Semisupervised score based matching algorithm to evaluate the effect of public health interventions (preprint)

Multivariate matching algorithms "pair" similar study units in an observational study to remove potential bias and confounding effects caused by the absence of randomizations. In one-to-one multivariate matching algorithms, a large number of "pairs" to be matched could mean both the information from a large sample and a large number of tasks, and therefore, to best match the pairs, such a matching algorithm with efficiency and comparatively limited auxiliary matching knowledge provided through a "training" set of paired units by domain experts, is practically intriguing. We proposed a novel one-to-one matching algorithm based on a quadratic score function $S_{\beta}(x_i,x_j)= \beta^T (x_i-x_j)(x_i-x_j)^T \beta$. The weights $\beta$, which can be interpreted as a variable importance measure, are designed to minimize the score difference between paired training units while maximizing the score difference between unpaired training units. Further, in the typical but intricate case where the training set is much smaller than the unpaired set, we propose a \underline{s}emisupervised \underline{c}ompanion \underline{o}ne-\underline{t}o-\underline{o}ne \underline{m}atching \underline{a}lgorithm (SCOTOMA) that makes the best use of the unpaired units. The proposed weight estimator is proved to be consistent when the truth matching criterion is indeed the quadratic score function. When the model assumptions are violated, we demonstrate that the proposed algorithm still outperforms some popular competing matching algorithms through a series of simulations. We applied the proposed algorithm to a real-world study to investigate the effect of in-person schooling on community Covid-19 transmission rate for policy making purpose.

Dynamic Estimation of Epidemiological Parameters of COVID-19 Outbreak and Effects of Interventions on Its Spread (preprint)

A key challenge for estimating the epidemiological parameters of the COVID-19 out-break in Wuhan is the discrepancy between the officially reported number of infections and the true number of infections. A common approach to tackling the challenge is to use the number of infections exported from Wuhan to infer the true number in the city. This approach can only provide a static estimate of the epidemiological parameters before Wuhan lockdown on January 23, 2020, because there are almost no exported cases thereafter. Here, we propose a method to dynamically estimate the epidemiological parameters of the COVID-19 outbreak in Wuhan by recovering true numbers of infections from day-to-day official numbers. Using the method, we provide a comprehensive retrospection on how the disease had progressed in Wuhan from January 19 to March 5, 2020. Particularly, we estimate that the outbreak sizes by January 23 and March 5 were 11,239 [95% CI 4,794-22,372] and 124,506 [95% CI 69,526-265,113], respectively. The effective reproduction number attained its maximum on January 24 (3.42 [95% CI 3.34-3.50]) and became less than 1 from February 7 (0.76 [95% CI 0.65-0.92]). We also estimate the effects of two major government interventions on the spread of COVID-19 in Wuhan. In particular, transportation suspension and large scale hospitalization respectively prevented 33,719 and 90,072 people from getting infected in the nine-day time period right after its implementation.