Method for processing medical data and assessing eligibility for a medical study

Abstract

The invention relates to a method for processing medical data of a patient during a medical consultation and for assessing a match with a medical study, the method being implemented by a computer system (1) and comprising: (E1) connecting to a medical data database (42) in order to obtain a medical dataset (DMi) of the patient; (E2) analyzing the medical data (DMi) in order to generate a summary of care episodes (S); (E4) connecting to a medical study database (52) in order to obtain sets of eligibility criteria (El_JCm), each set of eligibility criteria being associated with a medical study (El); (E5) assessing a match score (SCm) between the summary and each set of criteria; (E6) selecting or excluding each medical study on the basis of the associated match score; (E7) generating and transmitting a patient eligibility notification (NE) for the selected medical studies.

Description

Description Title of the invention: Method for processing medical data and assessing eligibility for a medical study

[0001] technical field

[0002] The invention relates to the technical field of health and medical data processing, and more particularly to a method, a computer system and a computer program for the analysis of a patient's medical data during a medical consultation and the evaluation of the patient's eligibility for medical or clinical studies.

[0003] State of the art

[0004] The use of patients' medical data is a major challenge for improving the quality of care and advancing medical research. However, accessing and using patient data presents significant challenges, particularly given the ethical and legal framework surrounding this data and the lack of structured data collection.

[0005] For example, despite the implementation of the Shared Medical Record (DMP) in France, the information it contains remains largely underutilized. Indeed, the limited integration of DMP data into healthcare professionals' software and the difficulty in performing effective and relevant searches within these records constitute major obstacles to their optimal use. Furthermore, current healthcare software offers few innovations in the use of DMP data or data from the electronic medical record internal to healthcare software. Finally, the heterogeneity of the sources and the structure of this medical data make it complex to utilize. This situation significantly limits practitioners' ability to quickly obtain a relevant overview of a patient's medical history during a consultation.

[0006] Furthermore, identifying patients eligible for medical studies remains a complex and time-consuming process. Current methods, often manual, involve the tedious review of hundreds or even thousands of medical records, each potentially containing hundreds of pages of information. This approach is not only inefficient but also prone to errors, which can lead to the inappropriate recruitment or inclusion of patients in studies for which they are not relevant or qualified, thereby compromising the validity of the results and wasting valuable time in initiating the study.

[0007] Other methods involving one or more data warehouses, from which patient data are extracted for use for purposes not precisely defined, raise ethical questions about the secondary reuse of health data, without informed consent on uses, without transparency, without traceability of processing, without information on tangible benefits for contributing patients.

[0008] Furthermore, access to patient data by research organizations (CROs, the French National Institute of Health and Medical Research (INSERM), the French National Centre for Scientific Research (CNRS), University Hospitals (CHU), and public hospitals) for the purpose of establishing cohorts or panels is hampered by difficulties related to managing the ethical framework and ensuring data confidentiality. Current systems, such as those described in applications WO2022146838A1 and W02020243400A1, do not offer a satisfactory solution for facilitating the secure exchange of data between physicians, medical software vendors, and companies involved in clinical trials, while respecting patient consent and data traceability.

[0009] Furthermore, statistics on health data are essential for guiding public health policy decisions and are difficult for public institutions to access, such as, in France, the Digital Health Agency (ANS), the ministries (Health), the National Health Insurance Fund (CNAM), the High Authority for Health (HAS), the National Health Observatory (ONS), the Institute for Research and Documentation in Health Economics (IRDES), the National Agency for the Safety of Medicines (ANSM), the Directorate for Research, Studies, Evaluation and Statistics (DREES), the National Inter-Scheme Health Insurance Information System (SNIIRAM), the Medicalization of Information Systems Program (PMSI), the National Health Data System (SNDS), the Health Data Hub, and in Europe, the European Medicines Agency (EMA), the European Centre for Disease Prevention and Control (ECDC).The European Health Data Space (EHDS), the European Centre for Disease Prevention and Control (ECDC), the Directorate-General for Health and Food Safety (DG SANTE), the Food and Drug Administration (FDA) in the United States, the US Centers for Disease Control and Prevention (CDC), the National Health Commission (NHC) in China, the Ministry of Health, Labour and Welfare (MHLW) in Japan, the World Health Organization (WHO), and health networks (cancer, diabetes, rare diseases, etc.) are all examples of organizations that use this technology. Current solutions do not allow for the precise targeting of dispersed patient panels to identify health priorities and measure the impact of public health policies and epidemiological trends.

[0010] Finally, existing methods for extracting relevant information from electronic health records, such as those presented in application WO2022187628A1, are not designed to provide a clear and intelligent synthesis of the EHR or other data sources directly within the physician's software during the consultation. This limits the practitioner's ability to make informed decisions and promptly offer patients participation in medical studies. Furthermore, no solution exists that provides complete traceability of the data sources used for the synthesis, which limits trust in the decision-making process of both the physician and the patient.

[0011] There is therefore a need for a process for processing a patient's medical data and evaluating a match between the patient and at least one medical study, which allows for the efficient use of this medical data, particularly when it comes from the DMP or an electronic medical record of a practitioner or a health center or a specific hospital; rapid identification of the patient's eligibility for a medical study, and secure management of data exchanges in compliance with the ethical framework.

[0012] The invention therefore falls within this context and seeks to resolve all the aforementioned drawbacks by addressing this need.

[0013] Presentation of the invention

[0014] The invention thus relates to a method for processing a patient's medical data during a medical consultation by a healthcare professional and for evaluating a correlation between the patient and at least one medical study, the method being implemented by a computer system and comprising the following steps: a. Connecting the computer system to at least one medical database to obtain a set of medical data relating to the patient; b. Analysis by the computer system of the medical data from the provided game to generate a summary of patient care episodes; c. Connecting the computer system to at least one database of medical studies to obtain a plurality of sets of eligibility criteria, each set being associated with a given medical study; d. For each set of eligibility criteria provided, evaluation by the computer system of at least one correspondence score between said patient care episode summary and said set of eligibility criteria provided; e. For each medical study provided, selection or exclusion by the computer system of said medical study based on said match score determined for the set of eligibility criteria associated with that medical study; f. Generation and transmission of a patient eligibility notification for the selected medical study(ies) to the patient and / or healthcare professional.

[0015] The invention thus proposes, during a consultation, access to the patient's medical data, for example, through the healthcare software used by the healthcare professional or via an independent software solution connected to the healthcare professional's medical software. The computer system can then access a complete and up-to-date view of the patient's medical history. In a second step, the computer system performs an intelligent analysis of all or part of the collected data, and in particular of each individual medical record, to obtain a structured and relevant summary of the patient's care episodes, facilitating the healthcare professional's rapid understanding of the patient's medical history.This summary can be displayed, either during its generation or later during the implementation of the process, on the healthcare professional's computer terminal, particularly through their healthcare software or a third-party software solution. This allows them to view, browse, and inspect the care episodes they deem relevant to their specialty, as well as access the data and documents that generated the summary to complete their analysis and verify its reliability. The summary of care episodes is thus integrated into the healthcare professional's workflow, optimizing consultation time and enabling the efficient use of patient medical data while improving the traceability of data used for decision-making, thereby enhancing the quality of patient care.

[0016] Concurrently or sequentially with these initial steps, the computer system accesses a database of planned, pre-enrollment medical studies, allowing the retrieval of the cohort eligibility criteria or panel for each study.

[0017] By combining the previously generated patient care episode summary, or even all or part of the medical data retrieved during the first step, and each of the retrieved eligibility criteria sets, the computer system can thus estimate or quantify the patient's eligibility for each medical study and automatically select the most relevant medical studies with regard to the patient's medical history.

[0018] The selected studies are then notified either directly to the healthcare professional or to the patient, for example via a platform accessing their Personal Health Record (DMP) or a third-party software solution, so that the patient can be informed of their eligibility. This process allows for an automated and real-time assessment of the patient's eligibility for one or more medical studies during the patient consultation, within the healthcare professional's workflow. The healthcare professional can then clearly explain the medical summary and the objective of the study for which the patient is eligible, thus reliably obtaining the patient's informed consent during the consultation.The patient can thus give their consent to be included in the recruitment phase of the cohort or panel of the medical study, or for their medical data, relevant to the study, to be transmitted to the study's principal investigator. It is therefore possible to organize secure data exchange management, in compliance with ethical guidelines, by centralizing access and limiting its use to specific purposes.

[0019] Definitions

[0020] In the present invention, the term "computer system" means any set of interconnected hardware and software components designed to process, store, and manipulate digital data in order to implement the process according to the invention. A computer system may include one or more processing devices, storage means, input / output interfaces, applications, programs or software, and communication means. The computer system may operate autonomously or as a network, and may be physically located on a single site or geographically distributed.

[0021] The computer system may include, but is not limited to, a computer terminal such as a desktop computer, a laptop computer, a server, a tablet, a smartphone, a cell phone and / or equipment of a computer network; a local computer network comprising several computer terminals and one or more servers interconnected to the terminals; or a "cloud" type computer infrastructure comprising computer servers accessible by means of wireless communication and providing services to a remote computer terminal, in particular of the SaaS (Software As A Service) type.

[0022] Preferably, the computer system may include one or more remote computer servers from a computer terminal that can be operated by the practitioner, the services of said server(s) then being accessible by this computer terminal via wireless communication protocols, such as through an internet browser and HTTPS and IP protocol, or through an application programming interface or API, or via cloud computing methods.Where applicable, all or part of the programs, software, or applications necessary for implementing the method according to the invention may be installed on the computer server(s), either centrally or in a distributed manner. All or part of these programs, software, or applications may be accessible from the computer terminal via wireless communication protocols, such as through an internet browser and HTTPS and IP protocols, or through an application programming interface (API). Where applicable, each computer terminal of the computer system may be equipped with a communication unit capable of exchanging data with a communication unit of a computer server capable of implementing all or part of the method according to the invention, in particular using one or more wireless communication protocols.

