TL;DR
This study found that planned school breaks, like winter and spring breaks, significantly reduce student absenteeism due to influenza-like illness (a-ILI) by nearly 50%. This suggests that even relatively short breaks from in-person schooling can be an effective strategy for curbing influenza transmission in schools and communities.
What Problem Does This Paper Address?
Influenza transmission is a significant public health concern, and school-aged children play a crucial role in its spread within communities due to their extensive social networks and prolonged viral shedding. While reactive school closures during outbreaks have been studied, their effectiveness is debated. This paper focuses on a less-explored area: the impact of planned school breaks (holidays) on influenza transmission. The authors aim to quantify how these breaks, which occur during influenza seasons, affect student absenteeism due to influenza-like illness (a-ILI) and whether they can serve as a proactive public health intervention.
Methods
The researchers analyzed student absenteeism data from the Oregon School District (OSD) in Wisconsin over five academic years (September 2014 to June 2019). They focused on a-ILI, defined as missing school due to fever and at least one respiratory symptom.
The core methodology involved comparing a-ILI counts in the two-week periods before and after scheduled winter and spring breaks. To control for potential confounding factors like general time trends, they introduced a "pseudo-break" of nine days between winter and spring breaks each year, during which school was still in session. This pseudo-break served as a control group to isolate the effect of actual school closures.
Statistical analyses included:
- Cochran-Mantel-Haenszel (CMH) test: Used for crude association analysis, comparing a-ILI odds before and after breaks, stratified by school year.
- Generalized Linear Regression Models (GLM): Employed to assess the relationship between a-ILI counts and break periods, while accounting for community-level influenza risk (measured by weekly medically attended influenza counts) and temporal trends (linear and quadratic effects of time).
Datasets
- Oregon School District (OSD) Student Absenteeism Data: Daily counts of student absences, categorized as a-ILI, collected via an automated telephone system and recorded in the Infinite Campus® student information system. This data spanned five academic years (2014-2015 to 2018-2019).
- Wisconsin Influenza Incidence Surveillance Project (W-IISP) Data: Weekly counts of laboratory-confirmed medically attended influenza (MAI) from primary care clinics in and around the OSD. This served as a proxy for community influenza risk.
No new benchmarks were introduced; the study utilized existing surveillance data and established statistical methods.
Key Insights and Ideas
- Planned breaks are not just downtime; they're public health tools: The study provides strong evidence that scheduled school breaks can actively reduce influenza transmission, not just passively reflect it.
- Timing is everything: The impact of breaks on a-ILI was more pronounced when the breaks coincided with periods of higher local influenza activity, highlighting the importance of strategic timing for interventions.
- Pseudo-breaks validate real breaks: The lack of significant findings for the pseudo-breaks (when school was in session) strengthens the argument that the observed reductions during actual breaks are due to the cessation of in-person schooling.
- Short breaks can have a significant impact: The study suggests that even breaks of 9-16 days can lead to a substantial, albeit short-term, reduction in a-ILI.
Key Results
- Nearly 50% Reduction in a-ILI: Following planned winter and spring breaks, there was a nearly 50% reduction in a-ILI counts in the two weeks after the break compared to the two weeks before.
- Winter breaks: Odds ratio of 0.679 (95% CI: 0.600-0.769, p < 0.001) for a-ILI after vs. before break.
- Spring breaks: Odds ratio of 0.327 (95% CI: 0.283-0.378, p < 0.001) for a-ILI after vs. before break.
- Adjusted Models Confirm Impact: Regression models, after accounting for community influenza risk and temporal trends, showed a proportional change (PC) of 0.483 for winter breaks and 0.488 for spring breaks, indicating a substantial decrease in a-ILI after breaks.
- Pseudo-breaks Show No Significant Effect: The analysis of pseudo-breaks revealed no statistically significant difference in a-ILI before and after these periods, supporting the idea that actual school closures were the driving factor.
- Younger Grades More Affected: Figure 3 indicates higher levels of a-ILI in younger grades (kindergarten through elementary school), suggesting these age groups might be more susceptible or contribute more to transmission.
Limitations and Future Work
- Observational Data: The study is observational, meaning it suggests associations but cannot definitively prove causation. Unmeasured temporal confounders could still play a role.
- Self-Reported Data: a-ILI relies on parent/guardian reporting, which may lead to underestimation due to underreporting.
- Proxy for Influenza: a-ILI is used as a proxy for influenza, and while validated, it doesn't capture the nuances of different influenza types/subtypes.
- Generalizability: Findings from a single school district may not be generalizable to all populations or locations, especially given variations in influenza seasons and academic calendars.
- Community Risk Measurement: While community MAI data was used, it may not fully capture the complex dynamics of community-level risk.
Future Work: The authors suggest further investigation into the impact of well-timed shorter breaks, both planned and unplanned, to determine an optimal duration for brief school closures that could effectively suppress community transmission of influenza. They also note the potential for similar interventions (like transitions to distance learning) to have comparable effects.
Summary updated