Read on Twitter
How high is the risk of #COVID19 transmission among train passengers, by seat locations, co-travel time, seat re-used? See new @WorldPopProject paper on CID: https://academic.oup.com/cid/article/doi/10.1093/cid/ciaa1057/5877944#.XyKPb4Escpw.twitter
1/n Some further interpretations & comments: We have carefully checked our data/results again and it looks all good.
2/n Regarding the small attack rate in Group 2, only 33 cases might cause a high uncertainty of estimates, given travellers adjacent to index patient had only an average attack rate of 3.5% (95%CI 2.9% – 4.3%).
3/n Among the 1355 close contacts in the Group 2, most of them had travel times less than 1 hour (46%) and 2 hrs (77%), and only 153 close contacts (11%) travelled more than 3 hrs, with only 3% more than 5 hrs.
4/n The quick decreasing close contact number over time and small number of index patients in the Group 2 might explain the low attack rate, especially after 4 hours. Figure S4 shows the distribution of close contact number between the two groups.
5/n As most index patients and close contacts (97.3%) travelled before 25 January when interventions had not been widely implemented across China, the attack rate estimated is more likely to reflect the setting of no/limited face covering and temperature screening at stations.
6/n Due to the data limitations, we can’t identify social relationships of passengers in Group 1. However, if most of co-travellers in Group 1 were family members, their attach risk on train might not significantly increase over time in travel.
7/n As they might spend more time togethoer at home, there might be a higher risk of transmission at household compared to train. So, the attack rate in Group 1 could reflects both transmission risk among travellers of family members and other train passengers.
