Evaluation of influenza control measures frequently focuses on the efficacy of chemoprophylaxis and vaccination, while the effectiveness of non-pharmaceutical interventions (NPI) receives less emphasis. While influenza control measures are frequently reported for individual outbreaks, there have been few efforts to characterize the real-world effectiveness of these interventions across outbreaks.
To characterize influenza case and outbreak definitions and control measures reported by long-term care facilities (LTCFs) of elderly adults and estimate the reduction in influenza-like illness (ILI) attack rates due to chemoprophylaxis and NPI.
We conducted a literature search in PubMed including English-language studies reporting influenza outbreaks among elderly individuals in LTCFs. A Bayesian hierarchical logistic regression model estimated the effects of control measures on ILI attack rates.
Of 654 articles identified in the literature review, 37 articles describing 60 influenza outbreaks met the inclusion criteria. Individuals in facilities where chemoprophylaxis was used were significantly less likely to develop influenza A or B than those in facilities with no interventions [odds ratio (OR) 0·48, 95% CI: 0·28, 0·84]. Considered by drug class, adamantanes significantly reduced infection risk (OR 0·22, 95% CI: 0·12, 0·42), while neuraminidase inhibitors did not show a significant effect. Although NPI showed no significant effect, the results suggest that personal protective equipment may produce modest protective effects.
Our results indicate pharmaceutical control measures have the clearest reported protective effect in LTCFs. Non-pharmaceutical approaches may be useful; however, most data were from observational studies and standardized reporting or well-conducted clinical trials of NPI are needed to more precisely measure these effects.