Table 2.
Odds ratios (95% CI) of the effect of influenza interventions on attack rates among outbreaks consisting of influenza A only (Models 1 and 2) and outbreaks of influenza A or B (Models 3 and 4) in long-term care facilities, considering antiviral interventions together (Models 1 and 3) and as independent interventions (Models 2 and 4)
| Influenza A outbreaks (n = 51) | Influenza A or B outbreaks (n = 60) | |||
|---|---|---|---|---|
| Model 1 | Model 2 | Model 3 | Model 4 | |
| No intervention | Reference* | Reference | Reference | Reference |
| Personal protective equipment | 0·75 (0·33, 1·61) | 0·53 (0·25, 1·10) | 0·99 (0·49, 1·93) | 0·63 (0·33, 1·19) |
| Social distancing | 1·05 (0·53, 2·16) | 1·35 (0·72, 2·62) | 1·07 (0·58, 1·90) | 1·31 (0·78, 2·18) |
| Any antiviral drug | 0·52 (0.29, 0·93) | – | 0·48 (0·28, 0·84) | – |
| Adamantanes | – | 0·33 (0·17, 0·62) | – | 0·27 (0·14, 0·48) |
| Neuraminidase inhibitors | – | 1·55 (0·62, 3·98) | – | 1·27 (0·56, 2·76) |
| Adamantanes and Neuraminidase inhibitors | – | 0·64 (0·35, 1·28) | – | 0·57 (0·31, 1·03) |
| Influenza B | – | – | 0·55 (0·27, 1·15) | 0·39 (0·20, 0·82) |
CI, confidence interval; n, number of outbreaks.
Point estimates and 95% CIs were derived from hierarchical binomial models and are interpreted as the difference in the odds of becoming an influenza case in a facility with a particular policy versus a facility where no intervention was implemented.
Bolded estimates indicate statistically significant effects (P < 0·05).
Reference indicates that outbreaks using ‘No intervention’ served as the comparison group for calculating odds ratios.