京都大学 / 0048 / 新制・課程博士 / 博士(情報学) / 甲第22575号 / 情博第712号 / 新制||情||122(附属図書館) / 京都大学大学院情報学研究科社会情報学専攻 / (主査)教授 黒田 知宏, 教授 守屋 和幸, 教授 吉川 正俊 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

Intersection Advance Driver Assist Systems (I-ADAS) may prevent 25 to 93% of intersection crashes. The effectiveness of I-ADAS will be limited by driver's pre-crash behavior and other environmental factors. This study will characterize real-world intersection traversals to evaluate the effectiveness of I-ADAS while accounting for driver behavior in crash and near-crash scenarios. This study characterized real-world intersection traversals using naturalistic driving datasets: the Second Strategic Highway Research Program (SHRP-2) and the Virginia Traffic Cameras for Advanced Safety Technologies (VT-CAST) 2020. A step-by-step approach was taken to create an algorithm that can identify three different intersection traversal trajectories: straight crossing path (SCP); left turn across path opposite direction (LTAP/OD); and left turn across path lateral direction (LTAP/LD). About 140,000 intersection traversals were characterized and used to train a unique driver behavior model. The median average speed for all encounter types was about 7.2 m/s. The driver behavior model was a Markov Model with a multinomial regression that achieved an average 90.5% accuracy across the three crash modes. The model used over 124,000 total intersection encounters including 301 crash and near-crash scenarios. I-ADAS effectiveness was evaluated with realistic driver behavior in simulations of intersection traversal scenarios based on proposed US New Car Assessment Program I-ADAS test protocols. All near-crashes were avoided. The driver with I-ADAS overall helped avoid more crashes. For SCP and LTAP the collisions avoided increased as the field of view of the sensor increased in I-ADAS only simulations. There were 18% crash scenarios that were not avoided with I-ADAS with driver. Among near-crash scenarios, where NHTSA expects no I-ADAS activation, there were fewer I-ADAS activations (58.5%) due to driver input compared to the I-ADAS only simulations (0%). / Master of Science / Intersection Advance Driver Assist Systems (I-ADAS) may prevent 25-93% of intersection crashes. I-ADAS can assist drivers in preventing or mitigating these crashes using a collision warning system or automatically applying the brakes for the driver. One way I-ADAS may assist in crash prevention is with automatic emergency braking (AEB), which will automatically apply braking without driver input if the vehicle detects that a crash is imminent. The United States New Car Assessment Program (US-NCAP) has also proposed adding I-ADAS with AEB tests into its standard test matrix. The US-NCAP has proposed three different scenarios. All the tests have two crash-imminent configurations where the vehicles are set up to collide if no deceleration occurs and a near-miss configuration where the vehicles are set up to barely miss each other. This study will use intersection traversals from naturalistic driving data in the US to build a driver behavior model. The intersection travels will be characterized by their speed, acceleration, deceleration, and estimated time to collision. The driver behavior model was able to predict the longitudinal and lateral movements for the driver. The proposed US-NCAP test protocols were then simulated with varied sensors parameters where one vehicle was equipped with I-ADAS and a driver. The vehicle with I-ADAS with a driver was more successful than a vehicle only equipped with I-ADAS at preventing a crash.