[0023] Alternatively, one or more computer applications, computer programs or software, enabling the implementation of the process according to the invention, may be installed directly on a computer terminal that can be operated by the healthcare professional, and / or installed on a server on a local network so that they can be accessed by a computer terminal on that local network, or hosted on a computer server from which they are accessible through software installed on a computer terminal, such as an internet browser or an application programming interface.

[0024] Preferably, the IT system may include one or more healthcare software programs specifically designed for processing medical data and managing patient records. These programs may be installed locally on a healthcare professional's computer terminal, such as a workstation, tablet, mobile device, or laptop, or accessed from that terminal via a client-server architecture.In this configuration, during the connection steps, the computer terminal can connect, through health software, to databases hosted on remote servers, including the medical database and the medical studies database, the health software being able to use specific services, including via application programming interfaces (APIs), to perform the data processing steps of the process according to the invention, such as the steps of analyzing and generating summaries of care episodes, evaluating correspondence scores, selecting and excluding medical studies and generating eligibility notifications, these notifications being able to be displayed on the health professional's computer terminal.

[0025] In the present invention, the term "medical dataset" means any collection of data, structured or unstructured, in raw, processed, digitized, or consolidated form, relating directly or indirectly to the patient's health. This data may be generated manually and / or automatically during a patient's care pathway or following various types of medical interventions, by one or more healthcare professionals and / or by one or more healthcare facilities and / or organizations and / or by one or more computer programs, all at once or sequentially, and is not limited to a particular medical field.

[0026] Medical data may include, but is not limited to: a. A medical document such as a consultation or procedure report; the result of a clinical, biological, radiological, or other analysis or examination; a prescription or order, or a vaccination certificate; or any other document, regardless of its format (text, table, image, sound, video, graphic) containing medical information relating to an investigation, diagnosis, treatment, or medical follow-up. b.Medical administrative data of the patient, such as patient identification data, like their surname(s), first name(s), age, date of birth, place of birth, sex, place of residence, medical identification number, such as a social security number, family or legal affiliation, or care pathway data such as a date of medical care or discharge from an establishment, a place, or a type of care; or medical follow-up data such as a file number or a follow-up file; or data relating to the identity of a healthcare professional who has interacted directly or indirectly with the patient, such as a name, a specialty, a department or a care unit; or any other data relating to the patient's medical history and care pathway; c. Medical financial data of the patient, such as billing data, reimbursement data and coverage by health insurance, mutuals and insurance companies, in particular on the dates and level of coverage of long-term illnesses; d. data relating to the patient's medical history, such as allergies, vaccinations, and current or past treatments and / or health conditions, information relating to family history, lifestyle habits (social history); e. continuous, discontinuous or one-off medical monitoring data of the patient, such as data from a connected health device, software or a remote monitoring or teleconsultation application. f. Prevention data, provided by the patient, as part of an assisted or self-administered prevention assessment, including quantitative indicators or risk scores.

[0027] The said medical data may be formatted, structured or normalised according to one or more standards for the exchange and / or interoperability of medical documents, such as an HL7 (Health Level 7), FHIR (Fast Healthcare Interoperability Resources), DICOM (Digital Imaging and Communications in Medicine) or IHE XDS (Cross-Enterprise Document Sharing) standard.

[0028] In the present invention, "medical study" means scientific and / or socio-economic research conducted in the field of health, aimed at increasing medical knowledge, evaluating a treatment or intervention or health policy or epidemiological surveillance, or improving medical practices and involving the collection of medical data from one or more patients before or during the medical study, from a database or directly from the patient(s), as well as the analysis and interpretation of this medical data, the medical study being carried out within a strict ethical framework, in particular with regard to the informed consent of the patients participating in the study and the protection of their privacy.

[0029] A medical study can be either interventional, meaning it involves administering a treatment, intervention, or therapeutic or preventative strategy to a patient, cohort, or panel of patients, or observational, consisting solely of collecting and analyzing medical data from a cohort or panel of patients. A medical study can be, in particular, a cross-sectional study, a case-control study, a prospective or retrospective cohort or panel study, a randomized controlled trial, a non-randomized study, a crossover trial, a cluster study, or a qualitative study.

[0030] In the context of the present invention, a medical study may be a medical study in the planning or pre-enrollment phase, prior to the recruitment, execution, and analysis phases, and in which the study objectives and a set of eligibility criteria are defined, among other things. A medical study may also be a simple data collection for statistical purposes.

[0031] In the present invention, "eligibility criterion" means a condition according to which a patient may or may not be considered a candidate for a medical study.

[0032] This could be an inclusion criterion defining a necessary condition for a patient to be considered a candidate in the medical study, or an exclusion criterion defining a sufficient condition for a patient to be excluded from the study. An eligibility criterion could be a strict condition, relating to the absence or presence of a patient characteristic or strict adherence to a range of values, or conversely, an approximate condition or a set of strict and / or relative conditions, which could, for example, be linked to a demographic, geographic, pathological, genetic, or behavioral characteristic of the patient, to a clinical or therapeutic condition of the patient, or to the patient's clinical history.

[0033] A set of eligibility criteria could, for example, be of the following type: a. Inclusion criteria: i. Age between 18 and 65 years, ii. Confirmed diagnosis of chronic gastritis, iii. Presence of persistent gastrointestinal symptoms for at least 3 months; b. Exclusion criteria: i. Comorbidity: Gastric ulcer, ii. Current treatment: Proton pump inhibitors.

[0034] In the present invention, "database" means any computer system for collecting, storing, organizing, displaying and managing data.

[0035] A medical database can therefore contain medical data relating to one or more patients. It can be, interchangeably, a database centralizing medical data from different sources, such as the Shared Medical Record (DMP) or an electronic medical record, or a database managed by a health insurance company or a public or private health or research organization; a database of a single medical structure, such as a doctor's office, clinic, hospital, healthcare facility, pharmacy, laboratory, medical imaging center, or any other healthcare provider; a database of a medical software solution provider; or a database responsible for collecting data acquired by connected devices or mobile health or wellness applications used by the patient.It can be expected that this data will be stored or exchanged in standardized formats, for example conforming to an HL7, FHIR, DICOM or IHE XDS standard.

[0036] A medical study database can contain information relating to medical studies and can be hosted by a public or private institution, a research organization, a medical study sponsor, an investigator site, a coordinator, or a contract research organization (CRO). It can be either a central database for medical studies from different organizations or a database designed to manage medical studies from a single organization.

[0037] It can be conceived that, during the step of connecting the computer system to at least one medical or medical study database, the computer system connects, simultaneously or sequentially, to different databases to collect the patient's medical data or to obtain different sets of eligibility criteria.

[0038] In the present invention, "episode of care" means a delimited set of uses of care or follow-up of care, or interventions by a health professional, mobilized during a defined period of time, for a patient in response to the same reason for using care, such as a health problem, a pathology, a health condition, a state of health, or prevention, such as screening or vaccination.

[0039] A care episode can be defined as all the resources and services, such as hospitalizations, treatments, procedures, consultations, follow-ups and remote consultations, remote monitoring, and telemedicine equipment, dedicated to the same reason for seeking care over a specific period, which can range from a few days to several years, without requiring a direct link between these resources and services or continuity of care during this period. A care episode can be delimited temporally, either by the resolution of the health problem, by a maximum duration, or by a maximum time interval separating two consecutive visits to care within the episode. A care episode may coexist, in whole or in part, with other care episodes related to distinct reasons for seeking care.

[0040] In the context of the present invention, a summary of a patient's care episodes can be formed by any structured and chronological representation of one or more patient care episodes, capable of providing an overview of their health journey.

[0041] A summary of care episodes may include, but is not limited to: the reason for seeking care for each episode, the chronology and duration of each episode, all or part of the care visits for each episode, the temporal or causal relationships between the care visits for each episode, and key information for each episode, such as diagnoses, treatments, and outcomes. A summary of care episodes may be presented in various formats, including graphical, textual, or tabular.

[0042] As an example, a summary of care episodes for a given patient might look like this: a. "Hypertension" care episode (01 / 01 / 2020 - ongoing): i. 01 / 01 / 2020 General practitioner consultation - Initial diagnosis ii. 15 / 01 / 2020 Blood test - Lipid profile iii. 01 / 02 / 2020 Cardiologist consultation - Diagnostic confirmation iv. 15 / 02 / 2020 Medication prescription - Start of treatment v. 01 / 05 / 2020 Follow-up consultation - Treatment adjustment vi. 01 / 11 / 2020 Follow-up consultation - Stabilization b. Treatment episode "Wrist Fracture" (15 / 03 / 2020 - 30 / 06 / 2020): i. 15 / 03 / 2020: Emergency consultation - Diagnosis and cast ii. 22 / 03 / 2020: Follow-up X-ray iii. 15 / 04 / 2020: Orthopedic consultation - Cast removal iv. 20 / 04 / 2020 - 30 / 06 / 2020: Physiotherapy sessions

[0043] In the present invention, "at least one correspondence score between a patient's episode of care summary and a set of eligibility criteria associated with a medical study" means one or more quantitative or qualitative values ​​representing the degree of adequacy between the information contained in the patient's episode of care summary and the eligibility criteria defined for the medical study.

