Indiana University-Purdue University Indianapolis (IUPUI) / Phenotyping definitions are essential in cohort identification when conducting clinical research, but they become an obstacle when they are not readily available. Developing new definitions manually requires expert involvement that is labor-intensive, time-consuming, and unscalable. Moreover, automated approaches rely mostly on electronic health records’ data that suffer from bias, confounding, and incompleteness. Limited efforts established in utilizing text-mining and data-driven approaches to automate extraction and literature-based knowledge discovery of phenotyping definitions and to support their scalability. In this dissertation, we proposed a text-mining pipeline combining rule-based and machine-learning methods to automate retrieval, classification, and extraction of phenotyping definitions’ information from literature. To achieve this, we first developed an annotation guideline with ten dimensions to annotate sentences with evidence of phenotyping definitions' modalities, such as phenotypes and laboratories. Two annotators manually annotated a corpus of sentences (n=3,971) extracted from full-text observational studies’ methods sections (n=86). Percent and Kappa statistics showed high inter-annotator agreement on sentence-level annotations. Second, we constructed two validated text classifiers using our annotated corpora: abstract-level and full-text sentence-level. We applied the abstract-level classifier on a large-scale biomedical literature of over 20 million abstracts published between 1975 and 2018 to classify positive abstracts (n=459,406). After retrieving their full-texts (n=120,868), we extracted sentences from their methods sections and used the full-text sentence-level classifier to extract positive sentences (n=2,745,416). Third, we performed a literature-based discovery utilizing the positively classified sentences. Lexica-based methods were used to recognize medical concepts in these sentences (n=19,423). Co-occurrence and association methods were used to identify and rank phenotype candidates that are associated with a phenotype of interest. We derived 12,616,465 associations from our large-scale corpus. Our literature-based associations and large-scale corpus contribute in building new data-driven phenotyping definitions and expanding existing definitions with minimal expert involvement.

Indiana University-Purdue University Indianapolis (IUPUI) / Periodontal disease (PD) affects 42% of US population resulting in compromised quality of life, the potential for tooth loss and influence on overall health. Despite significant understanding of PD etiology, limited longitudinal studies have investigated PD change in response to various treatments. A major barrier is the difficulty of conducting randomized controlled trials with adequate numbers of patients over a longer time. Electronic dental record (EDR) data offer the opportunity to study outcomes following various periodontal treatments. However, using EDR data for research has challenges including quality and missing data. In this dissertation, I studied a cohort of patients with PD from EDR to monitor their disease status over time. I studied retrospectively 28,908 patients who received comprehensive oral evaluation at the Indiana University School of Dentistry between January 1st-2009 and December 31st-2014. Using natural language processing and automated approaches, we 1) determined PD diagnoses from periodontal charting based on case definitions for surveillance studies, 2) extracted clinician-recorded diagnoses from clinical notes, 3) determined the number of patients with disease improvement or progression over time from EDR data. We found 100% completeness for age, sex; 72% for race; 80% for periodontal charting findings; and 47% for clinician-recorded diagnoses. The number of visits ranged from 1-14 with an average of two visits. From diagnoses obtained from findings, 37% of patients had gingivitis, 55% had moderate periodontitis, and 28% had severe periodontitis. In clinician-recorded diagnoses, 50% patients had gingivitis, 18% had mild, 14% had moderate, and 4% had severe periodontitis. The concordance between periodontal charting-generated and clinician-recorded diagnoses was 47%. The results indicate that case definitions for PD are underestimating gingivitis and overestimating the prevalence of periodontitis. Expert review of findings identified clinicians relying on visual assessment and radiographic findings in addition to the case definition criteria to document PD diagnosis. / 2021-08-10

Blockchain is an interesting technology, with different projects developing every day since it first gained its light back in 2008. More and more research finds blockchain useful in several different sectors. One of the sectors being healthcare, specifically for electronic health records (EHR). EHR contains highly sensitive data which is critical to protect and, just in the year 2019, 41,232,527 records were deemed stolen. Blockchain can provide several benefits when it comes to EHR, such as increased security, availability, and privacy, however, it needs to be done correctly. Due to blockchain being a rather novel technology, there is room for improvement when it comes to integrating blockchain with EHR. In this thesis a framework for EHR in the healthcare sector is proposed, using Ethereum based smart contracts together with decentralized off-chain storage using InterPlanetary File System (IPFS) and strong symmetric encryption. The framework secures the records and provides a scalable solution. Furthermore, a discussion and evaluation regarding several security aspects that the framework excels on as well as what the framework could improve on.

