Guide to using EARS Algorithm

Start YearInsert the data start date
Start MonthInsert the start month of the data
Start DayInsert the start month of the data
Range of Data(min)Insert the beginning of range under review
Range of Data(max)Insert the end of the range under review

String indicating which method to use:

“C1” for EARS C1-MILD method (Default),

“C2” for EARS C2-MEDIUM method,

“C3” for EARS C3-HIGH method

BaselineHow many time points to use for calculating the baseline,
Min Sigma

By default 0. If minSigma is higher than 0, for C1 and C2, the quantity zAlpha

* minSigma is then the alerting threshold if the baseline is zero.
Howard Burkom suggests using a value of 0.5


An approximate (two-sided) (1 − α) prediction interval is calculated

Choose xlsx FileIn this section, upload the data file. The data format should be according to the data structure.

Early Aberration Reporting System

The Early Aberration Reporting System (EARS) is designed to provide advanced surveillance for short periods around events such as the Olympic Games, for which there is generally little or no prior information. EARS was first introduced by the centers for disease control and prevention, and since September 11, 2001, has been used as a standard surveillance system in many US local health departments; EARS also applies to New Zealand notifiable disease surveillance data and is updated weekly. The primary purpose of EARS is to provide multiple aberration detection methodsto national, state, and local health departments and allow users to change sensitivity and specificitythresholds to values considered important to public health by state and local health departments after selecting valid aberration detection methods. In this algorithm, three early detection methods named C1-MILD, C2-MEDIUM and C3-HIGH have been implemented. The terms mild, medium and high refer to the level of sensitivity of the three statistical methods. C1-MILD and C2- MEDIUM are actually types of Shewhart charts that use the moving average and sample standard deviation to standardize each observation, and C3 is the result of combining information based on C2. Threshold values in all three methods, C1-MILD, C2-MEDIUM and C3-HIGH, are obtainedusing the one-way CUSUM method.


The base period for C1-MILD is determined based on data from the previous 7 days, from day (t– 7) to day (t – 1). In this method, if a deviation alert is observed on a specific day (t), it is less likely to have a subsequent deviation alert on the following day (t + 1) because the increased values from the previous day are immediately included in the new base period. The C1-MILD method is highly beneficial for situations where detected alarms can be monitored and controlled on a daily basis and quickly (within 24 hours). Therefore, expecting a warning notification based on the same information for the next day would be incorrect. The threshold level calculation in the C1-MILD method is based on an equation related to the CUSUM threshold. It considers warning notifications for values that are more than 3 standard deviations higher than the average of the base values. If greater sensitivity is required, such as when state or local health facilities are alerted, the threshold can be adjusted for values that are more than 2 standard deviations above the baseline mean.


The second approach, C2-MEDIUM, utilizes data from a 7-day base period, specifically from day (t-9) to day (t-3). For instance, if the current day is the tenth day of active monitoring, the base data used for comparison would consist of the collected data from the first to the seventh day of the monitoring period. The difference in the base period of C2-MEDIUM compared to C1-MILD, with a two-day lag, increases the likelihood of C2-MEDIUM flagging consecutive high values. This is because these values are not immediately included in the new base period after the initial alert occurrence, resulting in a difference in the base comparison period. This difference allows C2-MEDIUM to provide better performance in identifying the duration of a rapid increase in cases, such as during the peak phase of an epidemic, by issuing timely alert notifications. The threshold calculations for C2-MEDIUM are similar to C1 MILD, with the threshold value based on CUSUM and values exceeding 3 standard deviations. Adjustments to the threshold are made for more sensitive surveillance objectives.


The third approach, C3-ULTRA, theoretically exhibits the highest sensitivity among the three proposed methods and is designed to detect gradual deviations over short time periods. C3- ULTRA utilizes the same base period as C2-MEDIUM, which spans from day (t-3) to day (t-9). However, the threshold is based on the three-day average run length of the CUSUM.


Hutwagner, L. C., Thompson, W. W., Seeman, G. M., & Treadwell, T. (2005). A simulation model for assessing aberration detection methods used in public health surveillance for systems with limited baselines. Statistics in medicine, 24(4), 543-550.


– Designing a set of evaluation tools for outbreak detection algorithms in the timely discovery of single-source and progressive epidemics. doctorate thesis Hamedan University of Medical Sciences.