How Dependable is COVID-19 Data During First Wave? Disclosure of Inconsistencies in Daily Reportage Confirmed Cases and Deaths

The global crisis triggered by the COVID-19 pandemic necessitated precise data monitoring and rigorous analysis efforts from worldwide health authorities and governments, particularly during the pandemic’s initial surge. This study employs Newcomb Benford’s Law specifically to identify potential anomalies in the reporting of COVID-19 data during the pandemic’s first wave. Our methodology encompasses the application of Newcomb Benford’s Law to the first digit analysis focusing on three (3) key statistics [d^∗, α− statistic (α^∗) and ω statistic (ω^∗)] in conjunction with the Kolmogorov-Smirnov test to unveil possible inconsistencies within world continental COVID-19 data reportage. By evaluating the actual distribution of leading digits across various COVID-19 data categories such as cumulative confirmed cases, deaths, recoveries, and active cases against the theoretical distribution proposed by Newcomb Benford’s Law, possible significant deviations were identified. We used the deviation from the Newcomb Benford’s law of anomalous numbers as a proxy for data accuracy. The findings reveal that except for the Australia/Oceania continent which exhibited pronounced deviations due to its unique data structure, the COVID-19 data from all other continents maintained a possible high level of reliability during the initial outbreak. The study concludes that while Benford’s Law is a valuable tool for anomaly detection in diverse data, its use in COVID-19 reportage data shows potential pitfalls. To enhance the effectiveness and reliability of detecting anomalies, the study advocates for integrating additional anomaly detection strategies, like density and boundary based approaches encompassing the local outlier factor and one-class SVM, alongside the Newcomb Benford analysis.