Methods for detecting seasonal influenza epidemics using a school absenteeism surveillance system

Abstract

Background: School absenteeism data have been collected daily by the public health unit in Wellington-Dufferin-Guelph, Ontario since 2008. To date, a threshold-based approach has been implemented to raise alerts for community-wide and within-school illness outbreaks. We investigate several statistical modelling approaches to using school absenteeism for influenza surveillance at the regional level, and compare their performances using two metrics.

Methods: Daily absenteeism percentages from elementary and secondary schools, and report dates for influenza cases, were obtained from Wellington-Dufferin-Guelph Public Health. Several absenteeism data aggregations were explored, including using the average across all schools or only using schools of one type. A 10% absence threshold, exponentially weighted moving average model, logistic regression with and without seasonality terms, day of week indicators, and random intercepts for school year, and generalized estimating equations were used as epidemic detection methods for seasonal influenza. In the regression models, absenteeism data with various lags were used as predictor variables, and missing values in the datasets used for parameter estimation were handled either by deletion or linear interpolation. The epidemic detection methods were compared using a false alarm rate (FAR) as well as a metric for alarm timeliness.

Results: All model-based epidemic detection methods were found to decrease the FAR when compared to the 10% absence threshold. Regression models outperformed the exponentially weighted moving average model and including seasonality terms and a random intercept for school year generally resulted in fewer false alarms. The best-performing model, a seasonal logistic regression model with random intercept for school year and a day of week indicator where parameters were estimated using absenteeism data that had missing values linearly interpolated, produced a FAR of 0.299, compared to the pre-existing threshold method which at best gave a FAR of 0.827.

Conclusions: School absenteeism can be a useful tool for alerting public health to upcoming influenza epidemics in Wellington-Dufferin-Guelph. Logistic regression with seasonality terms and a random intercept for school year was effective at maximizing true alarms while minimizing false alarms on historical data from this region.

Keywords: Absenteeism surveillance system; Disease modelling; Epidemic detection; Influenza; Seasonal logistic regression.

Fig. 1

Influenza and absenteeism in WDGPH. Average absenteeism and laboratory-confirmed influenza cases (“Flu Cases”) reported to WDGPH for the WDG region from January 2008 to June 2018

Fig. 2

Evaluation metrics time-line. Illustration of terms used in the definitions of the evaluation metrics, ADD and FAR

Fig. 3

Characteristics of best-performing models. Representation of the proportion of a) model types and b) absenteeism data aggregation types within the 50 lowest FAR-producing epidemic detection methods. See Table 2 for aggregation abbreviations

Fig. 4

Characteristics of regression models. Effects of different model factors and absenteeism data aggregations on FAR averaged over all regression model types with optimized parameters. Each row represents a different aggregation for absenteeism data and each column represents either a data handling method or whether an additional predictor aside from absenteeism was included in the model. The pairs of columns separated by spaces can be compared to view the effect on FAR across the different data aggregations, where a lighter shade indicates preferable (lower) FAR. The value within each cell is the mean FAR

Fig. 5

Alarms of the top-performing model. True and false alarms for the best performing model, faceted by school year: the seasonal logistic random intercept model using ES-allavg data with l=7, Θ=0.20, interpolated missing values in the training data. Averaged absenteeism is plotted as grey bars, with actual laboratory-confirmed influenza case counts overlaid as black bars, and the epidemic reference day is indicated by the dashed yellow line for each school year