Automatic de-identification of French electronic health records: a cost-effective approach exploiting distant supervision and deep learning models.

BACKGROUND: Electronic health records (EHRs) contain valuable information for clinical research; however, the sensitive nature of healthcare data presents security and confidentiality challenges. De-identification is therefore essential to protect personal data in EHRs and comply with government regulations. Named entity recognition (NER) methods have been proposed to remove personal identifiers, with deep learning-based models achieving better performance. However, manual annotation of training data is time-consuming and expensive. The aim of this study was to develop an automatic de-identification pipeline for all kinds of clinical documents based on a distant supervised method to significantly reduce the cost of manual annotations and to facilitate the transfer of the de-identification pipeline to other clinical centers. METHODS: We proposed an automated annotation process for French clinical de-identification, exploiting data from the eHOP clinical data warehouse (CDW) of the CHU de Rennes and national knowledge bases, as well as other features. In addition, this paper proposes an assisted data annotation solution using the Prodigy annotation tool. This approach aims to reduce the cost required to create a reference corpus for the evaluation of state-of-the-art NER models. Finally, we evaluated and compared the effectiveness of different NER methods. RESULTS: A French de-identification dataset was developed in this work, based on EHRs provided by the eHOP CDW at Rennes University Hospital, France. The dataset was rich in terms of personal information, and the distribution of entities was quite similar in the training and test datasets. We evaluated a Bi-LSTM + CRF sequence labeling architecture, combined with Flair + FastText word embeddings, on a test set of manually annotated clinical reports. The model outperformed the other tested models with a significant F1 score of 96,96%, demonstrating the effectiveness of our automatic approach for deidentifying sensitive information. CONCLUSIONS: This study provides an automatic de-identification pipeline for clinical notes, which can facilitate the reuse of EHRs for secondary purposes such as clinical research. Our study highlights the importance of using advanced NLP techniques for effective de-identification, as well as the need for innovative solutions such as distant supervision to overcome the challenge of limited annotated data in the medical domain.

A framework for validating AI in precision medicine: considerations from the European ITFoC consortium.

BACKGROUND: Artificial intelligence (AI) has the potential to transform our healthcare systems significantly. New AI technologies based on machine learning approaches should play a key role in clinical decision-making in the future. However, their implementation in health care settings remains limited, mostly due to a lack of robust validation procedures. There is a need to develop reliable assessment frameworks for the clinical validation of AI. We present here an approach for assessing AI for predicting treatment response in triple-negative breast cancer (TNBC), using real-world data and molecular -omics data from clinical data warehouses and biobanks. METHODS: The European "ITFoC (Information Technology for the Future Of Cancer)" consortium designed a framework for the clinical validation of AI technologies for predicting treatment response in oncology. RESULTS: This framework is based on seven key steps specifying: (1) the intended use of AI, (2) the target population, (3) the timing of AI evaluation, (4) the datasets used for evaluation, (5) the procedures used for ensuring data safety (including data quality, privacy and security), (6) the metrics used for measuring performance, and (7) the procedures used to ensure that the AI is explainable. This framework forms the basis of a validation platform that we are building for the "ITFoC Challenge". This community-wide competition will make it possible to assess and compare AI algorithms for predicting the response to TNBC treatments with external real-world datasets. CONCLUSIONS: The predictive performance and safety of AI technologies must be assessed in a robust, unbiased and transparent manner before their implementation in healthcare settings. We believe that the consideration of the ITFoC consortium will contribute to the safe transfer and implementation of AI in clinical settings, in the context of precision oncology and personalized care.

Prevalence and nature of statin drug-drug interactions in a university hospital by electronic health record mining.

