To the Editor: Hsiao and colleagues (Dec. 14 issue)¹ report the results of a study of influenza vaccine implementation. We are concerned that intracluster correlations were ignored at the hospital, week, household, and within-patient (repeated vaccination) levels. Supplementary analyses that were stratified according to the first two potential correlations ignore interactions (which would occur, e.g., if a nursing home organizes group visits every other week) and ignore other potential correlations. Mixed-effects models may provide more accurate assessments.² To the Editor: In their article about recombinant or standard-dose inactivated influenza vaccines in adults under 65 years of age, Hsiao et al. . . .

致编辑： Hsiao 及其同事（12 月 14 日）¹报告了一项流感疫苗实施研究的结果。我们担心医院、周、家庭和患者内部（重复疫苗接种）级别的簇内相关性被忽视。根据前两个潜在相关性分层的补充分析忽略了相互作用（例如，如果疗养院每隔一周组织一次小组访问，就会发生这种情况）并忽略其他潜在相关性。混合效应模型可以提供更准确的评估。² 致编辑： Hsiao 等人在关于 65 岁以下成人使用重组或标准剂量灭活流感疫苗的文章中。 。 。 。