[0044] A correspondence score can thus be expressed as a single value, as a weighted combination of the results of the correspondence between the summary of care episodes and each eligibility criterion, or as a percentage of compliance with the eligibility criteria, or as a level of relevance on a defined scale.

[0045] A matching score can also be expressed as a binary result indicating whether or not the patient meets all the eligibility criteria.

[0046] A matching score may also be expressed as several individual scores, each corresponding to a specific eligibility criterion or a combination of several eligibility criteria, these scores being grouped for example in a vector or matrix representation or any other representation relevant to assess the patient's eligibility for the medical study.

[0047] Methods of implementation

[0048] In one embodiment of the invention, the method includes a preliminary authentication step, implemented by the computer system, of the patient's identity, the healthcare professional's identity, and the recording of the patient's consent to obtaining this medical data. The step of connecting the computer system to the medical database to obtain the set of medical data relating to the patient is conditional upon the success of the authentication of the patient's and healthcare professional's identities and the successful recording of the patient's consent.

[0049] In the present invention, "recording of consent to the obtaining of medical data" means any method enabling the computer system to record a free, informed, and unambiguous agreement from the patient, given to the healthcare professional or collected by the computer system, for the obtaining of their medical data. This agreement, within the meaning of Articles L.Articles III-7 et seq. of the Public Health Code (France), and Article 4(11) of the General Data Protection Regulation (GDPR) of the European Union, as adopted by the European Health Data Space (EHDS) or within the meaning of the Health Insurance Portability and Accountability Act (HIPAA) and the 21st Century Cures Act in the United States, the Personal Information Protection and Electronic Documents Act (PIPEDA) in Canada, the Privacy Act 1988 and the Australian Privacy Principles (APPs) in Australia or the Personal Information Protection Act (APPI) in Japan, may also in some cases be a simple oral non-opposition, if this is preceded by complete and understandable information, and if the patient has clearly understood the implications of the non-opposition obtained.

[0050] Preferably, the preliminary step of recording patient consent may involve technical means and / or sub-steps enabling the computer system to ensure that the patient's consent is: a. free, that is to say freely expressed by the patient; b. informed, the patient having received beforehand, in particular from the health professional, all the information necessary to understand the implications of his / her agreement; c. unambiguous, that is to say expressed clearly and unambiguously by the patient.

[0051] Preferably, the preliminary step of recording patient consent may involve technical means and / or sub-steps enabling the computer system to collect the patient's consent to the acquisition of their medical data, and where applicable, to delimit: a. the scope of the patient's consent, including the authorizations, limitations and prohibitions relating to the use of medical data expressed by the patient, b. the duration of the patient's consent, including any time limit set by the patient and / or any method of revocation expressed by the patient.

[0052] It can be anticipated that the computer system will include technical means allowing the patient to manage the lifecycle of their consent, and in particular to modify the scope and / or duration of the consent, or even to revoke this consent.

[0053] In one embodiment of the invention, the computer system may include means for reading a healthcare professional's card equipped with digital identification means and means for reading a patient's card also equipped with digital identification means. Where applicable, said cards may be smart cards, or Healthcare Professional Cards (CPS), containing a digital authentication certificate, in particular of type X.509, comprising a double public-private key pair, the computer system being capable of communicating with one or more authentication servers capable of verifying that the digital authentication certificates of a smart card are valid.

[0054] It may also be envisaged that the means of digital identification of the healthcare professional will be dematerialized, and may in particular include one or more of the following means: a national identity federation system, an internal system of the healthcare structure to which the healthcare professional belongs, an e-CPS card, or any other dematerialized version of the CPS card, accessible via a secure mobile application, ProSanté Connect identifiers, or any other strong authentication service allowing healthcare professionals to access digital health services securely, a FIDO2 key (Fast IDentity Online), or a European Health Professional Card (EHP card), an identifier of the type National Provider Identifier (NPI) in the United States or of the type NHS Digital Identity and Access Management (IAM) in the United Kingdom; or of the type Healthcare Provider Identifier - Individual (HPI-I) in Australia.It will also be possible to implement digital patient identification methods, which could include AppCV identifiers or any other digital version of the traditional health insurance card, enabling secure patient identification via smartphone; a digital European Health Insurance Card (EHIC); a German electronic health insurance card (eGK); identifiers linked to electronic health systems (EHRs) compliant with HIPAA standards in the United States; an NHS identifier in the United Kingdom; or an Individual Healthcare Identifier (IHI) in Australia. These various digital identification methods will allow for robust and secure authentication of system users, while adapting to technological developments in the field of digital health.

[0055] In this example, the authentication step for the patient's identity, the healthcare professional's identity, and the recording of the patient's consent to the collection of this medical data includes, after the reading of each patient's and healthcare professional's card by the reading device, a sub-step of receiving validation, by the authentication server(s), of each of the digital authentication certificates for these cards. The patient's and healthcare professional's identities are thus verified. This fulfills the prerequisite for qualifying the digital identity to carry the patient's consent information.

[0056] Alternatively, other technical means and / or sub-steps may be provided to verify the identities of the patient and the healthcare professional and to record or even collect the patient's consent, these technical means and / or sub-steps including digital forms, electronic double authentication methods, electronic signature methods, digital medical data management platforms equipped with consent management tools or audio, video or biometric authentication and / or validation methods.

[0057] Preferably, the computer system may include computer memory in which information relating to the identities of the patient and the healthcare professional and the patient's consent is stored in a logged manner and which is exchanged in the computer system during the implementation of said step of authentication of the identity of the patient, the identity of the healthcare professional and recording of the patient's consent to obtaining this medical data.

[0058] In one embodiment of the invention, the medical data analysis step of the provided game includes a sub-step of extracting, from each medical data point, at least one pair of information that can be linked to a patient care episode and a date related to that information, said summary of care episodes being generated from said information and dates extracted from said medical data of the provided game.

[0059] It will be possible to anticipate that several extraction methods can be applied to all or part of the patient's medical data, or even to the same patient's medical data.

[0060] In one embodiment of the invention, the extraction substep comprises a segmentation step, from the content of at least one of said medical data, of information relating to a pathology, a symptom, a medical history, a reason for consultation, a diagnosis, a medical observation, an auscultation result, a biomedical measurement, a result of a medical examination, a treatment, a prescription, and / or a reason for prescribing. Said segmented information may thus form said information that can be linked to a patient care episode, or may be combined and / or transformed to form said information that can be linked to a patient care episode, and / or may allow the extraction of said information that can be linked to a patient care episode.It can be anticipated that other types of information likely to contribute to the direct or indirect identification of a care episode may be segmented from the content of said medical data.

[0061] Where appropriate, said extraction substep includes a segmentation step, from the content of at least one of said medical data, of at least one date of said medical data that can be linked to the segmented information, such as a date of a medical report, a date of a prescription, a date of a consultation, a date of admission or discharge from a health facility.

[0062] In an example of an embodiment of the invention, in which the medical data is structured data, organized according to a predefined format in which each element of the medical data is stored in a specific field of the medical data identified by a given identifier, the segmentation substep involves the selection of one or more elements stored in fields of the medical data identified by predetermined identifiers.According to this example, it is possible to directly extract information, dated and likely to be linked to a patient care episode, from medical data when this medical data is structured, which can be the case for medical data such as medical analysis results which can be formatted according to an HL7 or FHIR standard; medical prescriptions which can be formatted according to a drug database such as a VIDAL® or THERIAQUE® database or the Single Drug Interoperability Reference (RUIM); or one or more ANSM databases; medical reimbursement data, which can be formatted according to a NOEMIE® standard or in which medical acts can be coded according to a classification such as the CCAM, the General Nomenclature of Professional Acts (NGAP) and health conditions can be coded according to a classification such as ICD, and in particular ICD9, ICD-10, ICD-11, or SNOMED.

[0063] In another alternative or cumulative embodiment of the invention, the segmentation step is implemented using at least one machine learning algorithm. Where appropriate, the machine learning algorithm may be trained beforehand to segment, from the content of a patient's medical data, information likely to contribute to the direct or indirect identification of a patient care episode, and in particular information relating to a pathology, symptom, medical history, reason for consultation, diagnosis, medical observation, auscultation result, biomedical measurement, result of a medical examination, treatment, prescription, and / or reason for prescription. It is thus possible to extract information likely to contribute to the direct or indirect identification of a patient care episode from unstructured medical data.

[0064] In the present invention, the term "machine learning algorithm," also called an artificial intelligence (AI) model or a machine learning system, refers to a set of mathematical instructions and rules designed to analyze data and make predictions with minimal human intervention. These algorithms are characterized by their ability to improve iteratively through a training process, during which they adjust their internal parameters based on input data and expected results, in order to optimize their performance on a specific task.

[0065] In the context of the present invention, at least one piece of information, capable of contributing to the direct or indirect identification of a patient care episode, may be segmented or predicted using one or more machine learning algorithms, implemented in parallel or sequentially, to which the aforementioned medical data is provided as input. By way of non-limiting example, this or these machine learning algorithms may be chosen from among: artificial neural networks, deep neural networks (DNNs), convolutional neural networks (CNNs), recurrent neural networks (RNNs), transformers or encoder-decoders, large language models (LLMs), support vector machines (SVMs), random forests, or any appropriate combination of these models.

[0066] In an example of an embodiment of the invention, the segmentation step is implemented using at least one of the following algorithms: a. a natural language processing (NLP) type machine learning algorithm trained to transform textual medical data into a set of text tokens, or tokenizer, and to convert each text token into a feature vector including in particular vector representations of words, or embeddings, pre-trained from a medical corpus; b. a named entity recognition (NER) type machine learning algorithm, trained to generate contextual representations of text tokens from a text according to the two reading directions of the text, to label from these representations each token according to a predetermined type of medical entity of which it is a part and to identify sequences of labels corresponding to predetermined medical entities.

