webiomed
play

WEBIOMED Company K -SkAI CLINICAL DECISION SUPPORT SYSTEM WITH - PowerPoint PPT Presentation

Oct 03, 2023 •382 likes •576 views

WEBIOMED Company K -SkAI CLINICAL DECISION SUPPORT SYSTEM WITH MACHINE LEARNING PROBLEM NON-COMMUNICABLE DISEASES ( NCD ) ARE THE MOST IMPORTANT CAUSE OF MORTALITY IN ALL COUNTRIES: of all deaths 71% $ Trillions - in the world National

  1. WEBIOMED Company «K -SkAI » CLINICAL DECISION SUPPORT SYSTEM WITH MACHINE LEARNING

  2. PROBLEM NON-COMMUNICABLE DISEASES ( NCD ) ARE THE MOST IMPORTANT CAUSE OF MORTALITY IN ALL COUNTRIES: of all deaths 71% $ Trillions - in the world National healthcare costs Proportional NCD mortality 4 TYPES OF DISEASES Cardiovascular diseases 20% ( every third death in the world ) 44% Main GOAL is to reduce the total mortality rate of these diseases by 25% until 2025. Cancer diseases 4% Total NCD deaths DIRECTIONS: 41 million people • Comprehensive prevention 10% Respiratory diseases • Risk factors prediction • Informing the population about Diabetes 22% suspected diseases at an early stage Other diseases (Alzheimer's d., 2 Parkimson’s d.) World Health Organization, 2018. https://apps.who.int/iris/handle/10665/274512

  3. Webiomed Webiomed is cloud-based Clinical Decision Support System Technology features: integration with Electronic health records • (EHR) extraction of full EHR documents and Clinical advantages: • features • various disease assessment complexed approach to the analysis of • • integration of different approaches to the text data (including NLP ) the clinical conditions producing Data Sets for ML • • own methods to the healthcare analysis de-identified electronic health • management records to determine diseases and clinical • targeted recommendations to conditions physicians and patients risk factors prediction by machine • learning and deep learning models 3 http://webiomed.ai/

  4. ADVANTAGES & POINTS OF DIFFERENTIATION WE OFFER CDSS Webiomed help physicians and healthcare providers to determine various diseases and clinical conditions. Webiomed is able to assess various risks, diseases and health related factors. MISSION • to reduce medical errors • to provide high speed processing of EHR big data • to improve the quality of the diagnostic process • to predict high risk diseases probability and clinical conditions WE USE • clinical assessment scales We use machine learning methods • clinical guidelines & scientific documents to produce new healthcare knowledge • machine learning (neural networks) FUTURE • the use of the unified ontological platform compatible with any other CDSS 4

  5. HOW DOES WEBIOMED WORK? Data preprocessing (format-logical control, NLP) Data analysis (algorithms, scales, neural models) Extraction of risk factors De-identified Predictions of the group risks electronic health records for different diseases (JSON) Webiomed Creation of targeted recommendations to Results are sent to EHR physicians and patients EHR, (report, JSON, HTML) according to the clinical guidelines EMR 5

  6. USED METHODS RESULTS: INPUT: Analysis methods Group risk predictions of diseases: Health examinations very high, high, moderate, low Clicinal scale based analysis Lab tests and diagnostics Extraction of additional risk factors Regulatory requirements analysis Instrumental data Identification of hidden diseases Clinical recommendation algorithms Clinical examination results Ambulance calls Prediction of the diseases suspicion ML Forecast of a critical event (complications) Types of diseases 6

  7. EVIDENCE-BASED MEDICINE Clinical scales Guidelines & scientific documents • RISK OF ATHEROSCLEROSIS RISK OF OTHER CARDIOVASCULAR DISEASES 7

  8. DEEP AND MACHINE LEARNING MODELS TO IMPROVE CARDIVASCULAR RISK PREDICTION GOAL: to compare both methods to CVD risk prediction based on extracted EHR data - machine learning and traditional risk scales ML TECHNOLOGY RESULTS ELECTRONIC HEALTH UNSTRUCTURED NLP DATA SET 1 True Positive Rate RECORD DATA 0,8 АМБУЛАТОРНАЯ КАРТА № 27916 0,6 Deep Learning ( ROC AUC = 0,75-0,76) 31 517 patients Logistic Regression ( ROC AUC =0,74-0,76) 3 652 all features patients 0,4 Framingham ( ROC AUC = 0,62-0,72) NEURAL NETWORKS MODEL CLINICAL DECISION SCORE ( ROC AUC = 0,66-0,73) SUPPORT SYSTEM (CDSS) 0,2 PROCAM ( ROC AUC = 0,60-0,69) 0 0,2 0,4 0,6 0,8 1 Individual risks False Positive Rate prediction CONCLUSION PATIENT COHORT  The machine learning outperformed a traditional clinically-used predictive model • Total – 3 652 (have all features: for CVD risk prediction. vital signs ,diagnoses, medications)  This approach was used to create a CDSS. It uses both methods: traditional risk scales and • Average age – 49,4 (21-75) models based on neural network. The system can calculate the CVD risks automatically and • Female - 68,2% recalculate immediately after adding new information to the EHR. EI. Korsakov, A. Gusev, T. Kuznetsova, D. Gavrilov, R. Novitskiy « Deep and machine learning models to improve risk prediction of cardiovascular disease using data extraction from electronic health record » /ESC Congress, Paris.2019