Recent progress in treatment interventions has resulted in increased survival rates and longevity for diagnosed heart failure patients. However, heart failure still remains one of the leading causes of rehospitalization worldwide, where emergency readmissions continue to be a common occurrence. The multifactorial complexity of heart failure makes clinical judgment difficult and may lead to erroneous discharge prognoses and estimates in recovery trajectories. Recognizing emergency readmissions among heart failure patients who have been discharged is crucial within the critical six-month post-discharge period to proactively address additional support needs. To address the research question, “To what extent can machine learning models predict emergency readmissions in Chinese heart failure patients within six months post-discharge?”, this paper uses electronic health records obtained from a single healthcare center in China, containing 2,008 validated heart failure patients. This study adopts an experimental research methodology, where four machine learning models are developed to explore the research question. To ensure robustness, 10-fold cross-validation with stratified sampling and a two-step feature selection process is performed in addition to evaluation through metrics such as the area under the receiving operating curve and F1 Score. The findings indicate only modest predictive capability among the classifiers in the validation cohort. The best-achieved area under the receiving operating curve and F1 Score are obtained from separate classifiers with scores of 0.682 and 0.577, respectively. The findings provide valuable insights into future research on the effectiveness of ML-based prediction models for emergency readmission in Chinese heart failure patients.

Medical record systems are used whenever caregiving is practiced. The medical records serve an important role in establishing patient safety. It is not possible to prevent honest-but-curious doctors from accessing records since it is legally required to allow doctors to access health records for emergency cases. However, it is possible to log accesses to records and mitigate malicious behaviour through rate limiting. Nevertheless, many of the records systems today are lacking good authentication, logging and auditing and existing proposals for securing medical records systems focus on the context of multiple different healthcare providers. In this thesis, an architecture for an electronic health records system for a telemedicine provider is designed. The architecture is based on several requirements from both the legal perspective and general security conventions, but also from a doctor’s perspective. Unlike the legal and general security conventions perspective, doctor requirements are more functionality and usability concerns rather than security concerns. The architecture is evaluated based on two main threat models and one secondary threat model, i.e. insider adversaries. Almost all requirements are satisfied by the solution design, but the two main threat models can not be entirely mitigated. It is found that confidentiality can be violated by the two main threat models, but the impact is heavily limited through audit logging and rate limiting. / Journalsystem är en central del inom vården och patientjournaler har en stor roll i att uppnå bra patientsäkerhet. Det är inte möjligt att förhindra läkare från att läsa särskilda journaler eftersom läkare behöver tillgång till journaler vid nödsituationer. Däremot går det att logga läkarnas handlingar och begränsa ondsint beteende. Trots det saknar många av dagens journalsystem bra metoder för autentisering, loggning och granskning. Befintliga förslag på att säkra journalsystemen fokuserar på sammanhang där flera olika vårdgivare är involverade. I den här rapporten presenteras en arkitektur för ett patientjournalsystem till en telemedicinsk leverantör. Arkitekturen utgår från flertalet krav baserade på både ett legalt perspektiv och generella säkerhetskonventioner, men även läkares perspektiv. Arkitekturen är evaluerad baserat på två huvudsakliga hotmodeller och en sekundär hotmodell. Arkitekturen uppfyller så gott som alla krav, men de två huvudsakliga hotmodellerna kan inte mitigeras helt och hållet. De två huvudsakliga hotmodellerna kan bryta sekretessen, men genom flödesbegränsning och granskning av loggar begränsas påverkan.

No description available.

