19/06/2019 Almost all areas of health are becoming e-health - PDF document

19/06/2019 Almost all areas of health are becoming e-health … Clinical decision support algorithms: how to ensure their safety and usefulness? Prof. Valérie D’Acremont , MD, PhD Centre for primary care and public health, UNIL (Unisanté) Swiss Tropical and Public Health Institute, UNIBAS (SwissTPH) CISTM16, 9 June 2019 Labrique et al, Global Health: Science and Practice 2013 1 2 What is available on the market? Are electronic clinical decision support algorithms really new..? Play Store Symptom checker Babylon ada Isabel “The times have passed when a single human mind could even pretend to know all that might be useful in aiding patients.” 1964 L.C. Payne, The role of the computer in refining diagnosis, The lancet 2013: Dr. Watson (IBM) 3 4 Babylon at the heart of controversies… Babylon Babylon signs a contract with NHS. “We are one of the safest primary care provider of UK.” 8 complaints filed by GPs in UK. Some would like to see us fail and use 51% of actual diagnoses anonymous and wrong allegations. Some The UK Care Quality Commission concludes even pretend to be physicians… that in some areas Babylon is not safe. were among the top 3 diagnoses (report censured by High Court). provided by the algorithm “Babylon technology is certified as a medical device.” (classe 1) Letter to the BMJ: “Could Babylon please supply evidence?” Internal study with 50 case scenarios: Semigran et al, BMJ 2015 “Babylon do better diagnosis than human Letter to the Lancet: “serious beings.” methodological problems” 2.5 millions people in UK, Rwanda and Ireland are presently using Babylon… 5 6 1

19/06/2019 Ethics and challenges around digital health What type of algorithm for whom? Patient Front-line physician/clinician Specialized physician “A key challenge is to ensure that all people enjoy the benefits of digital technologies for everyone. We must make sure that innovation and technology helps to reduce the inequities in our world, instead of becoming another reason people are left behind. Countries must be guided by evidence to establish sustainable harmonized digital systems, not seduced by every new gadget. SHOULD I GOTO A PHYSICIAN? 1. Refer or admit? You don’t have to go to a physician 2. Which lab test? Ultimately, digital technologies are not ends in themselves; they You don’t need to stay in bed are vital tools to promote health, keep the world safe, and serve 3. Meaning of the result but avoid physical efforts. the vulnerable.” Don’t stay in the sun and drink in the clinical context a lot. Take paracetamol. (specific treatment)? Dose of paracetamol WHO guideline digital interventions 2019 7 8 2 nd step: Structured review of the literature First step: Define target user and patient Community health worker Pharmacist 12’124 articles Child 2 months – 5 years Traveler or migrant upon return from the with history of fever tropics with fever or high temperature Physician at hospital Primary care clinician Kristina Keitel et al, Plos Medicine 2017 9 10 3 rd step: studies to measure disease prevalence 4 th step: CART analyses to best combine clinical predictors Febrile adult travellers Febrile Tanzanian children Skin & soft tissue Mononucleosis 2% 3% Noninfectious 2% Neurological infections Other 3% Sensitivity Specificity 0.6% Genitourinary 46% 93% 4% LR+ LR- Viral diseases 5% 6.57 0.58 Tropical Respiratory infections infections 45% 13% Acute febrile diarrhoea 15% Unknown 14% Based on 25’743 biological tests Buss et al., in preparation D’Acremont et al., NEJM 2014 De Santis et al. Plos One 2017 11 12 2

19/06/2019 5 th step: novel host biomarkers that predict disease 6 th step: clinical decision support algorithm (CDSA) 28-day mortality in Radiological pneumonia in febrile Tanzanian adults febrile Tanzanian children Richard-Greenblattet et al, CID 2019 Erdman et al Plos One 2015 ALMANACH, Clotilde Rambaud et al, Plos One 2013 (Adapted by Olga de Santis) 13 14 7 th step: transform medical thinking into software coding MedAl-C: a software to transform your algorithm into an App   0100*+%=x011 ‐    %0110+x%1001x@01  101+00%@01 15 16 The validation cycle of electronic clinical decision support algorithms I Validity and user-friendliness in the IT lab II Adaptation time & place, Clinical safety and using generated efficacy data Clinical and epidemiological context IV III Impact, Clinical including effectiveness on costs Keitel & D’Acremont, Clin Microb & Infect 2018 17 18 3