[0067] It can be anticipated that the segmentation step will include a step of selecting at least one machine learning algorithm from a set of predetermined models depending on the type of medical data to be segmented, the segmentation step being implemented by means of said at least one selected machine learning algorithm.

[0068] Where appropriate, said medical data will be provided as input to the first algorithm and the vector representations from the first algorithm will be provided to the second algorithm, the medical entities recognized by the second algorithm allowing to segment said information likely to participate in the identification, direct or indirect, of a patient care episode.

[0069] The first algorithm could be a transformer, such as a BERT (Bidirectional Encoder Representations from Transformers) model, trained unsupervised, for example using a hidden language modeling method, on a training corpus of unstructured medical texts, such as anonymized electronic medical records, synthetic data, hospital reports, and medical scientific articles, and an initial dictionary composed of a set of tokens belonging to the medical vocabulary. The training corpus could, for example, consist of documents from shared databases or partner data warehouses.

[0070] For example, we can predict that the second algorithm is in NLP, or a BiLSTM-CRF (Bidirectional Long Short-Term Memory - Conditional Random Fields) for example comprising two hidden layers of 256 LTSM units each, trained in a supervised manner, on a training corpus composed of a part of the previous training corpus, each text being previously manually annotated to identify a medical entity of the type PATHOLOGY, SYMPTOM, ANTECEDENT, REASON FOR CONSULTATION, DIAGNOSIS, MEDICAL OBSERVATION, AUSCULTATION RESULT, BIOMEDICAL MEASUREMENT, EXAMINATION RESULT, TREATMENT, PRESCRIPTION.

[0071] Advantageously, it may be stipulated that the step of segmenting at least one date of said medical data that may be linked to the segmented information be implemented using at least one machine learning algorithm. Where appropriate, this may be the same machine learning algorithm(s) used for the segmentation of said information that may contribute to the direct or indirect identification of a patient care episode.

[0072] In yet another alternative or cumulative embodiment of the invention, the extraction substep includes a segmentation step, from metadata of at least one of said medical data, of information relating to the health establishment that originated said medical data, of information relating to the context of medical care concerned by said medical data, of information relating to the identity and / or the specialty of medical practice of one or more health professionals who participated in the generation of said medical data, of information relating to the nature of said medical data and / or the dates of said medical data.

[0073] In the present invention, "medical data metadata" means structured information that describes, dates, explains, locates, or facilitates the storage, use, or management of the medical data itself. This may include: a. Information about the origin of the medical data, such as its author or source or its dates of creation, validation, dissemination and modification or its version or its link with parent or related information; b. Information relating to the type and format of the medical data, such as the type of document (report, analysis result, prescription) or the file format (image, text, pdf); c. Information relating to confidentiality and access rights associated with medical data; d. Information relating to the context of acquisition of the medical data, such as an emergency service, hospitalization, outpatient service, or home care; e. Information relating to the classification or categorization of the medical data, such as keywords or classification codes. f. Patient information, such as identity details (identifiers, name(s), first name(s), date of birth, place of birth), administrative data (insurance identifiers, coverage of specific pathologies or long-term illness, reimbursement information).

[0074] It is therefore possible to identify or infer, from this metadata, information that may be linked to a patient care episode.

[0075] In one embodiment of the invention, it may be provided that the information potentially linked to a patient care episode is inferred from the information relating to a pathology, symptom, medical history, reason for consultation, diagnosis, medical observation, auscultation result, biomedical measurement, medical examination result, treatment, prescription and / or reason for prescription, and from one or more additional databases. It is thus possible to enrich the segmented data derived from the patient's medical data, regardless of the segmentation method used, with data not belonging to the patient's medical data, in order to further improve the identification of the patient's care episodes.

[0076] Where appropriate, said extraction substep may include a subsequent step of connecting the computer system to a database associating a plurality of information of the same type likely to be linked to a patient care episode and each associated with at least one piece of information relating to a pathology, a symptom, a medical history, a reason for consultation, a diagnosis, a medical observation, an auscultation result, a biomedical measurement, a result of a medical examination, a treatment, a prescription and / or a reason for prescription, and a step of selecting, from said database, information likely to be linked to a patient care episode from said information segmented from the content of said patient medical data.

[0077] For example, said segmentation step may be a segmentation step of a drug identifier from the content of the medical data and said information likely to be linked to a patient care episode said may be a therapeutic indication inferred from a database of therapeutic indications each associated with one or more drug identifiers.

[0078] Advantageously, it can be foreseen that, during the implementation of the extraction substep, different segmentation methods, such as those described above, are iteratively implemented on each of the medical data in the set of medical data, these segmentation methods being able to be applied sequentially or in parallel, in whole or in part, on each of the medical data, in particular according to the type of medical data and according to the results obtained from the previous segmentation methods.

[0079] In one embodiment of the invention, the medical data analysis step of the provided set includes a substep of identifying, for each pair of information that can be linked to a patient care episode and a date relating to that information, at least one patient care episode to which said information is linked, the care episodes being ordered chronologically using said dates relating to the information to form said summary of care episodes.

[0080] In the identification substep, the aim is to associate each pair of information and a date with a care episode that can be inferred from that pair, and to group the information and date pairs relating to the same care episode. The synthesis of care episodes thus allows the healthcare professional to visualize the chronology of care episodes.

[0081] Preferably, in the identification substep, at least one patient care episode is identified from each information-date pair to which it is associated. This care episode may be associated only with that pair or with other information-date pairs. If necessary, the same information-date pair may be associated with several care episodes. This avoids having to manage conflicts in the associations of information-date pairs with care episodes, which could be detrimental when the healthcare professional analyzes the care summary.

[0082] Advantageously, in the identification substep, pairs of information and dates whose information is linked to an identical or similar reason for seeking care can be associated with the same care episode. If so, the care episode will be characterized by this reason for seeking care.

[0083] For example, in the identification substep, pairs of information and date whose information is linked to identical or similar reasons for consultation, such as symptoms that are wholly or partly identical or to reasons relating to the same organ or anatomical area, can be associated with the same episode of care.

[0084] Similarly, in the identification substep, pairs of information and data whose information is related to identical or similar diagnoses, such as the same diagnosed or probable health condition or health conditions targeting the same organ or anatomical area, can be associated with the same episode of care.

[0085] Similarly, in the identification substep, pairs of information and date whose information is related to identical or similar care, such as the same care prescribed regularly or performed by the same health professional, or similar or identical care prescribed or performed by different health professionals of similar or identical specialties, or care prescribed or performed by different health professionals of the same health facility, can be associated with the same episode of care.

[0086] Advantageously, in the identification substep, pairs of information and date whose dates are separated by a time interval less than a given threshold value can be associated with the same episode of care.

[0087] These associations of different couples with the same care episode can be conditioned on a similarity of information between these couples, such as similar reasons for consultation, patient symptoms, diagnoses, and / or prescribed treatments. This association can also be conditioned on the recurrence or regularity of the same event reported by these couples, such as a consultation, admission, medical examination, and / or prescription renewal. Furthermore, this association can be conditioned on a temporal and / or logical chain of a sequence of documents and / or events, such as a closely spaced sequence of consultations and / or medical examinations, or a closely spaced sequence of consultations and care within the same medical facility.

[0088] In an example of the implementation of the identification substep, each information-date pair can be associated with a care episode or grouped with another information-date pair associated with a patient care episode by means of at least one of the following identification and grouping, or clustering, algorithms: a. A grouping algorithm based on medical terminologies such as SNOMED, ​​ICD, the grouping being carried out according to relational terminological axes such as hierarchical relations, including parent-child inclusions, attributive relations, including the qualification of a main concept, associative relations including causality, anteriority, manifestation, evaluation criteria, relational groups, including functional groupings and according to a common context, conceptual relations, including a concept adapted to the context, and logical relations, including according to a structured categorization of pathology; b. A K-means clustering algorithm, DBSCAN, or a hierarchical model, trained to identify natural clusters, particularly those based on similarity and shared characteristics; c. A probabilistic algorithm such as an LDA (Latent Dirichlet Allocation) algorithm, trained to identify one or more care episodes from a given piece of information; d. A multiaxial density clustering algorithm, such as DBSCan, trained to identify patterns or chains of care episodes; e. A Self-Organizing Maps (SOM) type algorithm or any other machine learning algorithm capable of reducing the dimensional complexity of care episode groupings.

[0089] In one embodiment of the invention, the medical data analysis step of the provided dataset includes a substep of grouping the medical data into groups of medical data, each group of medical data being associated with the same care episode. The information groups are associated with the ordered care episodes to form said care episode summary. In this embodiment, the care episode summary allows the healthcare professional to visualize the chronology of care episodes and access the various medical data relating to these episodes, in order to refine their understanding of the patient's health status, improve their interpretation, and increase their confidence in the tool.

[0090] In one embodiment of the invention, the step of evaluating at least one matching score includes a substep of generating a matching predicate from each set of eligibility criteria provided, and a substep of evaluating an outcome of the predicate to which is provided as an argument said summary of care episodes.

[0091] In the present invention, "correspondence predicate" means a logical function or expression which, when applied to one or more arguments, returns a truth value, namely true or false.

[0092] It is possible to stipulate that each eligibility criterion in a set of eligibility criteria be transformed into a matching predicate to be evaluated using all or part of the summary of care episodes. Alternatively, it is possible to stipulate that only some of these eligibility criteria in a set are transformed into matching predicates to be evaluated using all or part of the summary of care episodes, with the other criteria being evaluated using other patient information not included in the summary of care episodes. Alternatively, it is also possible to stipulate that several, or even all, of the eligibility criteria in a set are transformed into a single matching predicate to be evaluated using all or part of the summary of care episodes.