AIM: Our aim was to describe prevalence, nature, and level of severity of potential statin drug-drug interactions in a university hospital. METHODS: In a cross-sectional study, statin drug-drug interactions were screened from medical record of 10,506 in-patients treated stored in the clinical data warehouse "eHOP." We screened drug-drug interactions using Theriaque and Micromedex drug databases. RESULTS: A total of 22.5% of patients were exposed to at least one statin drug-drug interaction. Given their lipophilicity and CYP3A4 metabolic pathway, atorvastatin and simvastatin presented a higher prevalence of drug-drug interactions while fluvastatin presented the lowest prevalence. Up to 1% of the patients was exposed to a contraindicated drug-drug interaction, the most frequent drug-drug interaction involving influx-transporter (i.e., OATP1B1) interactions between simvastatin or rosuvastatin with cyclosporin. The second most frequent contraindicated drug-drug interaction involved CYP3A4 interaction between atorvastatin or simvastatin with either posaconazole or erythromycin. Furthermore, our analysis showed some discrepancies between Theriaque and Micromedex in the prevalence and the nature of drug-drug interactions. CONCLUSIONS: Different drug-drug interaction profiles were observed between statins with a higher prevalence of CYP3A4-based interactions for lipophilic statins. Analyzing the three most frequent DDIs, the more significant DDIs (level 1: contraindication) were reported for transporter-based DDI involving OATP1B1 influx transporter. These points are of concern to improve prescriptions of statins.

Combining information from a clinical data warehouse and a pharmaceutical database to generate a framework to detect comorbidities in electronic health records.

BACKGROUND: Medical coding is used for a variety of activities, from observational studies to hospital billing. However, comorbidities tend to be under-reported by medical coders. The aim of this study was to develop an algorithm to detect comorbidities in electronic health records (EHR) by using a clinical data warehouse (CDW) and a knowledge database. METHODS: We enriched the Theriaque pharmaceutical database with the French national Comorbidities List to identify drugs associated with at least one major comorbid condition and diagnoses associated with a drug indication. Then, we compared the drug indications in the Theriaque database with the ICD-10 billing codes in EHR to detect potentially missing comorbidities based on drug prescriptions. Finally, we improved comorbidity detection by matching drug prescriptions and laboratory test results. We tested the obtained algorithm by using two retrospective datasets extracted from the Rennes University Hospital (RUH) CDW. The first dataset included all adult patients hospitalized in the ear, nose, throat (ENT) surgical ward between October and December 2014 (ENT dataset). The second included all adult patients hospitalized at RUH between January and February 2015 (general dataset). We reviewed medical records to find written evidence of the suggested comorbidities in current or past stays. RESULTS: Among the 22,132 Common Units of Dispensation (CUD) codes present in the Theriaque database, 19,970 drugs (90.2%) were associated with one or several ICD-10 diagnoses, based on their indication, and 11,162 (50.4%) with at least one of the 4878 comorbidities from the comorbidity list. Among the 122 patients of the ENT dataset, 75.4% had at least one drug prescription without corresponding ICD-10 code. The comorbidity diagnoses suggested by the algorithm were confirmed in 44.6% of the cases. Among the 4312 patients of the general dataset, 68.4% had at least one drug prescription without corresponding ICD-10 code. The comorbidity diagnoses suggested by the algorithm were confirmed in 20.3% of reviewed cases. CONCLUSIONS: This simple algorithm based on combining accessible and immediately reusable data from knowledge databases, drug prescriptions and laboratory test results can detect comorbidities.

Timeline representation of clinical data: usability and added value for pharmacovigilance.

BACKGROUND: Pharmacovigilance consists in monitoring and preventing the occurrence of adverse drug reactions (ADR). This activity requires the collection and analysis of data from the patient record or any other sources to find clues of a causality link between the drug and the ADR. This can be time-consuming because often patient data are heterogeneous and scattered in several files. To facilitate this task, we developed a timeline prototype to gather and classify patient data according to their chronology. Here, we evaluated its usability and quantified its contribution to routine pharmacovigilance using real ADR cases. METHODS: The timeline prototype was assessed using the biomedical data warehouse eHOP (from entrepôt de données biomédicales de l'HOPital) of the Rennes University Hospital Centre. First, the prototype usability was tested by six experts of the Regional Pharmacovigilance Centre of Rennes. Their experience was assessed with the MORAE software and a System and Usability Scale (SUS) questionnaire. Then, to quantify the timeline contribution to pharmacovigilance routine practice, three of them were asked to investigate possible ADR cases with the "Usual method" (analysis of electronic health record data with the DxCare software) or the "Timeline method". The time to complete the task and the data quality in their reports (using the vigiGrade Completeness score) were recorded and compared between methods. RESULTS: All participants completed their tasks. The usability could be considered almost excellent with an average SUS score of 82.5/100. The time to complete the assessment was comparable between methods (P = 0.38) as well as the average vigiGrade Completeness of the data collected with the two methods (P = 0.49). CONCLUSIONS: The results showed a good general level of usability for the timeline prototype. Conversely, no difference in terms of the time spent on each ADR case and data quality was found compared with the usual method. However, this absence of difference between the timeline and the usual tools that have been in use for several years suggests a potential use in pharmacovigilance especially because the testers asked to continue using the timeline after the evaluation.