Electronic Health Records (EHR) data contain comprehensive longitudinal patient information, which is usually stored in databases in the form of either multi-relational structured tables or unstructured texts, e.g., clinical notes. EHR provides a useful resource to assist doctors' decision making, however, they also present many unique challenges that limit the efficient use of the valuable information, such as large data volume, heterogeneous and dynamic information, medical term abbreviations, and noisy nature caused by misspelled words. This dissertation focuses on the development and evaluation of advanced machine learning algorithms to solve the following research questions: (1) How to seek answers from EHR for clinical activity related questions posed in human language without the assistance of database and natural language processing (NLP) domain experts, (2) How to discover underlying relationships of different events and entities in structured tabular EHRs, and (3) How to predict when a medical event will occur and estimate its probability based on previous medical information of patients. First, to automatically retrieve answers for natural language questions from the structured tables in EHR, we study the question-to-SQL generation task by generating the corresponding SQL query of the input question. We propose a translation-edit model driven by a language generation module and an editing module for the SQL query generation task. This model helps automatically translate clinical activity related questions to SQL queries, so that the doctors only need to provide their questions in natural language to get the answers they need. We also create a large-scale dataset for question answering on tabular EHR to simulate a more realistic setting. Our performance evaluation shows that the proposed model is effective in handling the unique challenges about clinical terminologies, such as abbreviations and misspelled words. Second, to automatically identify answers for natural language questions from unstructured clinical notes in EHR, we propose to achieve this goal by querying a knowledge base constructed based on fine-grained document-level expert annotations of clinical records for various NLP tasks. We first create a dataset for clinical knowledge base question answering with two sets: clinical knowledge base and question-answer pairs. An attention-based aspect-level reasoning model is developed and evaluated on the new dataset. Our experimental analysis shows that it is effective in identifying answers and also allows us to analyze the impact of different answer aspects in predicting correct answers. Third, we focus on discovering underlying relationships of different entities (e.g., patient, disease, medication, and treatment) in tabular EHR, which can be formulated as a link prediction problem in graph domain. We develop a self-supervised learning framework for better representation learning of entities across a large corpus and also consider local contextual information for the down-stream link prediction task. We demonstrate the effectiveness, interpretability, and scalability of the proposed model on the healthcare network built from tabular EHR. It is also successfully applied to solve link prediction problems in a variety of domains, such as e-commerce, social networks, and academic networks. Finally, to dynamically predict the occurrence of multiple correlated medical events, we formulate the problem as a temporal (multiple time-points) and multi-task learning problem using tensor representation. We propose an algorithm to jointly and dynamically predict several survival problems at each time point and optimize it with the Alternating Direction Methods of Multipliers (ADMM) algorithm. The model allows us to consider both the dependencies between different tasks and the correlations of each task at different time points. We evaluate the proposed model on two real-world applications and demonstrate its effectiveness and interpretability. / Doctor of Philosophy / Healthcare is an important part of our lives. Due to the recent advances of data collection and storing techniques, a large amount of medical information is generated and stored in Electronic Health Records (EHR). By comprehensively documenting the longitudinal medical history information about a large patient cohort, this EHR data forms a fundamental resource in assisting doctors' decision making including optimization of treatments for patients and selection of patients for clinical trials. However, EHR data also presents a number of unique challenges, such as (i) large-scale and dynamic data, (ii) heterogeneity of medical information, and (iii) medical term abbreviation. It is difficult for doctors to effectively utilize such complex data collected in a typical clinical practice. Therefore, it is imperative to develop advanced methods that are helpful for efficient use of EHR and further benefit doctors in their clinical decision making. This dissertation focuses on automatically retrieving useful medical information, analyzing complex relationships of medical entities, and detecting future medical outcomes from EHR data. In order to retrieve information from EHR efficiently, we develop deep learning based algorithms that can automatically answer various clinical questions on structured and unstructured EHR data. These algorithms can help us understand more about the challenges in retrieving information from different data types in EHR. We also build a clinical knowledge graph based on EHR and link the distributed medical information and further perform the link prediction task, which allows us to analyze the complex underlying relationships of various medical entities. In addition, we propose a temporal multi-task survival analysis method to dynamically predict multiple medical events at the same time and identify the most important factors leading to the future medical events. By handling these unique challenges in EHR and developing suitable approaches, we hope to improve the efficiency of information retrieval and predictive modeling in healthcare.