[0093] It can be expected that the generation of a matching predicate from a set of eligibility criteria will be such that the predicate exactly matches the eligibility criteria or, alternatively, that the predicate will be more permissive or more exclusive than the eligibility criteria. For example, an eligibility criterion requiring "HPV infection" could lead to the generation of a predicate of the type "probability of HPV infection greater than 80%".

[0094] Preferably, said matching predicate may include a function for evaluating a value from the synthesis of care episodes and a comparison of said value to a given threshold value.

[0095] According to another example of implementation, it may be foreseen that the step of evaluating at least one correspondence score is implemented by at least one machine learning algorithm, in particular of a generative type, such as a transformer, to encode the set of eligibility criteria and the synthesis of care episodes in vector representations, the correspondence score being evaluated by predicting a probability of correspondence between said vector representations, in particular by means of a machine learning algorithm, for example of the random forest type, or by evaluating a similarity score between said vector representations.

[0096] In one embodiment of the invention, the selection or exclusion step of each medical study may include a comparison of the correspondence score determined for the set of eligibility criteria associated with that medical study to a predetermined threshold value.

[0097] In one embodiment of the invention, the method includes a step of displaying on a computer terminal of the computer system said summary of care episodes and the eligibility notification, each care episode being displayed with at least one graphic component with which a user of the computer terminal can interact to display the medical data relating to that care episode.

[0098] Advantageously, in the display stage, the eligibility notification may be displayed along with information about the selected medical study(ies) and / or information about the procedures for inclusion and participation in the selected medical study(ies). This facilitates the work of the healthcare professional in informing the patient and obtaining their informed consent for inclusion in a medical study.

[0099] Advantageously, in the display stage, the graphical component associated with a care episode can display the groups of medical data associated with that episode, and potentially the reasons for grouping this medical data. This allows healthcare professionals to easily navigate the summary of care episodes.

[0100] If desired, the process may include a step to evaluate a confidence index for the identification of each identified care episode, for example, calculated from the associations made during the identification substep. A weight may be assigned to each association method that could be used in this substep; the weights assigned to all the associations that enabled the identification of a care episode from the information-date pairs may be combined to calculate the confidence index, for example, by addition.

[0101] If applicable, in the display step, each care episode can be shown with the confidence index for the indication of that care episode. This allows the healthcare professional to easily assess the relevance of the summary of care episodes and interpret it appropriately.

[0102] In one embodiment of the invention, the method includes a step, implemented by the computer system, for recording the patient's consent to be contacted in connection with at least one of the selected medical studies and / or to have patient information transmitted to an organization responsible for at least one of the selected medical studies. The method thus enables the patient's inclusion in the recruitment phase of the medical study, whether interventional or observational.

[0103] It should be noted that the coupling of the recording of the patient's consent to be contacted or to have this information transmitted, with the confirmation of the patient's identity implemented in the prior authentication step, reinforces the legal security required for this collection of the patient's consent.

[0104] In the present invention, "recording of patient consent to be contacted or to have their information transmitted" means any method enabling the computer system to record the patient's free, informed, and unambiguous agreement, given to the healthcare professional or collected by the computer system, to be contacted in connection with at least one of the selected medical studies and / or to have patient information transmitted to an organization responsible for at least one of the selected medical studies. This agreement, in the case of an interventional study, must be explicit consent and may be, in the case of an observational study with anonymized data, a simple oral non-objection or explicit consent.

[0105] Preferably, the patient consent recording step may involve technical means and / or sub-steps enabling the computer system to ensure that the patient's consent is: a. free, that is to say freely expressed by the patient; b. informed, the patient having received beforehand, in particular from the health professional, all the information necessary to understand the implications of his / her agreement; c. unambiguous, that is to say expressed clearly and unambiguously by the patient.

[0106] Advantageously, the said consent registration step may be preceded by a display step on a computer terminal of the computer system, for example on the computer terminal of the healthcare professional or on a digital platform for managing the patient's medical data, of one or more of the following pieces of information, for each of the selected medical studies: objective of the study, duration of the study, identities and contact details of the sponsor and the person in charge of the study, benefits and risks, alternatives, compensation, inclusion procedures, type and / or list of data collected, conditions for processing medical data and / or any other information likely to inform the patient's consent, complying where appropriate with the regulations on information and consent specific to research involving human subjects (in particular the obligations related to the Jardé law in France);and / or an interface for collecting patient consent, including in particular an interface for selecting the means of communication of patient data.

[0107] Preferably, the patient consent registration step may involve technical means and / or sub-steps enabling the computer system to collect the patient's consent to the acquisition of their medical data, and where applicable, to delimit: a. the scope of the patient's consent, and in particular the authorizations, limitations and prohibitions relating to the use of medical data expressed by the patient,

[0108] the duration of the patient's consent, and in particular any time limit set by the patient and / or any method of revocation expressed by the patient.

[0109] Advantageously, the consent registration step may include a storage step, specifically within the database of the selected medical study for which the patient has given consent, of the patient's medical data necessary for said study and information enabling the patient's consent to be time-stamped. Where appropriate, this information enabling the time-stamping of the patient's consent may be provided for as an electronic signature associated with the patient's medical data necessary for said study.

[0110] It can be anticipated that the computer system will include technical means allowing the patient to manage the lifecycle of their consent, and in particular to modify the scope and / or duration of the consent, or even to revoke this consent.

[0111] Advantageously, the said consent registration step may be followed by a step of displaying or transmitting to a computer terminal of the computer system a notification to the patient including a set of information relating to the selected medical study and which was the subject of the patient's consent, such as the methods of sharing and confidentiality of data, the right of withdrawal and / or objection, the provisional schedule of the study.

[0112] The invention also relates to a computer system arranged to implement the process according to the invention.

[0113] The invention also relates to a computer program comprising program code which is designed to implement the method according to the invention.

[0114] The invention also relates to a data carrier on which the computer program according to the invention is recorded.

[0115] Brief description of the figures

[0116] Other advantages and features of the present invention are now described by means of purely illustrative and in no way limiting examples of the scope of the invention, and from the accompanying figures, in which:

[0117] [Fig. 1] represents, schematically and partially, a computer system for the implementation of a process according to an example of an embodiment of the invention;

[0118] [Fig. 2] represents, schematically and partially, a method for processing medical data of a patient during a medical consultation by a healthcare professional and for evaluating a match between the patient and at least one medical study according to an example of an embodiment of the invention, implemented using the computer system of [Fig. 1];

[0119] [Fig. 3] represents, schematically and partially, an example of medical data provided in a step of the process of [Fig. 2];

[0120] [Fig. 4] schematically and partially represents a segmentation of the medical data from [Fig. 3] implemented in a step of the process in [Fig. 2]; and

[0121] [Fig. 5] represents, schematically and partially, a grouping of medical data to identify a care episode in a step of the process of [Fig. 2];

[0122] [Fig. 6] represents, schematically and partially, an example of a synthesis of care episodes displayed in a step of the process of [Fig. 2];

[0123] [Fig. 7] schematically and partially represents examples of sets of eligibility criteria provided in a step of the process in [Fig. 2]; and

[0124] [Fig. 8] represents, schematically and partially, evaluations of correspondence scores between the sets of eligibility criteria of [Fig. 7] and the synthesis of care episodes of [Fig. 6] during a step of the process of [Fig. 2],

[0125] In the description that follows, identical elements, by structure or by function, appearing on different figures retain, unless otherwise specified, the same references.

[0126] The processes described below can also be implemented by software programs executable by a computer system. Furthermore, their implementation can be achieved through distributed processing and / or parallel processing, particularly for processing multiple data points simultaneously.

[0127] The figures described in this document are intended to provide a general understanding of the invention in various embodiments. These figures are not intended to serve as a complete description of all the elements and features of the devices, processors, and systems necessary for the invention. Many other embodiments of the invention, or combinations thereof, may be apparent to those skilled in the art upon reading this description, by combining the disclosed embodiments. Other embodiments may be derived from the description, so that structural and logical substitutions and changes may be made without departing from the scope of the present invention.

[0128] Furthermore, the description and figures should be considered illustrative rather than restrictive, and the appended claims are intended to cover all modifications, improvements, and other embodiments of the invention. Therefore, the scope of the following claims should be determined by the broadest possible interpretation of the claims and their equivalents and should not be restricted or limited by the preceding description.

[0129] Description of the implementation methods

[0130] [Fig. 1] A computer system 1 for implementing a method of processing a patient's medical data during a medical consultation by a healthcare professional and evaluating a match between the patient and at least one medical study according to an example of an embodiment of the invention.

[0131] In the example in [Fig. 1], the computer system 1 includes, on the healthcare professional's side, a computer terminal 2 that the healthcare professional can operate like a computer. Healthcare software 21, designed for processing medical data and managing patient records, is accessible from this terminal 2. The software 21 can be installed locally on the terminal 2 or it can be hosted on a remote server (not shown) as software as a service, or SaaS, this software being accessible via a web browser.

[0132] The computer system 1 also includes an automatic reading device 22 (smart card reader, RFID reader, QR code reader, barcode scanner), connected to the computer terminal 2 in order to exchange data with the software 21. It may be envisaged that the system 1 includes several automatic reading devices 22 in order to be able to read different formal user recognition formats (smart cards, RFID signals, QR code, barcode).

