Epidemiological Identification of A Novel Pathogen in Real Time: Analysis of the Atypical Pneumonia Outbreak in Wuhan, China, 2019-2020

Abstract

Virological tests have now shown conclusively that a novel coronavirus is causing the 2019-2020 atypical pneumonia outbreak in Wuhan, China. We demonstrate that non-virological descriptive characteristics could have determined that the outbreak is caused by a novel pathogen in advance of virological testing. Characteristics of the ongoing outbreak were collected in real time from two medical social media sites. These were compared against characteristics of eleven pathogens that have previously caused cases of atypical pneumonia. The probability that the current outbreak is due to "Disease X" (i.e., previously unknown etiology) as opposed to one of the known pathogens was inferred, and this estimate was updated as the outbreak continued. The probability (expressed as a percentage) that Disease X is driving the outbreak was assessed as over 29% on 31 December 2019, one week before virus identification. After some specific pathogens were ruled out by laboratory tests on 5 January 2020, the inferred probability of Disease X was over 49%. We showed quantitatively that the emerging outbreak of atypical pneumonia cases is consistent with causation by a novel pathogen. The proposed approach, which uses only routinely observed non-virological data, can aid ongoing risk assessments in advance of virological test results becoming available.

Keywords: Bayes’ theorem; causation; diagnosis; epidemic; prediction; statistical model.

Figure 1

Real-time estimation of the probability that the ongoing pneumonia outbreak is driven by each candidate pathogen, given available information on different days. The probability that the outbreak is due to an unknown pathogen (Disease X) increases as more information becomes available, for two reasons: (i) the current outbreak can be seen to exhibit characteristics that are not similar to those observed in previous outbreaks, and; (ii) previously observed pathogens are ruled out by laboratory test results. Arrows indicate new information available on each date. Results are shown for different metrics describing the distance between characteristics of the ongoing outbreak and each candidate pathogen, and either including or excluding initial exposure information for the current outbreak (i.e., worked at/visited a wet market), specifically: (A) Hamming distance (the sum of squares difference between the entries in the columns of Table 1 corresponding to the ongoing outbreak and the candidate pathogen considered) with wet market exposure; (B) Euclidean distance (the square root of the Hamming distance) with wet market exposure; (C) Hamming distance without wet market exposure; (D) Euclidean distance without wet market exposure. Dashed grey lines show the probability for every pathogen (including Disease X) if the only information included is the ruling out of different pathogens through laboratory tests (i.e., a probability of 1/(1 + number of candidate pathogens remaining on that day)). Note that the probability corresponding to different pathogens can be identical, for example, severe acute respiratory syndrome (SARS) and Mycoplasma pneumoniae were assessing as being equally likely as the causative pathogen from 30 December to 4 January, and Legionellosis and Chlamydia pneumoniae had equal probability from 30 December to 12 January (Details in Supplementary Materials Table S1).