Yes / We compare the performance of logistic regression with several alternative machine learning methods to estimate the risk of death for patients following an emergency admission to hospital based on the patients’ first blood test results and physiological measurements using an external validation approach. We trained and tested each model using data from one hospital (n=24696) and compared the performance of these models in data from another hospital (n=13477). We used two performance measures – the calibration slope and area under the curve (AUC). The logistic model performed reasonably well – calibration slope 0.90, AUC 0.847 compared to the other machine learning methods. Given the complexity of choosing tuning parameters of these methods, the performance of logistic regression with transformations for in-hospital mortality prediction was competitive with the best performing alternative machine learning methods with no evidence of overfitting. / Health Foundation; National Institute for Health Research (NIHR) Yorkshire and Humberside Patient Safety Translational Research Centre (NIHR YHPSTRC)

Information Systems (IS) innovations in healthcare sector are seen as panacea to control burgeoning demand on healthcare resources and lack of streamlining in care delivery. Two particular manifestations of such innovations are telehealth and electronic records in its two forms: the electronic medical records and the electronic health records. Deployment efforts concerning both of these IS-innovations have encountered a rough terrain and have been slow. Problems are also faced while evaluating the effectiveness of innovations on health and care delivery outcomes through strategies such as randomised controlled trials- particularly in case of telehealth. By taking these issues into account, this research investigates the issues that affect IS innovation deployment and its evaluation. The strategy adopted in this research was informed by recursive philosophy and theoretical perspectives within IS that strived to expound this recursive relationship. It involved conducting two longitudinal case studies that are qualitative in nature. The first study involved telehealth deployment and its evaluation in the UK, while the second case study involved the deployment of electronic medical/health records in the US. Data was collected through focus group discussions, interviews and online discussion threads; and was analysed thematically. The results of this research indicate that there are nine issues that arise and affect the deployment and evaluation of IS innovation in healthcare; and these are design, efficiency and effectiveness, optimality and equity, legitimacy, acceptance, demand and efficacy, expertise, new interaction patterns, and trust. These issues are attributes of relationships between the IS innovation, context of healthcare and the user. The significance of these attributes varies during the deployment and evaluation process, and due to iterative nature of IS innovation. This research further indicates that all the attributes have either direct or indirect impact on work practices of the user.

