSFoRm) project seeks to facilitate research using linked data from more than one source. We performed a requirements analysis which identified a wide range of data and business process requirements that need to be met before linking primary care and either genetic or disease registry data. OBJECTIVES: To develop a survey to assess the readiness of data repositories to participate in linked research - the Transform International Research Readiness (TIRRE) survey. METHOD: We develop the questionnaire based on our requirement analysis; with questions at micro-, meso- and macro levels of granularity, study-specific questions about diabetes and gastro-oesophageal reflux disease (GORD), and research track record. The scope of the data required was extensive. We piloted this instrument, conducting ten preliminary telephone interviews to evaluate the response to the questionnaire. RESULTS: Using feedback gained from these interviews we revised the questionnaire; clarifying questions that were difficult to answer and utilising skip logic to create different series of questions for the various types of data repository. We simplified the questionnaire replacing free-text responses with yes/no or picking list options, wherever possible. We placed the final questionnaire online and encouraged its use (www.clininf.eu/jointirre/info.html). CONCLUSION: Limited field testing suggests that TIRRE is capable of collecting comprehensive and relevant data about the suitability and readiness of data repositories to participate in linked data research.

Conducting requirements analyses for research using routinely collected health data: a model driven approach.

BACKGROUND: Medical research increasingly requires the linkage of data from different sources. Conducting a requirements analysis for a new application is an established part of software engineering, but rarely reported in the biomedical literature; and no generic approaches have been published as to how to link heterogeneous health data. METHODS: Literature review, followed by a consensus process to define how requirements for research, using, multiple data sources might be modeled. RESULTS: We have developed a requirements analysis: i-ScheDULEs - The first components of the modeling process are indexing and create a rich picture of the research study. Secondly, we developed a series of reference models of progressive complexity: Data flow diagrams (DFD) to define data requirements; unified modeling language (UML) use case diagrams to capture study specific and governance requirements; and finally, business process models, using business process modeling notation (BPMN). DISCUSSION: These requirements and their associated models should become part of research study protocols.

A Business Analytics Framework for Primary Care.

In this paper, we present a Business Analytics (BA) framework, which addresses the challenge of analysing primary care outcomes for both patients and clinicians from multiple data sources in an accurate manner. A review of the process monitoring literature has been conducted in the context of healthcare management and decision making and its findings have informed the formulation of a BA conceptual framework for process monitoring and decision support in primary care. Furthermore, a real case study is conducted to demonstrate the application of the BA framework to implement a BA dashboard tool within one of the largest primary care providers in England. Findings: The main contributions of the presented work are the development of a conceptual BA framework and a BA dashboard tool to support management and decision making in primary care. This was evaluated through a case study of the implementation of the BA dashboard tool in London's largest primary care provider. This BA tool provides real-time information to enable simpler decision-making processes and to inform business transformation in a number of areas. The resulting increased efficiency has led to significant cost savings and improved delivery of patient care.

Learning to Discover Explainable Clinical Features With Minimum Supervision.

Purpose: To compare supervised transfer learning to semisupervised learning for their ability to learn in-depth knowledge with limited data in the optical coherence tomography (OCT) domain. Methods: Transfer learning with EfficientNet-B4 and semisupervised learning with SimCLR are used in this work. The largest public OCT dataset, consisting of 108,312 images and four categories (choroidal neovascularization, diabetic macular edema, drusen, and normal) is used. In addition, two smaller datasets are constructed, containing 31,200 images for the limited version and 4000 for the mini version of the dataset. To illustrate the effectiveness of the developed models, local interpretable model-agnostic explanations and class activation maps are used as explainability techniques. Results: The proposed transfer learning approach using the EfficientNet-B4 model trained on the limited dataset achieves an accuracy of 0.976 (95% confidence interval [CI], 0.963, 0.983), sensitivity of 0.973 and specificity of 0.991. The semisupervised based solution with SimCLR using 10% labeled data and the limited dataset performs with an accuracy of 0.946 (95% CI, 0.932, 0.960), sensitivity of 0.941, and specificity of 0.983. Conclusions: Semisupervised learning has a huge potential for datasets that contain both labeled and unlabeled inputs, generally, with a significantly smaller number of labeled samples. The semisupervised based solution provided with merely 10% labeled data achieves very similar performance to the supervised transfer learning that uses 100% labeled samples. Translational Relevance: Semisupervised learning enables building performant models while requiring less expertise effort and time by using to good advantage the abundant amount of available unlabeled data along with the labeled samples.

Automated feature-based grading and progression analysis of diabetic retinopathy.