Today, Real World Data (RWD) is a popular topic in many studies. In particular, it is anticipated to be a significant resource for addressing issues brought on by drug development costs, lengthy development times, and safety concerns. The Swedish healthcare Quality Registries (QR) are studied to contribute to the improvement of health care with individual-based clinical data. Recorded data is used for quality improvement, guidance compliance monitoring, and research. However, the workflow for such a framework that applies RWD which is patient-related data that came from various sources to the new drug development field is currently not well-defined. Thus the main aim of this project is to establish a strategy for integrating RWD with pharmaceutical modeling. To achieve this aim, QRs were examined through an ontological approach. The data and procedures necessary for modeling the development of new drugs, as well as the correspondence of the pieces offered by QR, were studied to assess the feasibility. The modeling of new drug development was studied for three applications: Adverse Drug Event (ADE), Computer-based Simulation (CBS), and drug repurposing, and the analysis of QRs was conducted on seven diseases. After in-depth analysis, although there were differences between the registries, it showed enough feasibility in terms of how much the data provided in the studies on drug repurposing and computer-based simulation satisfied the items required for new drug development. However, in the case of rare diseases, given the lack of an automated method, the ethical ambiguity, and the speed of the process, there still seems to be potential for improvement.  Many registries have begun to support research on the development of novel medications, such as by independently recording the features of drugs. These initiatives could enable the future potential of new Real World Evidence (RWE) such as in the field of proteomics and genomics discovery. / Idag är Real World Data (RWD) ett populärt ämne i många studier. I synnerhet förväntas det vara en betydande resurs för att ta itu med problem som orsakas av kostnader för läkemedelsutveckling, långa utvecklingstider och säkerhetsproblem. Kvalitetsregistren studeras för att bidra till att förbättra hälso- och sjukvården med individbaserade kliniska data. Registrerade data används för kvalitetsförbättring, övervakning av efterlevnad av riktlinjer och forskning. Arbetsflödet för ett sådant ramverk som tillämpar RWD som är patientrelaterad data som kom från olika källor till det nya läkemedelsutvecklingsområdet är dock för närvarande inte väldefinierat. Därför är huvudsyftet med detta projekt att upprätta en strategi för att integrera RWD med läkemedelsmodellering. För att uppnå detta mål undersöktes kvalitetsregister genom ett ontologiskt tillvägagångssätt. De data och procedurer som krävs för att modellera utvecklingen av nya läkemedel, såväl som överensstämmelsen mellan de bitar som erbjuds av kvalitetsregister, studerades för att bedöma genomförbarheten. Modelleringen av utvecklingen av nya läkemedel studerades för tre tillämpningar: skadlig läkemedelseffekt, datorbaserad simulering och återanvändning av läkemedel, och analys av kvalitetsregister genomfördes på sju sjukdomar. Efter en djupgående analys, även om det fanns skillnader mellan registren, visade den tillräcklig genomförbarhet när det gäller hur mycket data som tillhandahållits i studierna om läkemedelsåteranvändning och datorbaserad simulering uppfyllde de krav som krävdes för utveckling av nya läkemedel. Men när det gäller sällsynta sjukdomar, med tanke på avsaknaden av en automatiserad metod, den etiska oklarheten och processens snabbhet, verkar det fortfarande finnas potential för förbättringar. Många register har börjat stödja forskning om utveckling av nya mediciner, till exempel genom att oberoende registrera drogens egenskaper. Dessa initiative skulle kunna möjliggöra den framtida potentialen för nya verkliga bevis, såsom inom området för proteomik och genomik.

Before market authorization of new medicines, their efficacy and safety are evaluated using randomized controlled trials. While there is no doubt about the scientific value of randomized trials, they are usually conducted in selected populations with questionable generalizability to routine care.  In the digital data revolution era, with healthcare data growing at an unprecedented rate, drug monitoring in routine care is still highly under-utilized. Although many countries have access to data on prescription drugs at the individual level in ambulatory care, such data are often missing for hospitals. This is a growing problem considering the clear trend towards more new and expensive drugs administered in the hospital setting. The aim of this thesis was therefore to develop methods for extracting data on drug use from a hospital-based electronic health record system and further to build and evaluate models for real-time monitoring of effectiveness and safety of new drugs in routine care using data from electronic health records and regional and national health care registers. Using the developed techniques, we were able to demonstrate drug use and health service utilization for inflammatory bowel disease and to evaluate the comparative effectiveness and safety of antiarrhythmic drugs. With a rapidly evolving drug development, it is important to optimize the evaluation of effectiveness, safety and health economic value of new medicines in routine care. We believe that the models described in this thesis could contribute to fulfil this need.

Proton pump inhibitors (PPIs) are commonly coadministered with clopidogrel, an antiplatelet agent, to patients with acute coronary syndrome (ACS). Mechanistic studies suggest that PPIs have the potential to competitively inhibit the bioactivation of clopidogrel and may attenuate its antiplatelet action in the body. The clinical implications of this drug-drug interaction have been extensively studied; however reported findings are inconsistent. More recently, several studies have questioned whether PPIs are associated with adverse cardiovascular events independent of clopidogrel. Given that PPIs and clopidogrel are widely used, it is critical to better understand the clinical impact of the concomitant treatment with both drugs. This thesis includes four studies that investigate the clinical effects of the drug-drug interaction between clopidogrel and PPIs. Chapter 2, a systematic review and meta-analysis, summarizes findings from 118 studies. Findings do not provide strong evidence for an association between adverse cardiovascular events and the use of PPIs when used alone, in combination with clopidogrel, or in combination with other antiplatelets. Chapters 3, 4, and 5 present analyses of real-world data comprised of electronic medical records. Results of these analyses demonstrate 1) that the concomitant use of clopidogrel and PPIs among inpatients was consistent with clinical guidelines suggested by the FDA (Chapter 3); 2) a lack of association between PPI use vs nonuse and four adverse cardiovascular outcomes among clopidogrel users (Chapter 4); and 3) a lack of association between PPI use vs nonuse and adverse cardiovascular outcomes among prasugrel users or ticagrelor users (Chapter 5). Collectively, our findings do not provide evidence of an elevated risk of adverse cardiovascular outcomes with the combined use of PPIs and clopidogrel. Although pharmacodynamic and pharmacokinetic studies have demonstrated an interaction between these two drugs, our findings support the opinion that the biological interaction does not translate into adverse clinical events among patients with acute coronary syndrome.

