Norrby S R
Scand J Infect Dis Suppl. 1984;42:169-76.
Documentation of clinical and microbiological efficacy and safety of antibiotics require well designed studies involving large number of patients. In principal, all studies should be comparative and the comparative agent(s) should be well documented in the literature. If possible, the studies should also be blind but if satisfactory blinding cannot be guaranteed, open studies will have to be performed, especially if injectible antibiotics are studied. The size of the trial, in terms of number of patients evaluable for clinical and especially for microbiological efficacy, must be sufficient either to prove significant differences between the regimens studied or to demonstrate equality with a reasonable type II error. Since the end-points used for efficacy and safety are not normally continuous, the patient materials must often be of a size which can only be achieved by multiple independent trials or multicentre trials. Thus, such studies are recommended.
抗生素临床和微生物学疗效及安全性的记录需要设计良好、涉及大量患者的研究。原则上，所有研究都应具有可比性，且对照药物应在文献中有充分记载。如有可能，研究还应采用盲法，但如果无法保证令人满意的盲法，则必须进行开放研究，尤其是在研究注射用抗生素时。就可评估临床疗效尤其是微生物学疗效的患者数量而言，试验规模必须足够大，以证明所研究方案之间存在显著差异，或在合理的II类错误情况下证明等效性。由于用于疗效和安全性的终点通常不是连续性的，患者样本量往往需要达到只有通过多个独立试验或多中心试验才能实现的规模。因此，推荐此类研究。