[0133] Alternatively, the computer system 1 can be configured to allow for paperless authentication of the patient and / or healthcare professional. In this case, the computer terminal 2 can be equipped with a specific software module, integrated into the software 21 or operating independently, capable of interacting, for example, via a wireless communication interface, with secure mobile applications installed on users' smartphones. This module can thus support authentication via smart cards carrying electronic certificates, cryptographic identification tokens, QR code reading based on an external identification system, or "eCards" such as the ApCV (digital health insurance card) for patients or the eCPS (digital healthcare professional card) for healthcare professionals.In the case of digital authentication, system 1 will also be able to integrate additional security mechanisms, such as biometric verification (fingerprint, facial recognition) via the patient's or healthcare professional's smartphone, or the use of a one-time code (OTP) generated by the mobile application, in order to strengthen the security of the authentication process.

[0134] The computer system 1 also includes a computer server 3, remote from terminal 2. The server 3 hosts software 31 offering data processing services for the implementation of certain steps of the process according to the invention, described later.

[0135] In the example described, the health software 21 can access the services of the software 31 through an application programming interface (API) and a wireless telecommunications network 6 to which the server 3 and the terminal 2 are connected.

[0136] Computer system 1 also includes other computer servers 4 and 5.

[0137] Server 4 hosts software 41 offering access services to a database 42 containing medical data relating to one or more patients.

[0138] Base 42 can be either a centralized medical data database from different sources, such as the Shared Medical Record (DMP), or a database managed by a single source, such as a health insurance company or a medical structure or a group of medical structures, or a database responsible for collecting data acquired by connected objects or mobile health or wellness applications used by patients.

[0139] Server 5 hosts software 51 offering access services to a database 52 containing data relating to medical studies.

[0140] Base 52 can be a database of a research organization, a medical study sponsor, an investigating center, a coordinator or a contract research organization (or CRO).

[0141] Software 21 can access, through software 31 and under certain conditions, the data in databases 42 and 52 through application programming interfaces (APIs) and the wireless telecommunications network 6 to which servers 4 and 5 are connected.

[0142] It can be foreseen that the computer system 1 comprises several servers 4 and / or 5, each hosting a medical database and / or a database of medical studies, which can be accessed, under certain conditions, by the software 21 through the software 31.

[0143] Other architectures for the computer system 1 than the one shown in [Fig. 1] may be considered without departing from the scope of the present invention. In particular, it may be possible to provide that one or both of the databases 42 and / or 52 are hosted by the server 3.

[0144] In connection with [Fig. 2], we will now describe a process for processing a patient's medical data during a medical consultation by a healthcare professional and for evaluating a match between the patient and at least one medical study, implemented using computer system 1.

[0145] The procedure described in [Fig. 2] is implemented during a medical consultation of a patient by a healthcare professional.

[0146] In stages not shown, the patient may continue their journey within a hospital or medical facility. This journey may include a reception stage, one or more examination stages involving medical imaging, sampling, analysis, surgery, care, or interviews with medical staff; these stages may feed into the healthcare professional's health software.

[0147] During the consultation, a component of the computer system 1 initiates the implementation of the process according to the invention. In the example described, this is the healthcare professional. Alternatively, the process could be initiated autonomously by the healthcare software, particularly when the patient's medical record is loaded.

[0148] Similarly, in unrepresented steps, the healthcare professional may perform operations on the healthcare software, before or during the consultation, such as entering patient information.

[0149] In an authentication step E0, the computer system 1 authenticates the identities of the patient and the healthcare professional and records the patient's consent to the collection of their medical data as part of the consultation, this consent being able to be collected orally by the healthcare professional to be recorded in the computer system.

[0150] In the example in [Fig. 2], this step E0 is implemented using an automatic card reader 22, into which a healthcare professional's health card (CPS) and a patient's health insurance card are successively inserted. These cards are smart cards, or CPSs, containing an X.509 type digital authentication certificate and a public-private key pair. The healthcare software 21 can thus communicate with one or more authentication servers (not shown) to verify that the digital authentication certificate of each of these cards is valid.

[0151] It is possible to have the healthcare professional's identity authenticated before the patient consultation, for example, during their first daily login to the healthcare software, and for only the patient's identity to be authenticated during the consultation. Alternatively, the healthcare professional could be equipped with two card readers, each designed to receive one of the smart cards, with the validity of the digital authentication certificates of the cards being verified simultaneously during the consultation, for example, when the patient's medical record is being loaded.

[0152] In the example described, it is assumed that the authentication of the patient's identity, through the validation of the digital certificate of their health insurance card, allows the computer system 1 to consider that the patient consents to the collection of their medical data, insofar as the patient must voluntarily provide their card to the practitioner. It is conceivable that the health software 21 could also include an interface for the healthcare professional to confirm that the patient has freely, clearly, and unambiguously expressed their consent or non-opposition to the collection of their medical data, and has previously received from the healthcare professional all the information necessary to understand the implications of their agreement.

[0153] Alternatively, other technical means and / or sub-steps may be provided to verify the identities of the patient and the healthcare professional and to obtain the patient's free, informed and unambiguous consent, these technical means and / or sub-steps including digital forms, electronic double authentication methods, electronic signature methods, digital medical data management platforms equipped with consent management tools or audio, video or biometric authentication and / or validation methods.

[0154] In step E01, information relating to the identities of the patient and healthcare professional and the patient's consent, which is exchanged in the computer system during step E0, is stored, in a logged manner, in a computer memory of server 3.

[0155] In step E1, the software 31 connects to the medical database 42 to retrieve a set of medical data (DMi) relating to the patient. Establishing this connection is conditional upon the successful authentication of the patient's and healthcare professional's identities and the successful recording of the patient's consent during the implementation of step E0.

[0156] It can be foreseen that the DMi medical data set will be collected using one or more patient identifiers, for example a social security number from the patient's health insurance card, transmitted by software 21 to software 31, this or these identifiers allowing the selection and extraction of the patient's medical data from database 42.

[0157] By way of non-limiting example, the medical data set may include one or more of the following data: a patient's medical document, a patient's medical administrative data, a patient's medical financial data, data relating to the patient's medical history, data relating to the patient's ongoing medical monitoring.

[0158] Figure 3 shows an example DMi of a patient's medical data that could be collected from database 42.

[0159] This DMi data is an emergency medical report written following the patient's admission to a hospital emergency department. It is unstructured data, whose characters have been pre-recognized, and which includes MTDi metadata indicating, among other things, the document type, the document creation date, the document author and their department, and the document recipients.

[0160] As another example, in step El, the software can also retrieve structured medical data, such as DM2 data stored in json, xml or FHIR ("Fast Healthcare Interoperability Resources") format and hierarchically organized with the following fields and values:

[0161] { { Reason for admission; { Date and time of admission; 2024-08-29, 14:30}; { Reason; Acute respiratory distress due to severe COVID-19}; { Unit; Intensive Care Unit (ICU)}; { Mode of admission; Transfer from the Emergency Department}}; { Clinical data on admission; { Body temperature; 39.2°C}; { MRC dyspnea scale; 4 / 5}}; { Planned monitoring; { Frequency of checks; Every 2 hours}}; { Parameters to monitor; { Respiratory rate}; { Renal function (urea, creatinine)}}; { Primary diagnosis; { ICD-10 code; U07.1 (COVID-19 confirmed)}}; { Attending physician; { Name; Dr. Martin Lefèvre}; { Specialty; Intensive Care Medicine}}.

[0162] In step E2, the patient's DMi medical data is analyzed by software 31 to generate a summary of patient care episodes S.

[0163] For these purposes, in an extraction substep E21, at least one pair of information that can be linked to a patient ESk care episode and a date related to that information is extracted from each medical data DMi.

[0164] Within this E21 extraction substep, different methods of segmenting DMij information that may participate in the direct or indirect identification of an ESk care episode and dDMij dates that may be linked to the segmented DMij information can be applied to each of the DMi medical data.

[0165] This information may, for example, include information relating to a pathology, a symptom, a medical history, a reason for consultation, a diagnosis, a medical observation, an auscultation result, a biomedical measurement, a result of a medical examination, a treatment, a prescription and / or a reason for prescription.

[0166] A first SI segmentation step is implemented using at least one machine learning algorithm to segment, from the content of a medical data DMi, at least one piece of information DMij and a date dDMi that can be linked to this segmented information DMij.

[0167] For example, the SI segmentation step can be implemented using at least: a. a first machine learning algorithm of the natural language processing (NLP) type; b. a second machine learning algorithm of the named entity recognition (NER) type.

[0168] In this example, the first algorithm is a transformer, such as a BERT (Bidirectional Encoder Representations from Transformers) model, trained to transform textual medical data into a set of text tokens, or tokenizer, and to convert each text token into a feature vector including, in particular, pre-trained vector representations of words, or embeddings, from a medical corpus.

[0169] The first algorithm can be trained unsupervised, for example, using a masked language modeling method. It can be trained using a training corpus of unstructured medical texts, such as anonymized electronic medical records, hospital reports, and medical scientific articles, and an initial dictionary composed of a set of tokens belonging to the medical vocabulary. The training corpus could, for example, be drawn from shared databases or partner data warehouses.

[0170] For example, we could predict that the second algorithm is a BiLSTM-CRF (Bidirectional Long Short-Term Memory - Conditional Random Fields) for example comprising two hidden layers of 256 LTSM units each, trained to generate contextual representations of text tokens from a text according to the two reading directions of the text, to label from these representations each token according to a predetermined type of medical entity of which it is a part and to identify sequences of labels corresponding to predetermined medical entities.