Clinical observations of individual patients are the cornerstones for furthering our understanding of the human body, diseases, and therapeutics. Traditionally, clinical observations were communicated through publishing case reports and case series. The effort of identifying and investigating unusual clinical observations has always rested on the shoulders of busy clinicians. To date, there has been little effort dedicated to increasing the efficiency of identifying unique and uncommon patient observations that may lead to valuable discoveries. In this thesis, we propose and implement an augmented intelligence framework to identify potential novel clinical observations by combining machine analytics through outlier analysis with the judgment of subject-matter experts. Preeclampsia is a significant cause of maternal and perinatal mortality and morbidity, and advances in its management have been slow. Considering the complex etiological nature of preeclampsia, clinical observations are essential in advancing our understanding of the disease and therapeutic approaches. Thus, the objectives and studies in this thesis aim to answer the hypothesis that using outlier analysis in preeclampsia-related medical data would lead to identifying previously uninvestigated clinical cases with new clinical insight. This thesis combines three articles published or submitted for publication in peer-reviewed journals. The first article (published) is a systematic review examining the extent to which case reports and case series in preeclampsia have contributed new knowledge or discoveries. We report that under one-third of the identified case reports and case series presented new knowledge. In our second article (submitted for publication), we provide an overview of outlier analysis and introduce the framework of augmented intelligence using our proposed extreme misclassification contextual outlier analysis approach. Furthermore, we conduct a systematic review of obstetrics-related research that used outlier analysis to answer scientific questions. Our systematic review findings indicate that such use is in its infancy. In our third article (published), we implement the proposed augmented intelligence framework using two different outlier analysis methods on two independent datasets from separate studies in preeclampsia and hypertensive disorders of pregnancy. We identify several clinical observations as potential novelties, thus supporting the feasibility and applicability of outlier analysis to accelerate clinical discovery.

Data is one of the most contributing factors for modern business operations. Having bad data could therefore lead to tremendous losses, both financially and for customer experience. This thesis seeks to find anomalies in real-world, complex, structured data, causing an international enterprise to miss out on income and the potential loss of customers. By using graph theory and similarity analysis, the findings suggest that certain countries contribute to the discrepancies more than other countries. This is believed to be an effect of countries customizing their products to match the market’s needs. This thesis is just scratching the surface of the analysis of the data, and the number of opportunities for future work are therefore many.

Advances in unsupervised learning and deep neural networks have led to increased performance in a number of domains, and to the ability to draw strong comparisons between the biological method of self-organization conducted by the brain and computational mechanisms. This thesis aims to use real-world data to tackle two areas in the domain of computer vision which have biological equivalents: feature detection and motion tracking. The aforementioned advances have allowed efficient learning of feature representations directly from large sets of unlabeled data instead of using traditional handcrafted features. The first part of this thesis evaluates such representations by comparing regularization and preprocessing methods which incorporate local neighbouring information during training on a single-layer neural network. The networks are trained and tested on the Hollywood2 video dataset, as well as the static CIFAR-10, STL-10, COIL-100, and MNIST image datasets. The induction of topography or simple image blurring via Gaussian filters during training produces better discriminative features as evidenced by the consistent and notable increase in classification results that they produce. In the visual domain, invariant features are desirable such that objects can be classified despite transformations. It is found that most of the compared methods produce more invariant features, however, classification accuracy does not correlate to invariance. The second, and paramount, contribution of this thesis is a biologically-inspired model to explain the emergence of motion tracking behaviour in early development using unsupervised learning. The model’s self-organization is biased by an original concept called retinal constancy, which measures how similar visual contents are between successive frames. In the proposed two-layer deep network, when exposed to real-world video, the first layer learns to encode visual motion, and the second layer learns to relate that motion to gaze movements, which it perceives and creates through bi-directional nodes. This is unique because it uses general machine learning algorithms, and their inherent generative properties, to learn from real-world data. It also implements a biological theory and learns in a fully unsupervised manner. An analysis of its parameters and limitations is conducted, and its tracking performance is evaluated. Results show that this model is able to successfully follow targets in real-world video, despite being trained without supervision on real-world video.