BACKGROUND: In diabetic retinopathy (DR) screening programmes feature-based grading guidelines are used by human graders. However, recent deep learning approaches have focused on end to end learning, based on labelled data at the whole image level. Most predictions from such software offer a direct grading output without information about the retinal features responsible for the grade. In this work, we demonstrate a feature based retinal image analysis system, which aims to support flexible grading and monitor progression. METHODS: The system was evaluated against images that had been graded according to two different grading systems; The International Clinical Diabetic Retinopathy and Diabetic Macular Oedema Severity Scale and the UK's National Screening Committee guidelines. RESULTS: External evaluation on large datasets collected from three nations (Kenya, Saudi Arabia and China) was carried out. On a DR referable level, sensitivity did not vary significantly between different DR grading schemes (91.2-94.2.0%) and there were excellent specificity values above 93% in all image sets. More importantly, no cases of severe non-proliferative DR, proliferative DR or DMO were missed. CONCLUSIONS: We demonstrate the potential of an AI feature-based DR grading system that is not constrained to any specific grading scheme.

The Hayes principles: learning from the national pilot of information technology and core generalisable theory in informatics.

BACKGROUND: There has been much criticism of the NHS national programme for information technology (IT); it has been an expensive programme and some elements appear to have achieved little. The Hayes report was written as an independent review of health and social care IT in England. OBJECTIVE: To identify key principles for health IT implementation which may have relevance beyond the critique of NHS IT. OUTCOME: We elicit ten principles from the Hayes report, which if followed may result in more effective IT implementation in health care. They divide into patient-centred, subsidiarity and strategic principles. The patient-centred principles are: 1) the patient must be at the centre of all information systems; 2) the provision of patient-level operational data should form the foundation - avoid the dataset mentality; 3) store health data as close to the patient as possible; 4) enable the patient to take a more active role with their health data within a trusted doctor-patient relationship. The subsidiarity principles set out to balance the local and health-system-wide needs: 5) standardise centrally - patients must be able to benefit from interoperability; 6) provide a standard procurement package and an approved process that ensures safety standards and provision of interoperable systems; 7) authorise a range of local suppliers so that health providers can select the system best meeting local needs; 8) allow local migration from legacy systems, as and when improved functionality for patients is available. And finally the strategic principles: 9) evaluate health IT systems in terms of measureable benefits to patients; 10) strategic planning of systems should reflect strategic goals for the health of patients/the population. CONCLUSIONS: Had the Hayes principles been embedded within our approach to health IT, and in particular to medical record implementation, we might have avoided many of the costly mistakes with the UK national programme. However, these principles need application within the modern IT environment. Closeness to the patient must not be interpreted as physical but instead as a virtual patient-centred space; data will be secure within the cloud and we should dump the vault and infrastructure mentality. Health IT should be developed as an adaptive ecosystem.

Procuring interoperability at the expense of usability: a case study of UK National Programme for IT assurance process.

BACKGROUND: The allure of interoperable systems is that they should improve patient safety and make health services more efficient. The UK's National Programme for IT has made great strides in achieving interoperability; through linkage to a national electronic spine. However, there has been criticism of the usability of the applications in the clinical environment. METHOD: Analysis of the procurement and assurance process to explore whether they predetermine usability. RESULTS: Processes separate developers from users, and test products against theoretical assurance models of use rather than simulate or pilot in a clinical environment. The current process appears to be effective for back office systems and high risk applications, but too inflexible for developing applications for the clinical setting. CONCLUSIONS: For clinical applications agile techniques are more appropriate. Usability testing should become an integrated part of the contractual process and be introduced earlier in the development process.

A system for solution-orientated reporting of errors associated with the extraction of routinely collected clinical data for research and quality improvement.

BACKGROUND: We have used routinely collected clinical data in epidemiological and quality improvement research for over 10 years. We extract, pseudonymise and link data from heterogeneous distributed databases; inevitably encountering errors and problems. OBJECTIVE: To develop a solution-orientated system of error reporting which enables appropriate corrective action. METHOD: Review of the 94 errors, which occurred in 2008/9. Previously we had described failures in terms of the data missing from our response files; however this provided little information about causation. We therefore developed a taxonomy based on the IT component limiting data extraction. RESULTS: Our final taxonomy categorised errors as: (A) Data extraction Method and Process; (B) Translation Layer and Proxy Specification; (C) Shape and Complexity of the Original Schema; (D) Communication and System (mainly Software-based) Faults; (E) Hardware and Infrastructure; (F) Generic/Uncategorised and/or Human Errors. We found 79 distinct errors among the 94 reported; and the categories were generally predictive of the time needed to develop fixes. CONCLUSIONS: A systematic approach to errors and linking them to problem solving has improved project efficiency and enabled us to better predict any associated delays.

Evidence Based Prediction and Progression Monitoring on Retinal Images from Three Nations.