  9. DEMONSTRATION OF MVP SERVICE « WEBIOMED.CHECK-UP » In response to this request Webiomed returns the identified risk factors and the When the physician working by EMR he can ask for artificial intelligence’s advice. It appropriate assessment of group patient risk requires him to push the bottom in the workflow. EMR automatically analyzes the The results are displayed on the system website page electronic record of the patient and sends to the Webiomed de-identified request for its analysis The answer contains detailed explanations and further recommendations for the doctor and the patient WEBIOMED 9

  10. Webiomed extracts DATASET Social data Anamnesis and signaling information MVP version screenshot Medical documents 10

  11. Who needs this service ?(costumers) INSURANCE COMPANY SERVICES FOR PATIENTS PUBLIC HEALTH SYSTEMS PHYSICIANS

  12. Opportunities for PHARMA As a system for analysis and full health records extraction CDSS Webiomed will provide: Health Records Analysis: • Diagnosis • Conditions • Drug therapy Physicians’ support Clinical trials, Research&Dev • Drug therapy prescribing • Data Analysis: individual and general • Drug therapy monitoring: effects, side • Data Sets implementation: important effects, complications, error reduction diseases, locations, nationalities and etc. and etc. • Drug compatibility PHARMACEUTICAL MANAGEMENT 12

  13. Intellectual property CERTIFICATES KEY PUBLICATIONS http://jtelemed.ru/article/iskusstvennyj-intellekt-v-ocenke-riskovrazvitija-serdechno-sosudistyh-zabolevanij Physician and information technologies. 2018. No. 3. pp. 45-60. Physician and information technologies, No. 3, 2017. pp. 92-105. Information society. No. 4-5, 2017, pp. 78-93 Certificate for registering Certificate for Webiomed Physician and information technologies, No. 2, 2017. pp. 60-72. PC software trademark “Clinical Decision Support System Webiomed ” Healthcare Manager. 2014. No. 1. P. 51-60. Software is being registered as a medical device in Roszdravnadzor Physician and information technologies. No.5, 2011 pp. 60-76 13 Medical academic journal. Volume 5. No. 3. 2005. Supplement 7. pp. 64-67

  14. MEDIA ABOUT PROJECT Artificial Intelligence helps physicians to Identify dangerous diseases Artificial intelligence has increased the detection of cancer risk https://www.rosminzdrav.ru/regional In Yamal, artificial intelligence examined _news/11278-iskusstvennyy- factors 30 thousand patients intellekt-pomozhet-yamalskim- https://rg.ru/2019/04/06/reg-urfo/na- https://ntinews.ru/in_progress/likbez/kak-iskusstvennyy- vracham-vyyavlyat-opasnye- iamale-iskusstvennyj-intellekt-obsledoval- intellekt-pomogaet-vracham-v-rabote-.html zabolevaniya-na-rannih-stadiyah 30-tysiach-pacientov.html CDSS must be implemented in healthcare Will artificial intelligence replace doctors? https://www.yanao.ru/presscenter/news/8918/ Increase in the identification of heart disease risk factors http://www.topnews.ru/news_id_129732.h https://www.kommersant.ru/doc/3984543 tml AI finds those who have а risk of a heart attack. https://zdrav.fom.ru/post/zhitelej-muravlenko-kotorym- grozit-infarkt-nahodit-intellektualnaya-sistema « Artificial Intelligence in medicine » Regional Scientific and Practical Conference was held in April, 2019. http://conf.nbmz.ru/ 14

  15. PROJECTS IN RUSSIAN REGIONS Regional pilot projects: • • • Industry projects : • • • 15

  16. OUR PARTNER S AWARDS and ACHIEVEMENTS NATIONAL BASE OF MEDICAL KNOWLEDGE The Association of Developers and Users of Artificial Intelligence Systems in Medicine Digital Health Awards SKOLKOVO The Skolkovo Innovation Center is a high technology business area in Russia ASSOCIATION OF CLINICAL PHARMACOLOGISTS PROF-IT.2019 This is the largest organization of clinical pharmacologists in Russia.Established in 2009 MEDICAL PREVENTION CENTER YNAO Health Organization for prevent diseases #I ТМ 2019 PETROZAVODSK STATE UNIVERSITY Petrozavodsk State University is the Flagship University of the Republic of Karelia

More recommend