Factors influencing the development of primary care data collection projects from electronic health records: a systematic review of the literature.

BACKGROUND: Primary care data gathered from Electronic Health Records are of the utmost interest considering the essential role of general practitioners (GPs) as coordinators of patient care. These data represent the synthesis of the patient history and also give a comprehensive picture of the population health status. Nevertheless, discrepancies between countries exist concerning routine data collection projects. Therefore, we wanted to identify elements that influence the development and durability of such projects. METHODS: A systematic review was conducted using the PubMed database to identify worldwide current primary care data collection projects. The gray literature was also searched via official project websites and their contact person was emailed to obtain information on the project managers. Data were retrieved from the included studies using a standardized form, screening four aspects: projects features, technological infrastructure, GPs' roles, data collection network organization. RESULTS: The literature search allowed identifying 36 routine data collection networks, mostly in English-speaking countries: CPRD and THIN in the United Kingdom, the Veterans Health Administration project in the United States, EMRALD and CPCSSN in Canada. These projects had in common the use of technical facilities that range from extraction tools to comprehensive computing platforms. Moreover, GPs initiated the extraction process and benefited from incentives for their participation. Finally, analysis of the literature data highlighted that governmental services, academic institutions, including departments of general practice, and software companies, are pivotal for the promotion and durability of primary care data collection projects. CONCLUSION: Solid technical facilities and strong academic and governmental support are required for promoting and supporting long-term and wide-range primary care data collection projects.

Using electronic health records for clinical research: the case of the EHR4CR project.

OBJECTIVES: To describe the IMI EHR4CR project which is designing and developing, and aims to demonstrate, a scalable, widely acceptable and efficient approach to interoperability between EHR systems and clinical research systems. METHODS: The IMI EHR4CR project is combining and extending several previously isolated state-of-the-art technical components through a new approach to develop a platform for reusing EHR data to support medical research. This will be achieved through multiple but unified initiatives across different major disease areas (e.g. cardiovascular, cancer) and clinical research use cases (protocol feasibility, patient identification and recruitment, clinical trial execution and serious adverse event reporting), with various local and national stakeholders across several countries and therefore under various legal frameworks. RESULTS: An initial instance of the platform has been built, providing communication, security and terminology services to the eleven participating hospitals and ten pharmaceutical companies located in seven European countries. Proof-of-concept demonstrators have been built and evaluated for the protocol feasibility and patient recruitment scenarios. The specifications of the clinical trial execution and the adverse event reporting scenarios have been documented and reviewed. CONCLUSIONS: Through a combination of a consortium that brings collectively many years of experience from previous relevant EU projects and of the global conduct of clinical trials, of an approach to ethics that engages many important stakeholders across Europe to ensure acceptability, of a robust iterative design methodology for the platform services that is anchored on requirements of an underlying Service Oriented Architecture that has been designed to be scalable and adaptable, EHR4CR could be well placed to deliver a sound, useful and well accepted pan-European solution for the reuse of hospital EHR data to support clinical research studies.

[Peranesthesic Anaphylactic Shocks: Contribution of a Clinical Data Warehouse]. / Chocs anaphylactiques peranesthésiques : contribution d'un entrepôt de données cliniques.