Purpose: The aim of this work is to demonstrate how a retinal image analysis system, DAPHNE, supports the optimization of diabetic retinopathy (DR) screening programs for grading color fundus photography. Method: Retinal image sets, graded by trained and certified human graders, were acquired from Saudi Arabia, China, and Kenya. Each image was subsequently analyzed by the DAPHNE automated software. The sensitivity, specificity, and positive and negative predictive values for the detection of referable DR or diabetic macular edema were evaluated, taking human grading or clinical assessment outcomes to be the gold standard. The automated software's ability to identify co-pathology and to correctly label DR lesions was also assessed. Results: In all three datasets the agreement between the automated software and human grading was between 0.84 to 0.88. Sensitivity did not vary significantly between populations (94.28%-97.1%) with specificity ranging between 90.33% to 92.12%. There were excellent negative predictive values above 93% in all image sets. The software was able to monitor DR progression between baseline and follow-up images with the changes visualized. No cases of proliferative DR or DME were missed in the referable recommendations. Conclusions: The DAPHNE automated software demonstrated its ability not only to grade images but also to reliably monitor and visualize progression. Therefore it has the potential to assist timely image analysis in patients with diabetes in varied populations and also help to discover subtle signs of sight-threatening disease onset. Translational Relevance: This article takes research on machine vision and evaluates its readiness for clinical use.

Information Theory and Medical Decision Making.

Information theory has gained application in a wide range of disciplines, including statistical inference, natural language processing, cryptography and molecular biology. However, its usage is less pronounced in medical science. In this chapter, we illustrate a number of approaches that have been taken to applying concepts from information theory to enhance medical decision making. We start with an introduction to information theory itself, and the foundational concepts of information content and entropy. We then illustrate how relative entropy can be used to identify the most informative test at a particular stage in a diagnosis. In the case of a binary outcome from a test, Shannon entropy can be used to identify the range of values of test results over which that test provides useful information about the patient's state. This, of course, is not the only method that is available, but it can provide an easily interpretable visualization. The chapter then moves on to introduce the more advanced concepts of conditional entropy and mutual information and shows how these can be used to prioritise and identify redundancies in clinical tests. Finally, we discuss the experience gained so far and conclude that there is value in providing an informed foundation for the broad application of information theory to medical decision making.

An instrument to identify computerised primary care research networks, genetic and disease registries prepared to conduct linked research: TRANSFoRm International Research Readiness (TIRRE) survey.

PURPOSE: The Translational Research and Patients safety in Europe (TRANSFoRm) project aims to integrate primary care with clinical research whilst improving patient safety. The TRANSFoRm International Research Readiness survey (TIRRE) aims to demonstrate data use through two linked data studies and by identifying clinical data repositories and genetic databases or disease registries prepared to participate in linked research. METHOD: The TIRRE survey collects data at micro-, meso- and macro-levels of granularity; to fulfil data, study specific, business, geographical and readiness requirements of potential data providers for the TRANSFoRm demonstration studies. We used descriptive statistics to differentiate between demonstration-study compliant and non-compliant repositories. We only included surveys with >70% of questions answered in our final analysis, reporting the odds ratio (OR) of positive responses associated with a demonstration-study compliant data provider. RESULTS: We contacted 531 organisations within the Eurpean Union (EU). Two declined to supply information; 56 made a valid response and a further 26 made a partial response. Of the 56 valid responses, 29 were databases of primary care data, 12 were genetic databases and 15 were cancer registries. The demonstration compliant primary care sites made 2098 positive responses compared with 268 in non-use-case compliant data sources [OR: 4.59, 95% confidence interval (CI): 3.93-5.35, p < 0.008]; for genetic databases: 380:44 (OR: 6.13, 95% CI: 4.25-8.85, p < 0.008) and cancer registries: 553:44 (OR: 5.87, 95% CI: 4.13-8.34, p < 0.008). CONCLUSIONS: TIRRE comprehensively assesses the preparedness of data repositories to participate in specific research projects. Multiple contacts about hypothetical participation in research identified few potential sites.

Smartphone-Based Real-Time Indoor Location Tracking With 1-m Precision.

Monitoring the activities of daily living of the elderly at home is widely recognized as useful for the detection of new or deteriorating health conditions. However, the accuracy of existing indoor location tracking systems remains unsatisfactory. The aim of this study was, therefore, to develop a localization system that can identify a patient's real-time location in a home environment with maximum estimation error of 2 m at a 95% confidence level. A proof-of-concept system based on a sensor fusion approach was built with considerations for lower cost, reduced intrusiveness, and higher mobility, deployability, and portability. This involved the development of both a step detector using the accelerometer and compass of an iPhone 5, and a radio-based localization subsystem using a Kalman filter and received signal strength indication to tackle issues that had been identified as limiting accuracy. The results of our experiments were promising with an average estimation error of 0.47 m. We are confident that with the proposed future work, our design can be adapted to a home-like environment with a more robust localization solution.