Introdução: Araraquara (SP) possui um programa de vacinação bem sucedido desde a década de 80, com o mais antigo Registro Informatizado de Imunização (RII) do país. Objetivos: Estimar a cobertura vacinal (CV) aos 12 e 24 meses de vida, em crianças nascidas em 2012, no município de Araraquara, investigar fatores associados à vacinação incompleta e analisar os eventos adversos pós-vacinação (EAPV). Métodos: Estudo observacional com componentes descritivo e analítico, abrangendo a coorte de crianças nascidas em 2012, residentes no município e registradas no Sistema de Informação de Nascidos Vivos (SINASC). Foram excluídas as crianças que faleceram no primeiro ano de vida e aquelas que se mudaram de Araraquara. As variáveis do estudo incluíram dados de vacinação e características maternas, de pré- natal/nascimento, do(s) serviço(s) de saúde e da área de residência. Estimou-se as CV e respectivos intervalos de confiança de 95 por cento (IC 95 por cento ) para cada vacina e esquema completo, conforme as normas vigentes no Estado de São Paulo em 2012/2013. A associação entre o esquema vacinal incompleto e as variáveis independentes foi investigada por meio da estimativa da odds ratio (OR) bruta e ajustada, por regressão logística múltipla não condicional hierarquizada, com os respectivos IC 95 por cento . Resultados: 2740 crianças estavam registradas no SINASC como residentes e 99,6 por cento dessas constavam no RII. Após excluir 30 óbitos (1,1 por cento ) e 98 crianças que se mudaram (3,6 por cento ), foram estudadas 2612 crianças. A CV para o esquema completo por doses recebidas aos 12 meses foi 67,9 por cento e aos 24 meses 79,7 por cento ; por doses oportunas foi 46,2 por cento e 32,8 por cento , respectivamente. A vacina sarampo, caxumba e rubéola apresentou a menor CV aos 12 meses por dose recebida (74,8 por cento ) e aos 24 meses por dose oportuna (53,5 por cento ). As vacinas com componente pertussis foram responsáveis por 58,8 por cento (10/17) dos casos de EAPV e febre foi a manifestação mais comum. A distribuição espacial da CV do esquema completo por área de residência não apresentou diferença estatística. No modelo final, mostraram-se independentemente associadas à vacinação incompleta: mães com idade entre 14 e 19 anos [aos 12 meses (OR:2,0); aos 24 meses (OR:2,5)]; com 12 anos ou mais de estudo [aos 12 meses (OR:1,9), aos 24 meses (OR:2,3)]; com três ou mais filhos [aos 12 meses (OR:3,2), aos 24 meses (OR:2,1)]; com menos de sete consultas de pré-natal [aos 12 meses (OR:1,7), aos 24 meses (OR:2,3)]; a criança ter frequentado unidade de saúde (US) pública e privada [aos 12 meses (OR:6,0), aos 24 meses (OR:8,0)], sem Estratégia Saúde da Família [aos 24 meses (OR:1,5)]; e ter vínculo fraco com a US [aos 24 meses (OR:1,4)]. Conclusão: Em Araraquara, a CV por vacina não é homogênea e há atraso vacinal. O uso do RII para o seu monitoramento pode constituir uma estratégia efetiva. A ausência de disparidades nas CV entre as distintas áreas de residência sugere a efetividade do programa de imunização na promoção da equidade em saúde. Recomenda-se priorizar ações de incentivo à vacinação de crianças filhas de mulheres com alta escolaridade e que apresentam vínculo mais frágil com os serviços públicos de saúde / Introduction: Araraquara (SP) has a successful vaccination program since the 80\'s, with the oldest Electronic Immunization Registry (EIR) in the country. Objectives: To estimate vaccination coverage (VC) at 12 and 24 months of life in children born in 2012 in the city of Araraquara, to investigate factors associated with incomplete vaccination and to analyze the adverse events following immunization (AEFI). Methods: An observational descriptive and analytical study comprising the cohort of children born in 2012, living in the city of Araraquara and recorded in the Live Births Information System (SINASC). Children who died in the first year of life or who moved from Araraquara were excluded. Study variables included vaccination data and characteristics of the mother, the antenatal/birth, the health unit (HU) and the area of residence. VC and the respective 95 per cent confidence intervals (95 per cent CI) were estimated for each vaccine and complete schedule, following the São Paulo\'s State recommendations in the years 2012/2013. The association between the incomplete vaccination schedule and the independent variables was investigated by estimating the crude and adjusted odds ratio (OR) by hierarchical non-conditional multiple logistic regression with the respective 95 per cent CI. Results: 2740 children were enrolled in the SINASC as residents and 99.6 per cent of them were in the EIR. After excluding 30 deaths (1.1 per cent ) and 98 children who moved (3.6 per cent ), 2612 children were studied. VC by received doses for the complete schedule at 12 months was 67.9 per cent and at 24 months was 79.7 per cent ; by timely doses was 46.2 per cent and 32.8 per cent , respectively. The measles, mumps and rubella vaccine had the lowest VC at 12 months per received dose (74.8 per cent ) and at 24 months per timely dose (53.5 per cent ). Vaccines with pertussis componente were responsible for 58.8 per cent (10/17) of AEFI cases and fever was the most common manifestation. The spatial distribution of VC of the complete schedule by area of residence did not present statistical difference. In the final model, incomplete vaccination was associated with mother between 14 and 19 years old [at 12 months (OR: 2.0); at 24 months (OR: 2.5)]; with 12 years or more of study [at 12 months (OR: 1.9), at 24 months (OR: 2.3) ]; with three or more children [at 12 months (OR: 3.2), at 24 months (OR: 2.1)]; with less than seven antenatal visits [at 12 months (OR: 1.7), at 24 months (OR: 2.3)]; the child has attended both public and private HU [at 12 months (OR: 6.0), at 24 months (OR: 8.0)], a HU without Family Health Strategy [at 24 months (OR: 1.5)]; and who had a weak link with the HU [at 24 months (OR: 1.4)]. Conclusion: VC per vaccine is not homogeneous in Araraquara and there is a vaccine delay. The use of RII for its monitoring can be an effective strategy. The lack of disparities in VC among the different areas of residence suggests the effectiveness of the immunization program in promoting health equity. It is recommended to prioritize actions to encourage children vaccination of mothers with high schooling and who have a more fragile link with public HU