1 School of Mathematics and Statistics, Guangxi Normal University, Guilin, China.
2 Academy of Mathematics and Systems Science, Chinese Academy of Sciences, Beijing, China.
Stat Methods Med Res. 2019 Jan;28(1):211-222. doi: 10.1177/0962280217719914. Epub 2017 Aug 10.
Pooled testing is useful to identify positive specimens for large-scale screening. Matrix pooling is one of the commonly used algorithms. In this work, we investigate the properties of matrix pooling and reveal that the efficiency of matrix pooling is related with the magnitude of overlapping among groups. Based on this property, we develop a new design to further improve the efficiency while taking into account of testing error. The efficiency, pooling sensitivity and specificity of this algorithm are explicitly derived and verified through plasmode simulation of detecting acute human immunodeficiency virus among patients who were suspected to have malaria in rural Ugandan. We show that the new design outperforms matrix pooling in efficiency while retain the pooling sensitivity and specificity.
混合检测对于大规模筛查中确定阳性样本非常有用。矩阵混合是常用的算法之一。在这项工作中，我们研究了矩阵混合的特性，并揭示了其效率与组间重叠程度有关。基于这一特性，我们设计了一种新的方案，在考虑检测误差的同时，进一步提高了效率。通过在乌干达农村地区疑似疟疾的患者中检测急性人类免疫缺陷病毒的质粒模拟，我们明确推导并验证了该算法的效率、混合灵敏度和特异性。结果表明，新设计在保持混合灵敏度和特异性的同时，提高了效率。