Accessible Modelling of Complexity in Health (AMoCH) and associated data flows: asthma as an exemplar.

Background Modelling is an important part of information science. Models are abstractions of reality. We use models in the following contexts: (1) to describe the data and information flows in clinical practice to information scientists, (2) to compare health systems and care pathways, (3) to understand how clinical cases are recorded in record systems and (4) to model health care business models.Asthma is an important condition associated with a substantial mortality and morbidity. However, there are difficulties in determining who has the condition, making both its incidence and prevalence uncertain.Objective To demonstrate an approach for modelling complexity in health using asthma prevalence and incidence as an exemplar.Method The four steps in our process are:1. Drawing a rich picture, following Checkland's soft systems methodology;2. Constructing data flow diagrams (DFDs);3. Creating Unified Modelling Language (UML) use case diagrams to describe the interaction of the key actors with the system;4. Activity diagrams, either UML activity diagram or business process modelling notation diagram.Results Our rich picture flagged the complexity of factors that might impact on asthma diagnosis. There was consensus that the principle issue was that there were undiagnosed and misdiagnosed cases as well as correctly diagnosed. Genetic predisposition to atopy; exposure to environmental triggers; impact of respiratory health on earnings or ability to attend education or participate in sport, charities, pressure groups and the pharmaceutical industry all increased the likelihood of a diagnosis of asthma. Stigma and some factors within the health system diminished the likelihood of a diagnosis. The DFDs and other elements focused on better case finding.Conclusions This approach flagged the factors that might impact on the reported prevalence or incidence of asthma. The models suggested that applying selection criteria may improve the specificity of new or confirmed diagnosis.

Using ontologies to improve semantic interoperability in health data.

The present-day health data ecosystem comprises a wide array of complex heterogeneous data sources. A wide range of clinical, health care, social and other clinically relevant information are stored in these data sources. These data exist either as structured data or as free-text. These data are generally individual person-based records, but social care data are generally case based and less formal data sources may be shared by groups. The structured data may be organised in a proprietary way or be coded using one-of-many coding, classification or terminologies that have often evolved in isolation and designed to meet the needs of the context that they have been developed. This has resulted in a wide range of semantic interoperability issues that make the integration of data held on these different systems changing. We present semantic interoperability challenges and describe a classification of these. We propose a four-step process and a toolkit for those wishing to work more ontologically, progressing from the identification and specification of concepts to validating a final ontology. The four steps are: (1) the identification and specification of data sources; (2) the conceptualisation of semantic meaning; (3) defining to what extent routine data can be used as a measure of the process or outcome of care required in a particular study or audit and (4) the formalisation and validation of the final ontology. The toolkit is an extension of a previous schema created to formalise the development of ontologies related to chronic disease management. The extensions are focused on facilitating rapid building of ontologies for time-critical research studies.

Feature combination networks for the interpretation of statistical machine learning models: application to Ames mutagenicity.

BACKGROUND: A new algorithm has been developed to enable the interpretation of black box models. The developed algorithm is agnostic to learning algorithm and open to all structural based descriptors such as fragments, keys and hashed fingerprints. The algorithm has provided meaningful interpretation of Ames mutagenicity predictions from both random forest and support vector machine models built on a variety of structural fingerprints.A fragmentation algorithm is utilised to investigate the model's behaviour on specific substructures present in the query. An output is formulated summarising causes of activation and deactivation. The algorithm is able to identify multiple causes of activation or deactivation in addition to identifying localised deactivations where the prediction for the query is active overall. No loss in performance is seen as there is no change in the prediction; the interpretation is produced directly on the model's behaviour for the specific query. RESULTS: Models have been built using multiple learning algorithms including support vector machine and random forest. The models were built on public Ames mutagenicity data and a variety of fingerprint descriptors were used. These models produced a good performance in both internal and external validation with accuracies around 82%. The models were used to evaluate the interpretation algorithm. Interpretation was revealed that links closely with understood mechanisms for Ames mutagenicity. CONCLUSION: This methodology allows for a greater utilisation of the predictions made by black box models and can expedite further study based on the output for a (quantitative) structure activity model. Additionally the algorithm could be utilised for chemical dataset investigation and knowledge extraction/human SAR development.

Dimensional reduction for a Bayesian filter.

An adaptive strategy is proposed for reducing the number of unknowns in the calculation of a proposal distribution in a sequential Monte Carlo implementation of a Bayesian filter for nonlinear dynamics. The idea is to solve only in directions in which the dynamics is expanding, found adaptively; this strategy is suggested by earlier work on optimal prediction. The construction should be of value in data assimilation, for example, in geophysical fluid dynamics.