AIM: To evaluate the performance of the collection of cases of anaphylactic shock during anesthesia in the Regional Pharmacovigilance Center of Rennes and the contribution of a query in the biomedical data warehouse of the French University Hospital of Rennes in 2009. METHODS: Different sources were evaluated: the French pharmacovigilance database (including spontaneous reports and reports from a query in the database of the programme de médicalisation des systèmes d'information [PMSI]), records of patients seen in allergo-anesthesia (source considered as comprehensive as possible) and a query in the data warehouse. RESULTS: Analysis of allergo-anesthesia records detected all cases identified by other methods, as well as two other cases (nine cases in total). The query in the data warehouse enabled detection of seven cases out of the nine. CONCLUSION: Querying full-text reports and structured data extracted from the hospital information system improves the detection of anaphylaxis during anesthesia and facilitates access to data.

Phenotyping of heart failure with preserved ejection faction using electronic health records and echocardiography.

Aims: Patients presenting symptoms of heart failure with preserved ejection fraction (HFpEF) are not a homogenous population. Different phenotypes can differ in prognosis and optimal management strategies. We sought to identify phenotypes of HFpEF by using the medical information database from a large university hospital centre using machine learning. Methods and results: We explored the use of clinical variables from electronic health records in addition to echocardiography to identify different phenotypes of patients with HFpEF. The proposed methodology identifies four phenotypic clusters based on both clinical and echocardiographic characteristics, which have differing prognoses (death and cardiovascular hospitalization). Conclusion: This work demonstrated that artificial intelligence-derived phenotypes could be used as a tool for physicians to assess risk and to target therapies that may improve outcomes.

[Use of the PMSI for the detection of adverse drug reactions]. / Utilisation du PMSI pour la détection d'effets indésirables médicamenteux.

AIM: To evaluate the performance of a query on international classification of diseases 10(th) version (ICD10) codes in the database of the programme for the medicalisation of information systems (programme de médicalisation des systèmes d'information, PMSI) to identify serious adverse drug reactions (ADR). METHODS: The query concerned hospital stays of patients discharged from the French University Hospital of Rennes in 2009. All the hospitalization summaries including a selected ICD10 code were analysed to validate ADR. RESULTS: Out of 383 cases, 142 cases were validated (37.1%). Performance of some ICD10 codes was particularly interesting, above 40% (T88.6, L27.0, J70.4, G62.0 and N14.1) and 79.5% of the ADR were detected by these five codes. During the study period, 98 ADR of the same type were spontaneously reported by physicians, 22 of which were common with the ICD10 query. CONCLUSIONS: The use of PMSI can be a tool for signal detection of serious ADR, in addition to spontaneous reporting.

Leveraging Clinical Data Warehouses to Measure Impact of Update Prescription Guidelines of Polyvalent Immunoglobulins in 2018 in France.

In France and in other countries, we observed a significant growth in human polyvalent immunoglobulins (PvIg) usage. PvIg is manufactured from plasma collected from numeral donors, and its production is complex. Supply tensions have been observed for several years, and it is necessary to limit their consumption. Therefore, French Health Authority (FHA) provided guidelines in June 2018 to restrict their usage. This research aims to assess the guidelines' impact of the FHA on the use of PvIg. We analyzed data from Rennes University Hospital, where all PvIg prescriptions are reported electronically with quantity, rhythm, and indication. From the clinical data warehouses of RUH, we extracted comorbidities and lab results to evaluate the more complex guidelines. We globally noticed a reduction in the consumption of PvIg after the guidelines. Compliance with the recommended quantities and rhythms have also been observed. By combining two sources of data, we have been able to show an impact of FHA's guidelines on the consumption of PvIg.

Gastroenteritis Forecasting Assessing the Use of Web and Electronic Health Record Data With a Linear and a Nonlinear Approach: Comparison Study.

