PT - JOURNAL ARTICLE AU - Theodoros Evgeniou AU - Mathilde Fekom AU - Anton Ovchinnikov AU - Raphael Porcher AU - Camille Pouchol AU - Nicolas Vayatis TI - Epidemic Models for Personalised COVID-19 Isolation and Exit Policies Using Clinical Risk Predictions AID - 10.1101/2020.04.29.20084707 DP - 2020 Jan 01 TA - medRxiv PG - 2020.04.29.20084707 4099 - http://medrxiv.org/content/early/2020/05/05/2020.04.29.20084707.short 4100 - http://medrxiv.org/content/early/2020/05/05/2020.04.29.20084707.full AB - Background In mid April 2020, with more than 2 · 5 billion people in the world following social distancing measures due to COVID-19, governments are considering relaxing lock-down. We combined individual clinical risk predictions with epidemic modelling to examine simulations of isolation and exit policies.Methods We developed a method to include personalised risk predictions in epidemic models based on data science principles. We extended a standard susceptible-exposed-infected-removed (SEIR) model to account for predictions of severity, defined by the risk of an individual needing intensive care in case of infection. We studied example isolation policies using simulations with the risk-extended epidemic model, using COVID-19 data and estimates in France as of mid April 2020 (4 000 patients in ICU, around 7 250 total ICU beds occupied at the peak of the outbreak, 0·5% percent of patients requiring ICU upon infection). We considered scenarios varying in the discrimination performance of a risk prediction model, in the degree of social distancing, and in the severity rate upon infection. Confidence intervals were obtained using an Approximate Bayesian Computation approach. The framework may be used with other epidemic models, with other risk predictions, and for other epidemic outbreaks.Findings Based on the data for France as of mid April 2020, simulations indicated that an exit policy considering clinical risk predictions starting on May 11, as planned by the government, could enable to immediately relax restrictions for an extra 10% (6 700 000 people) or more of the lowest-risk population, and consequently relax the restrictions on the remaining population up to two times (or several months) faster, with only a small proportion of the population remaining in isolation for an extended period of time – while abiding to the current ICU capacity. In contrast, implementing the same exit policy without risk predictions would exceed the ICU capacity by a multiple. Sensitivity analyses showed that when the assumed percentage of severe patients among the population decreased, or the prediction model discrimination improved, or ICU capacity increased, policies based on risk models had a greater impact on the results of epidemic simulations. At the same time, sensitivity analyses also showed that differential isolation policies require that higher risk individuals comply with recommended restrictions. In general, our simulations showed that risk prediction models could always improve policy effectiveness, keeping everything else constant, in line with value of information arguments, even for models with moderate discrimination power.Interpretation Clinical risk prediction models should be considered to manage outbreaks using a framework as the one developed. They can inform personalised isolation policies, for example by gradually restricting (relaxing) isolation from the highest (lowest) to the lowest (highest) predicted risk individuals, when such policies are considered. This may lead to both safer and faster outcomes than what can be achieved without such prediction models. They enable personalisation of policies, which are known to improve effectiveness in other non-healthcare contexts.Funding No funding was used for this research.Evidence before this study Several countries have implemented non-pharmaceutical interventions based on social distancing and isolation measures in order to limit the spread of COVID-19. There has been limited differentiation in the degree of isolation measures, except for those critical for the functioning of the healthcare system and other key services. There is limited evidence about the impact of relaxing these measures as this has happened only recently and in only a few countries. Investigating the potential impact of differential restrictions depending on medical factors, such as the risk of severe symptoms if infected by Sars-Cov-2, may inform policies for imposing or relaxing isolation policies when these are considered.Added value of this study This study investigates incorporating clinical risk predictions in epidemic models, allowing to explore isolation policies that consider individual clinical risks using simulations.Implications of all the available evidence Epidemic simulations of isolation policies that consider predicted clinical risks in order to differentiate restrictions depending on risks indicate the feasibility of new, possibly gradual, policies that may be riskier to implement without undertaking this type of risk-based approach.Competing Interest StatementThe authors have declared no competing interest.Funding StatementNo funding was used for this research.Author DeclarationsAll relevant ethical guidelines have been followed; any necessary IRB and/or ethics committee approvals have been obtained and details of the IRB/oversight body are included in the manuscript.YesAll necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived.YesI understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).YesI have followed all appropriate research reporting guidelines and uploaded the relevant EQUATOR Network research reporting checklist(s) and other pertinent material as supplementary files, if applicable.YesNo data. Code is available at https://reine.cmla.ens-cachan.fr/boulant/seair https://reine.cmla.ens-cachan.fr/boulant/seair