Pierre Squara1. 1. ICU, Clinique Ambroise Paré, 27bd Victor Hugo, 92200, Neuilly-sur-Seine, France. pierre.squara@orangr.fr.
Abstract
BACKGROUND: It is possible to monitor an epidemic evolution by plotting the number of patients, or its log-transform, as a function of time. However, these representations do not allow quick identifications of significant changes in the outbreak; a key information for estimating the needs for hospital and ICU beds, for decision-making, and resource allocation. Moreover, an epidemic is characterised by a heterogeneous evolution that depends on many unpredictable factors, coming from the virus itself or from its ecosystem. Simulations are very complex and based on hypotheses that are impossible to certify a priori, since each outbreak is different and has specific characteristics. A validation phase is necessary that may delay the usefulness of these tools. We tested a simpler method for monitoring the epidemic and rapidly predicting the peak. RESULTS: We present here a simple and easy-to-draw figure by plotting the daily rate of change in the number of patients as a function of time. This allows: (1) to rapidly identify the changes in the infection growth, (2) to extrapolate the regression lines for predicting the peaks, and (3) to use simple statistical models for identifying the significant inflexions and deriving the uncertainties. This figure predicted confidently the peak epidemic of the three waves in France. CONCLUSION: Plotting the daily rate of change in the number of patients as a function of time is a simple tool for monitoring an epidemic growth, allowing to quickly identify significant changes and to help in predicting the peak of the infection, with its confidence interval.
BACKGROUND: It is possible to monitor an epidemic evolution by plotting the number of patients, or its log-transform, as a function of time. However, these representations do not allow quick identifications of significant changes in the outbreak; a key information for estimating the needs for hospital and ICU beds, for decision-making, and resource allocation. Moreover, an epidemic is characterised by a heterogeneous evolution that depends on many unpredictable factors, coming from the virus itself or from its ecosystem. Simulations are very complex and based on hypotheses that are impossible to certify a priori, since each outbreak is different and has specific characteristics. A validation phase is necessary that may delay the usefulness of these tools. We tested a simpler method for monitoring the epidemic and rapidly predicting the peak. RESULTS: We present here a simple and easy-to-draw figure by plotting the daily rate of change in the number of patients as a function of time. This allows: (1) to rapidly identify the changes in the infection growth, (2) to extrapolate the regression lines for predicting the peaks, and (3) to use simple statistical models for identifying the significant inflexions and deriving the uncertainties. This figure predicted confidently the peak epidemic of the three waves in France. CONCLUSION: Plotting the daily rate of change in the number of patients as a function of time is a simple tool for monitoring an epidemic growth, allowing to quickly identify significant changes and to help in predicting the peak of the infection, with its confidence interval.