Weekly Nowcasting of New COVID-19 Cases Using Past Viral Load Measurements

Abstract

The rapid spread of the coronavirus disease COVID-19 has imposed clinical and financial burdens on hospitals and governments attempting to provide patients with medical care and implement disease-controlling policies. The transmissibility of the disease was shown to be correlated with the patient’s viral load, which can be measured during testing using the cycle threshold (Ct). Previous models have utilized Ct to forecast the trajectory of the spread, which can provide valuable information to better allocate resources and change policies. However, these models combined other variables specific to medical institutions or came in the form of compartmental models that rely on epidemiological assumptions, all of which could impose prediction uncertainties. In this study, we overcome these limitations using data-driven modeling that utilizes Ct and previous number of cases, two institution-independent variables. We collected three groups of patients (n = 6296, n = 3228, and n = 12,096) from different time periods to train, validate, and independently validate the models. We used three machine learning algorithms and three deep learning algorithms that can model the temporal dynamic behavior of the number of cases. The endpoint was 7-week forward number of cases, and the prediction was evaluated using mean square error (MSE). The sequence-to-sequence model showed the best prediction during validation (MSE = 0.025), while polynomial regression (OLS) and support vector machine regression (SVR) had better performance during independent validation (MSE = 0.1596, and MSE = 0.16754, respectively), which exhibited better generalizability of the latter. The OLS and SVR models were used on a dataset from an external institution and showed promise in predicting COVID-19 incidences across institutions. These models may support clinical and logistic decision-making after prospective validation.

Systematic Review on the Correlation Between SARS-CoV-2 Real-Time PCR Cycle Threshold Values and Epidemiological Trends

“…[ 2 ] 2021 USA To develop a new method that uses information inherent in Ct values from RT-qPCR tests to robustly estimate the epidemic trajectory from multiple or even a single cross-section of positive samples Real-time Single unit/closed system Single unit followed by regional The local epidemic trajectory of SARS-CoV-2 was accurately estimated from Ct values of routine hospital admissions Khalil et al. [ 26 ] 2022 Lebanon To use data-driven modeling that utilizes Ct values and previous number of cases to forecast the trajectory of the spread of COVID-19 Retrospective analysis Single unit / local National A polynomial regression and support vector machine regression model using Ct values demonstrated potential for predicting COVID-19 incidences in institutions Lin et al. [ 24 ] 2022 China To analyze the viral load data on confirmed cases during two local epidemics in Hong Kong to explore the possibility of a correlation between temporal changes in the distribution of viral loads (measured by RT-qPCR Ct values) and estimates of R t based on case counts Real-time Local Local A log linear regression was fitted to daily incidence-based R t , on daily mean and skewness of Ct values at sampling during the third wave (considered the training period for this study) for real-time assessment of COVID-19 transmission in the community using Ct values.…”

“…The GP model provided growth-rate estimates that followed those estimated using observed case counts for the whole state Khalil et al. [ 26 ] Hospital NR NR NR NR NR NR National daily COVID-19 confirmed case counts were obtained from the Lebanese Ministry of Public Health and Worldometers website NR 7-day There was a temporal delay between the observed Ct values and the incidence rate with a trough in mean Ct values (on October 8 2020) followed by an increase in case numbers 3 weeks later (on October 29 2020) There was an inverse correlation between mean Ct values and number of cases ( P < 0.001), (Spearman correlation) Six data-driven models that utilized Ct values and number of cases from a previous wave (training dataset) were used to predict the epidemic trajectory. This was evaluated using MSE The sequence-to-sequence model MSE = 0.025 The polynomial regression (OLS) and SVR MSE = 0.1596, and MSE = 0.16754, respectively Lin et al.…”

“…The studies were conducted in the USA [ 2 , 13 – 17 ], Europe [ 18 – 21 ], Africa [ 22 ], Australia [ 23 ], China [ 24 ], India [ 25 ], the Kingdom of Bahrain [ 3 ], and Lebanon [ 26 ]. RT–PCR Ct values were measured from public samples only (national or local), or from hospital and public samples in ten studies [ 3 , 14 , 18 – 25 ], and from hospital only or hospital and care home samples only in six studies [ 2 , 13 , 15 – 17 , 26 ]. Eleven studies included nasopharyngeal swabs [ 2 , 3 , 13 – 18 , 20 – 22 ], one study used nose and throat swabs [ 19 ], and four studies did not record the sample type [ 23 – 26 ].…”

“…RT–PCR Ct values were measured from public samples only (national or local), or from hospital and public samples in ten studies [ 3 , 14 , 18 – 25 ], and from hospital only or hospital and care home samples only in six studies [ 2 , 13 , 15 – 17 , 26 ]. Eleven studies included nasopharyngeal swabs [ 2 , 3 , 13 – 18 , 20 – 22 ], one study used nose and throat swabs [ 19 ], and four studies did not record the sample type [ 23 – 26 ]. At least 30 different polymerase chain reaction (PCR) assays were included in the studies, but not all were described in sufficient detail to give a precise count, with two studies not recording the PCR assay used [ 25 , 26 ].…”

Using Genome Sequence Data to Predict SARS-CoV-2 Detection Cycle Threshold Values

“…For example, Ct values for cross-sectional samples collected from patient populations over a given time period are often indicative of the state of the epidemic, with lower average Ct values indicating a growing pandemic (12)(13)(14). Similar studies have also found that crosssectional trends in Ct values can act as indicators of the future trajectory of the pandemic (15,16). At the patient level, Ct values have been shown to provide a good estimate of how long a patient will remain contagious (10,17,18), and several studies have shown that lower Ct values (i.e., higher viral loads) can be correlated with symptomatic infection, morbidity, and mortality (9,14,19,20).…”