[0171] The second algorithm can be trained in a supervised manner, on a training corpus composed of part of the previous training corpus, each text being previously manually annotated to identify for example a medical entity of type PATHOLOGY (Pa), SYMPTOM (S), ANTECEDENT (Ant), REASON FOR CONSULTATION (M_C), DIAGNOSIS (D), MEDICAL_OBSERVATION (O_M), AUSCULTATION_RESULT (R_A), BIOMEDICAL_MEASURE (M_B), EXAMINATION_RESULT (R_E), TREATMENT (Tr), PRESCRIPTION (Pr), and to identify dates.

[0172] Thus, the medical data DMi is provided as input to the first algorithm and the vector representations from the first algorithm will be provided to the second algorithm, the medical entities recognized by the second algorithm allowing to segment DMij information likely to participate in the identification, direct or indirect, of a patient care episode as well as dDMij dates likely to be linked to the segmented DMij information.

[0173] In [Fig. 4], we have represented an example of segmentation of the medical data DMi, several medical entities of type M_C, O_M, M_B, Pa, R_E, D and Pr as well as a date having been segmented from the content of this data.

[0174] Furthermore, entities of type S and Tr were segmented and then reclassified using other databases to which the software 31 can connect. These databases contain classifications of symptoms (ICD-10) and medications (Vidal) in order to refine the segmentation of the DMij information. It is anticipated that the other recognized entities, and in particular entities of type Pa, will also be reclassified using another database.

[0175] The segmented entities thus form DMi information that can participate in the direct or indirect identification of an ESk care episode, and the segmented dates form dDMij dates that can be linked to the segmented DMij information.

[0176] It will be possible to segment the DMij information from the content of unstructured medical data (DMi) using machine learning algorithms other than those described, including artificial neural networks, deep neural networks (DNNs), convolutional neural networks (CNNs), recurrent neural networks (RNNs), transformers or encoder-decoders, large language models (LLMs), support vector machines (SVMs), random forests, or any appropriate combination of these models. Note that, in cases where medical data (DMi) contains unrecognized text, the SI segmentation step may include a preliminary optical character recognition (OCR) step.

[0177] A second S2 segmentation step is implemented to segment, from MTDi metadata of a medical data DMi, at least one DMij piece of information likely to participate in the identification, direct or indirect, of an ESk care episode and / or a dDMij date likely to be linked to this segmented DMij information.

[0178] In the example of the DMi data in [Fig. 3], it will thus be possible to access the various pieces of information and dates contained in the MTDi metadata of the DMi data, such as the name of the healthcare professional who wrote the report and the date the document was created, as well as the recipients of the document.

[0179] A third segmentation step S3 is implemented, replacing the SI step when the medical data DMi is structured data, to segment at least one piece of information DMij and one date dDMi that may be linked to this segmented information DMij.

[0180] Thus, in the case of DM2 data, it is possible to access, in particular, the patient's admission date and the main diagnosis.

[0181] In the different segmentation steps that have been described, it will be possible to infer other DMij information that may participate in the direct or indirect identification of an ESk care episode from information extracted from a DMi medical data and from one or more additional databases to which the software can connect 31.

[0182] It can be anticipated that, during the implementation of the E21 extraction substep, different S segmentation steps, such as the SI, S2 and S3 steps described above, will be iteratively implemented on each of the patient's DMi medical data, these segmentation steps being able to be applied sequentially or in parallel, in whole or in part, on each of the DMi medical data, depending on the type of DMi medical data and according to the results obtained from the segmentation methods previously applied to this DMi medical data.

[0183] At the end of step E21, the software 31 thus obtains, for each medical data DMi, at least one pair of a segmented information DMij and a segmented date dDMij linked to this information, it being understood that several DMij information, or even all of the DMij information segmented from a medical data DMi, can be linked to the same date dDMij segmented from this medical data DMi.

[0184] The set of segmented DMij information and segmented dDMij dates linked to this information can then be used to identify an episode of care for the ESk patient.

[0185] For these purposes, for each medical data DMi, the analysis step E2 includes a sub-step E22 of identification of at least one episode of care ESk of the patient to which these segmented information DMij and these segmented dates dDMij are linked.

[0186] In this sub-step E22, the software 31 seeks to associate each pair of information DMij and date dDMij with an ESk care episode that can be inferred from this pair and to group the pairs of information DMij and date dDMij relating to the same ESk care episode.

[0187] Several association methods are combined in this substep E22, until each pair of information DMij and date dDMij is associated with an ESk care episode.

[0188] According to a first method for associating the E22 identification substep, pairs of DMi information and dDMij date information, where one or more DMij pieces of information are linked to an identical or similar reason for seeking care, can be associated with the same ESk care episode. If so, the ESk care episode will be characterized by this reason for seeking care.

[0189] In the case of the DMi and DM2 data, the diagnostic information for both of these data points to a SARS-CoV-2 coronavirus infection, so this information may possibly be related to the same ESk care episode characterized by a SARS-CoV-2 coronavirus infection.

[0190] According to other methods of association: a. pairs of information DMij and date dDMij where the DMij information(s) are linked to identical or similar reasons for consultation, such as symptoms that are wholly or partly identical or to reasons relating to the same organ or anatomical area, can be associated with the same ESk care episode; b. pairs of DMij information and dDMij date where one or more DMij pieces of information are related to identical or similar diagnoses, such as the same diagnosed or probable health condition or health conditions affecting the same organ or anatomical area, may be associated with the same ESk care episode; c. pairs of DMij information and dDMij date where one or more DMij pieces of information are related to identical or similar care, such as the same care prescribed regularly or performed by the same health professional, or similar or identical care prescribed or performed by different health professionals of similar or identical specialties, or care prescribed or performed by different health professionals in the same health facility, may be associated with the same ESk care episode; d. pairs of information DMij and date dDMij whose dates dDMij are separated by a time interval less than a given threshold value can be associated with the same ESk care episode.

[0191] It should be noted that, in the case of the DMi and DM2 data, the information relating to the healthcare facility that generated the medical data is identical, the prescription information in the DMi data and the unit information in the DM2 data are identical, the information on the patient's symptoms is similar, and the dates relating to this data are identical. All of these similarities make it highly probable that the DM1 and DM2 medical data are linked to the same episode of careESk, characterized by a SARS-CoV-2 coronavirus infection.

[0192] At the end of sub-step E22, the different associations thus make it possible to identify different episodes of care ESk to which are linked, with a certain probability, the pairs of information DMij and date dDMij extracted from the medical data DMi.

[0193] It should be noted that, in the example of [Fig. 2], no sorting is carried out in these links between the DMij - dDMij pairs and ESk care episodes, so that at least one ESk care episode of the patient is identified from each DMi -dDMij pair, to which it is associated, and that the same DMij - dDMij pair can be associated with several ESk care episodes.

[0194] Finally, the analysis step E2 includes a sub-step E23 of grouping the medical data DMi into groups of medical data Gk, each group of medical data Gk being associated with the same care episode ESk identified from links between the DMi data and the care episodes ESk.

[0195] The dDMij dates allow us to establish a start and end of the ESk care episode and to order the DMi medical data in this ESk care episode.

[0196] We have thus represented in [Fig. 5] a chain of medical data DMi to DM5, forming a group Gi associated with an episode of care ESi, characterized by a SARS-CoV-2 coronavirus infection extending from August 29, 2024 to September 10, 2024.

[0197] At the end of the analysis step E2, the software 31 thus results in a synthesis S of patient care episodes formed by the care episodes ESk, ordered chronologically, and by the groups of medical data Gk associated with these care episodes.

[0198] In an E3 step, the health software 21 can display said summary S.

[0199] Figure 6 shows an example of the graphical interface of the health software 21, in which the S summary is displayed.

[0200] Each ESk care episode is displayed, with several graphical components each representing one of the DMi medical data points from the Gk group associated with that ESk care episode. The healthcare professional can then interact with each graphical component to display the DMi medical data point it represents, as well as the relationships of that medical data point with other DMi medical data points in the same Gk group.

[0201] Displaying the summary S of care episodes allows the healthcare professional to view and navigate the chronology of ESk care episodes and access the various DMi medical data relating to these ESk care episodes, and in particular the documents of this DMi medical data, to refine their understanding of the patient's health status, improve their interpretation and increase their confidence in the tool.

[0202] It will be possible to foresee other modes of representation of the synthesis S of care episodes than that of [Fig. 6] without departing from the scope of the present invention.

[0203] Sequentially or in parallel with steps E1 to E3, in a step E4, the software 31 connects to the medical database 52 to obtain a plurality of sets of eligibility criteria Ei_JC m , each Ei_JC game m being associated with a given medical study Ei, in the planning or pre-enrollment phase and which can be either interventional or observational.

[0204] Each Ei_JC game m includes a predetermined number of inclusion criteria Ei_JC m _in and / or exclusion Ei_JC m _e n It can be anticipated that the same medical study Ei may be associated with several sets of eligibility criteria Ei_JC m .

[0205] Thus, three sets of eligibility criteria are represented in [Fig. 7] in tabular form, two sets of which, Ei_JCi and EI_JC2, are associated with the same study Ei, and one set, E2_JCI, is associated with another study E2. It should be noted that some of the inclusion criteria Ei_JC m _in and / or exclusion Ei_JC m _e n These could be non-medical criteria, such as demographic or geographic criteria.

[0206] In step E5, for each set of eligibility criteria Ei_JC m Software 31 evaluates at least one SC correspondence score m between the patient care episode synthesis S, from step E2, and this set of eligibility criteria Ei_JC m .

[0207] To this end, step E5 includes a substep E51 for generating one or more correspondence predicates P m based on each set of Ei_JC eligibility criteria m from stage E4.