19/06/2019 Online prospective study with 539 patient/clinician pairs Development of FeverTravel practice guidelines • Literature review • Construction of decision charts with documentation - quality of evidence “No death was recorded and all - strength of recommendation complications could be attributed to • Reviewed by 15 experts at the CISTM7 (Insbruck, the underlying illness rather than to 2001) adherence to guidelines.» • Publication in Journal of Travel Medicine 2003 • Update of guidelines in 2013 and in 2019 (new assessment by experts planned) Mueller et al, J. Travel Med. 2014 19 20 Pilot study with GPs using FeverTravelApp on simulated patients Algorithm to manage febrile children at primary care level (ePOCT) Glucometer Oximeter Hemoglobinometer Number of questions proposed by the App Malaria 19/31 algorithm CRP Exposures 11/31  diseases to Symptoms be considered PCT Clinical data General questions Recommendation for treatment and/or admission Without App With App Vibert et al, in preparation 21 22 Impact of e-POCT implementation on cure rate Randomized control trial of e-POCT 100 100 3739 children 2 months - 5 years (9 facilities, Dar es Salaam) 98% 98 80 96% Routine ALMANACH e-POCT 96 60 Other Potential impact of ePOCT 95% in children in Tanzania: Pneumonia 94 Severe disease 1 million clinical failures 40 averted per year 92 20 90 Cure rate at D3 and D7; 2 nd hospitalisations and deaths by D30 0 Routine ALMANACH ePOCT Kristina Keitel et al, Plos Medicine 2017 Kristina Keitel et al, Plos Medicine 2017 23 24 4

19/06/2019 Impact of e-POCT implementation on antibiotic prescriptions The validation cycle of electronic clinical decision algorithms I 95% 100 Validity and user-friendliness in the IT lab Other 80 Pneumonia II Adaptation Severe disease time & place, Clinical 60 Potential impact of ePOCT safety and using efficacy in children in Tanzania: generated ePOCT clinical trial data Clinical and 28 million unnecessary epidemiological 40 30% context antibiotics saved per year DYNAMIC project 20 11% IV III Impact, Clinical 0 including effectiveness on costs Routine ALMANACH ePOCT Kristina Keitel et al, Clin Microb & Infect 2018 Kristina Keitel et al, Plos Medicine 2017 25 26 Ecology and durability of smartphones/tablets implementation The Tanzanian DYNAMIC project e-POCT ePOCT : - extended medical content - new software - full connection to biosensors and rapid tests Dynamic algorithm : Validation: Through machine-learning and - 70 health facilities optimization Beneficiaries : - 2 semi-urban districts 500,000 sick children in Tanzania per year attending Health system : primary care facilities Use a FAIRPHONE ! Don’t store useless data on long term ! Enhanced M&E, disease surveillance, epidemic detection 10110100 1010101010 Data sciences : High number and variability of data 27 28 Impact of algorithms beyond health Each technical innovation is doubled sided, not due to the good or bad way of using it, but due to the change in the distribution of power. It removes « If you record in the REC, you learn at the same time. power from some to give it to others, changing the reality for all. If one day there is no tablet, you will still be able to correctly manage the child. » « Yes, it teaches us, as you cannot retain René Berger & Solange Ghernaouti-Hélie, Accoucheuse ,Centre de Santé de Boulma everything in your head. But with the REC, it It improves communication ‘Technocivilisation, pour une philosophie du numérique’, 2010 reminds you at any time. At any time, you have it in front of you and it allows you to master. » Infirmier, Centre de Santé de Samba « Now the clinicians ask us more questions on the child and touch him more. » «On était dans les ténèbres. Président comité de gestion, village de Yako Maintenant, on est dans la lumière. » Chef du village de Yako It brings back pride and autonomy 30 IeDA project from Terre des hommes in Burkina Faso Bessat et al, BMC Public Health 2019 29 30 5