BACKGROUND: Disease surveillance systems capable of producing accurate real-time and short-term forecasts can help public health officials design timely public health interventions to mitigate the effects of disease outbreaks in affected populations. In France, existing clinic-based disease surveillance systems produce gastroenteritis activity information that lags real time by 1 to 3 weeks. This temporal data gap prevents public health officials from having a timely epidemiological characterization of this disease at any point in time and thus leads to the design of interventions that do not take into consideration the most recent changes in dynamics. OBJECTIVE: The goal of this study was to evaluate the feasibility of using internet search query trends and electronic health records to predict acute gastroenteritis (AG) incidence rates in near real time, at the national and regional scales, and for long-term forecasts (up to 10 weeks). METHODS: We present 2 different approaches (linear and nonlinear) that produce real-time estimates, short-term forecasts, and long-term forecasts of AG activity at 2 different spatial scales in France (national and regional). Both approaches leverage disparate data sources that include disease-related internet search activity, electronic health record data, and historical disease activity. RESULTS: Our results suggest that all data sources contribute to improving gastroenteritis surveillance for long-term forecasts with the prominent predictive power of historical data owing to the strong seasonal dynamics of this disease. CONCLUSIONS: The methods we developed could help reduce the impact of the AG peak by making it possible to anticipate increased activity by up to 10 weeks.

Prescreening in oncology trials using medical records. Natural language processing applied on lung cancer multidisciplinary team meeting reports.

Defining profiles of patients that could benefit from relevant anti-cancer treatments is essential. An increasing number of specific criteria are necessary to be eligible to specific anti-cancer therapies. This study aimed to develop an automated algorithm able to detect patient and tumor characteristics to reduce the time-consuming prescreening for trial inclusions without delay. Hence, 640 anonymized multidisciplinary team meetings (MTM) reports concerning lung cancers from one French teaching hospital data warehouse between 2018 and 2020 were annotated. To automate the extraction of eight major eligibility criteria, corresponding to 52 classes, regular expressions were implemented. The RegEx's evaluation gave a F1-score of 93% in average, a positive predictive value (precision) of 98% and sensitivity (recall) of 92%. However, in MTM, fill rates variabilities among patient and tumor information remained important (from 31% to 100%). Genetic mutations and rearrangement test results were the least reported characteristics and also the hardest to automatically extract. To ease prescreening in clinical trials, the PreScIOUs study demonstrated the additional value of rule based and machine learning based methods applied on lung cancer MTM reports.

Coupling K-nearest neighbors with logistic regression in case-based reasoning.

Case-based reasoning (CBR) systems use similarity functions to solve new problems with past situations. K-nearest neighbors algorithm (K-NN) have been used in CBR systems to define new cases status according to characteristics of past nearest cases. We proposed a new hybrid approach combining logistic regression (LR) with K-NN to optimize CBR classification. First, we analyzed the knowledge database by LR procedures and the Pearson residuals of the LR model were used to define cases' utility of the knowledge database into K-NN. Secondly, we compared the classification performances of LR model and K-NNs coupled or not with LR. Our results showed that the information provided by the residuals could be used to optimize the settings of K-NN and to improve CBR classification.

RAVEL: retrieval and visualization in ELectronic health records.

Because of the ever-increasing amount of information in patients' EHRs, healthcare professionals may face difficulties for making diagnoses and/or therapeutic decisions. Moreover, patients may misunderstand their health status. These medical practitioners need effective tools to locate in real time relevant elements within the patients' EHR and visualize them according to synthetic and intuitive presentation models. The RAVEL project aims at achieving this goal by performing a high profile industrial research and development program on the EHR considering the following areas: (i) semantic indexing, (ii) information retrieval, and (iii) data visualization. The RAVEL project is expected to implement a generic, loosely coupled to data sources prototype so that it can be transposed into different university hospitals information systems.

Phenotyping of Heart Failure with Preserved Ejection Faction Using Health Electronic Records and Echocardiography.

Patients suffering from heart failure (HF) symptoms and a normal left ventricular ejection fraction (LVEF 50%) present very different clinical phenotypes that could influence their survival. This study aims to identify phenotypes of this type of HF by using the medical information database from Rennes University Hospital Center. We present a preliminary work, where we explore the use of clinical variables from health electronic records (HER) in addition to echocardiography to identify several phenotypes of patients suffering from heart failure with preserved ejection fraction. The proposed methodology identifies 4 clusters with various characteristics (both clinical and echocardiographic) that are linked to survival (death, surgery, hospitalization). In the future, this work could be deployed as a tool for the physician to assess risks and contribute to support better care for patients.

