Back

An investigation of the impact of community controls on commonly reported epidemiological estimates in tuberculosis household contact study

Shi, H.; Li, T.

2023-01-18 infectious diseases
10.1101/2023.01.18.23284705 medRxiv
Show abstract

BackgroundTuberculosis (TB) has long been a major public health problem worldwide. In particular, during the period of the raging covid-19 epidemic, the situation of tuberculosis prevention and control has been critical. However, current TB household contact study describes the general risk of TB in target population and are unable to characterize the individual risk following exposure to active TB cases. MethodWe designed a dynamic simulation program for TB transmission to generate simulated datasets based on historical data on TB infection in four regions of Brazil, conducted a household contact study of household contacts with active TB (n=1711), and added matched (n=1362) and unmatched (n=1276) community control households to generate different datasets, respectively. We estimated the Second attack rate (SAR), Odds ratio, relative risk for each dataset. ResultEnrolling community controls extends the classical model of infectious disease SAR to TB in a household contact setting. Allowing us to separate the risk of household exposure from the risk of TB infection in the community, thus obtaining separate estimates of SAR and risk of contact with active TB cases. But over-matching for community control can lead to a reduction in the amount of data and even mask certain risk factors.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.