[0208] It can be predicted that the Ei_JC criteria m i n and Ei_JC m _e n of a set of eligibility criteria Ei_JC m are directly transformed into correspondence predicates P m or that several, or even all, of the Ei_JC eligibility criteria m _in and Ei_JC m _e n from an Ei_JC game m are transformed into a single correspondence predicate P m .

[0209] In the example in [Fig. 7], each criterion Ei_JC m i n and Ei_JC m _e n is thus transformed into a predicate P m to which the synthesis S can be passed as an argument, and whose result can be 100% if the evaluation of the predicate is true and 0% if the evaluation of the predicate is false. We can predict that the results of a predicate P m not be binary and be evaluated on a different scale.

[0210] In a substep E52, the synthesis S of patient care episodes is passed as an argument to each of the predicates P m determined from each set of Ei_JC eligibility criteria m , in order to evaluate a RP result m of each predicate P m for this summary S.

[0211] In this substep E52, a SC correspondence score m can thus be evaluated based on the RP results m .

[0212] In the example described, each SC correspondence score m This corresponds to a weighted average of the RP results m of each predicate P m from an Ei_JC game m evaluated from the S synthesis, the weights being determined beforehand, for example by the sponsors of the Ei studies, to favor certain Ei_JC criteria m i n and Ei_JC m e n .

[0213] The RP results are thus represented in [Fig. 8]m obtained for each of the Ei_JC criteria m _in and Ei_JC m _e n from the synthesis S represented in [Fig. 6], as well as the SC correspondence scores m obtained from these results. The patient thus obtains an SCi correspondence score of 100% for the first set Ei_JCi of study Ei, an SC2 score of 75% for the second set EI_JC2 of study Ei and an SC3 score of 50% for the set E2_JCI of study E2.

[0214] We can expect to express a SC correspondence score m in other forms, and in particular as a binary result or as a vector or matrix representation, for example, of the RP results m .

[0215] In step E6, software 31 selects Ei studies for which the patient is eligible based on the SC matching score m determined for each set of eligibility criteria Ei_JC massociated with this Ei medical study.

[0216] In the example in [Fig. 8], software 31 thus selects only the Ei studies for which at least one of the SC correspondence scores m determined for the Ei_JC eligibility criteria sets m If the percentage associated with this medical study Ei exceeds a given threshold value, for example 80%, it excludes all other studies. In other words, the software 31 selects the medical study Ei.

[0217] In a step E7, the software 31 can thus generate and transmit to the health software 21 a NE eligibility notification of the patient for the selected medical study(ies), which can then be displayed by the health software 21, for example with the summary S during the step E3.

[0218] This eligibility notification may be displayed along with information about the selected medical study(ies) and / or the procedures for inclusion and participation in the selected study(ies), such as the study objective, duration, the sponsor's and principal investigator's identities and contact information, benefits and risks, compensation, conditions for processing medical data, and / or any other information likely to inform the patient's informed consent. The healthcare professional can thus inform the patient of their eligibility and offer them the opportunity to participate in recruitment for this or these medical studies, providing them with all the necessary information to ensure the patient's informed consent.

[0219] This allows for the collection of consent or non-opposition by the healthcare professional, freely given by the patient, in an informed, clear, and unambiguous manner, to be contacted for at least one of the selected medical studies and / or to have their information transmitted to an organization responsible for at least one of the selected medical studies. This consent can be recorded by the software in a database, for example, by linking it to identity authentication, thereby strengthening the legal security required for obtaining the patient's consent.

[0220] In an unrepresented variant of step E7, the NE notification may be transmitted to a digital platform for managing the patient's medical data.

[0221] In step E8, if the patient informs the healthcare professional that they wish to be included in the recruitment phase of a medical study for which they are eligible, software 31 can display, through healthcare software 21, an interface for obtaining the patient's consent to be contacted regarding at least one of the selected medical studies and / or to have patient information transmitted to an organization responsible for at least one of the selected medical studies. This process thus initiates the patient's inclusion in the recruitment phase of the medical study, whether interventional or observational.

[0222] In a step E9, in the case of an observational study and in the event of the patient agreeing to participate in a recruitment phase of a medical study Ei, it may be provided that the patient's medical data DMi necessary for this medical study Ei and information allowing the patient's consent to transmit this data to be timestamped are stored, for example in a database on server 3.

[0223] It should be noted that different aspects of the various interfaces that have been represented can be envisaged and / or other functionalities added, without going out of the scope of the present invention.

[0224] The preceding description clearly explains how the invention achieves its stated objectives, namely to provide a method for processing a patient's medical data and evaluating a match between the patient and at least one medical study, which allows for the efficient use of this medical data, particularly when it comes from the DMP or an electronic medical record; rapid identification of the patient's eligibility for a medical study, and secure management of data exchanges in compliance with the ethical framework.These objectives are achieved in particular through the generation, during a medical consultation and from this medical data, of a summary of patient care episodes which can be cross-referenced with eligibility criteria for medical studies in order to assess in real time, during the patient consultation, the patient's eligibility for these medical studies and, where appropriate, organize secure management of data exchanges, in compliance with the ethical framework and the patient's consent.

[0225] In any event, the invention cannot be limited to the embodiments specifically described in this document, and extends in particular to all equivalent means and to any technically operative combination of these means.

Claims

Demands

1. A method for processing a patient's medical data during a medical consultation by a healthcare professional and for evaluating a match between the patient and at least one medical study, the method being implemented by a computer system (1) and comprising the following steps: a. (E1) Connection of the computer system to at least one medical database (42) to obtain a medical data set (MDi) relating to the patient; b. (E2) Analysis by the computer system of the medical data in the provided set (MDi) to generate a summary of care episodes (S) of the patient; c. (E4) Connection of the computer system to at least one medical studies database (52) to obtain a plurality of sets of eligibility criteria (Ei_JC m ), each game being associated with a given medical study (Ei); d. (E5) For each set of eligibility criteria provided, evaluation by the computer system of at least one correspondence score (SC m ) between said patient care episode summary and said set of eligibility criteria provided; e. (E6) For each medical study provided, selection or exclusion by the computer system of said medical study based on said match score determined for the set of eligibility criteria associated with that medical study; f. (E7) Generation and transmission of a patient eligibility notification (EN) for the selected medical study(ies) to the patient and / or the healthcare professional.

2. A method according to the preceding claim, characterized in that it comprises a prior authentication step (EO, E01), implemented by the computer system (1), of the identity of the patient, the identity of the healthcare professional and of recording the patient's consent to obtaining this medical data (MDi), the connection step (El) of the computer system to the medical database (42) to obtain the medical data set (MDi) relating to the patient being conditional on the success of the authentication of the identities of the patient and the healthcare professional and on the success of the recording of the patient's consent.

3. A method according to any one of the preceding claims, characterized in that the analysis step (E2) of the medical data (DMi) of the supplied set comprises a substep of extraction (E21), from each medical data, of at least one pair of information (DMij) that can be linked to a patient care episode (ESk) and a date (dDMij) relating to this information, said summary of care episodes (S) being generated from said information and dates extracted from said medical data of the supplied set.

4. A method according to the preceding claim, characterized in that the extraction substep (E21) comprises a segmentation step (SI), from the content of at least one of said medical data (MDi), of information (MDi) relating to a pathology, a symptom, a medical history, a reason for consultation, a diagnosis, a medical observation, an auscultation result, a biomedical measurement, a result of a medical examination, a treatment, a prescription and / or a reason for prescription.

5. A method according to the preceding claim, characterized in that the segmentation step (SI) is implemented by means of at least one machine learning algorithm.

6. A method according to any one of claims 3 to 5, characterized in that the extraction substep (E21) comprises a segmentation step (S2), from metadata (MTDi) of at least one of said medical data (DMi), of information (DMi) relating to the health establishment that originated said medical data, of information relating to the context of medical care concerned by said medical data, of information relating to the identity and / or the specialty of medical practice of one or more health professionals who participated in the generation of said medical data, of information relating to the nature of said medical data and / or dates (dDMi) of said medical data.

7. A method according to any one of claims 3 to 6, characterized in that the analysis step (E2) of the medical data (DMi) of the supplied set includes an identification substep (E22), for each pair of information (DMij) that can be linked to a patient care episode (ESk) and date (dDMij) relating to that information, of at least one patient care episode (ESk) to which said information is linked, the care episodes being ordered chronologically using said dates relating to the information to form said summary of care episodes (S).

8. Method according to claim 7, characterized in that the analysis step (E2) of the medical data (DMi) of the supplied set includes a substep of grouping the medical data into groups (Gk) of medical data, each group of medical data being associated with the same episode of care (ESk), the groups of information being associated with the episodes of care ordered to form said synthesis of episodes of care (S).

9. A method according to any one of the preceding claims, characterized in that the evaluation step (E5) of at least one correspondence score (SC m ) includes a substep of generation (E51) of a matching predicate (P m ) from each set of eligibility criteria (Ei_JC m ) provided, and an evaluation substep (E52) of an outcome (RP m ) of the predicate to which the said synthesis of care episodes is provided as an argument.

10. A method according to any one of the preceding claims, characterized in that it comprises a display step (E3) on a computer terminal (2) of the computer system (1) of said summary of care episodes (S) and of the eligibility notification (NE), each care episode (ESk) being displayed with at least one graphical component with which a user of the computer terminal can interact to display the medical data (DMi) relating to that care episode.

11. A method according to any one of the preceding claims, characterized in that it comprises a step (E8, E9), implemented by the computer system -1), of recording the patient's consent to be contacted in the context of at least one of the selected medical studies (Ei) and / or to have information (DMi) relating to the patient transmitted to a structure in charge of at least one of the selected medical studies.

Images