Drug-Drug Interactions with Oral Anticoagulants as Potentially Inappropriate Medications: Prevalence and Outcomes in Elderly Patients in Primary Care and Hospital Settings.

Direct oral anticoagulants and vitamin K antagonists are considered as potentially inappropriate medications (PIM) in several situations according to Beers Criteria. Drug-drug interactions (DDI) occurring specifically with these oral anticoagulants considered PIM (PIM-DDI) is an issue since it could enhance their inappropriate character and lead to adverse drug events, such as bleeding events. The aim of this study was (1) to describe the prevalence of oral anticoagulants as PIM, DDI and PIM-DDI in elderly patients in primary care and during hospitalization and (2) to evaluate their potential impact on the clinical outcomes by predicting hospitalization for bleeding events using machine learning methods. This retrospective study based on the linkage between a primary care database and a hospital data warehouse allowed us to display the oral anticoagulant treatment pathway. The prevalence of PIM was similar between primary care and hospital setting (22.9% and 20.9%), whereas the prevalence of DDI and PIM-DDI were slightly higher during hospitalization (47.2% vs. 58.9% and 19.5% vs. 23.5%). Concerning mechanisms, combined with CYP3A4-P-gp interactions as PIM-DDI, were among the most prevalent in patients with bleeding events. Although PIM, DDI and PIM-DDI did not appeared as major predictors of bleeding events, they should be considered since they are the only factors that can be optimized by pharmacist and clinicians.

The Challenges of Implementing Comprehensive Clinical Data Warehouses in Hospitals.

Digital health, e-health, telemedicine-this abundance of terms illustrates the scientific and technical revolution at work, made possible by high-speed processing of health data, artificial intelligence (AI), and the profound upheavals currently taking place and yet to come in health systems [...].

Indexing Imaging Reports for Data Sharing: A Study of Mapping Using RadLex Playbook and LOINC.

New use cases and the need for quality control and imaging data sharing in health studies require the capacity to align them to reference terminologies. We are interested in mapping the local terminology used at our center to describe imaging procedures to reference terminologies for imaging procedures (RadLex Playbook and LOINC/RSNA Radiology Playbook). We performed a manual mapping of the 200 most frequent imaging report titles at our center (i.e. 73.2% of all imaging exams). The mapping method was based only on information explicitly stated in the titles. The results showed 57.5% and 68.8% of exact mapping to the RadLex and LOINC/RSNA Radiology Playbooks, respectively. We identified the reasons for the mapping failure and analyzed the issues encountered.

Out-of-Hospital Cardiac Arrest Detection by Machine Learning Based on the Phonetic Characteristics of the Caller's Voice.

INTRODUCTION: Out-of-hospital cardiac arrest (OHCA) is a major public health issue. The prognosis is closely related to the time from collapse to return of spontaneous circulation. Resuscitation efforts are frequently initiated at the request of emergency call center professionals who are specifically trained to identify critical conditions over the phone. However, 25% of OHCAs are not recognized during the first call. Therefore, it would be interesting to develop automated computer systems to recognize OHCA on the phone. The aim of this study was to build and evaluate machine learning models for OHCA recognition based on the phonetic characteristics of the caller's voice. METHODS: All patients for whom a call was done to the emergency call center of Rennes, France, between 01/01/2017 and 01/01/2019 were eligible. The predicted variable was OHCA presence. Predicting variables were collected by computer-automatized phonetic analysis of the call. They were based on the following voice parameters: fundamental frequency, formants, intensity, jitter, shimmer, harmonic to noise ratio, number of voice breaks, and number of periods. Three models were generated using binary logistic regression, random forest, and neural network. The area under the curve (AUC) was the primary outcome used to evaluate each model performance. RESULTS: 820 patients were included in the study. The best model to predict OHCA was random forest (AUC=74.9, 95% CI=67.4-82.4). CONCLUSION: Machine learning models based on the acoustic characteristics of the caller's voice can recognize OHCA. The integration of the acoustic parameters identified in this study will help to design decision-making support systems to improve OHCA detection over the phone.