目的 探讨平行性检验在未涉及生理活性生物体的ELISA四参数模型和平行线模型数据分析中的应用。方法 在未进行平行性检验的2种ELISA，肠道病毒71型体外相对效力插值法检测的四参数模型和肠道病毒71型抗原含量检测的平行线模型中，选取不同水平供试品检测数据，对不进行平行性检验获得的结果、平行性检验后偏离平行不显著的结果、平行性检验后偏离平行显著的结果进行一致性比较。结果 不同的数据分析模式均未导致结果波动，变异系数为2.9%~17.9%；将四参数模型插值法与偏离平行不显著结果、插值法与偏离平行显著结果进行配对t检验分析，除插值法与偏离平行显著结果对比中2.0水平P<0.001（t=0.11）外，其他组差异均无统计学意义；将平行线模型原始计算方法结果分别与偏离平行不显著结果和偏离平行显著结果进行配对t检验分析，差异均无统计学意义。结论 无论平行性检验偏离平行显著与否，其计算结果均与原始计算结果一致，经过数据分析模型筛选、进行过完整的方法学验证的ELISA，能够选择性使用平行性检验。

Objective To explore the application of parallelism test in the data analysis of four-parameter model and parallel line model ELISAs for non-physiologically active organisms.Methods In 2 ELISA methods without parallelism test, the four-parameter model of enterovirus 71 (EV71) in vitro relative efficacy measured by interpolation method, and the parallel line model of EV71 antigen content detection, test data of different levels were selected. The results obtained without parallelism test, and results with non-significant or significant deviation from parallel after parallelism test, were compared.Results Different data analysis modes did not cause result fluctuation, with coefficients of variation at 2.9% to 17.9%. The paired t-test analysis was carried out between the interpolation results of four-parameter model and results with non-significant or significant deviation from parallel, and showed no statistically significant difference except the comparison with significant deviation parallel results at 2.0 level (t=0.11， P<0.001). There was no statistically significant difference between the original calculation results of parallel line model and results with non-significant or significant deviation from parallel.Conclusion  Since the calculation results are consistent with original calculation results whether or not deviation from parallel in parallelism test is significant, ELISA can selectively use parallelism test after data analysis model screening and complete methodology validation.