Study on early warning for tuberculosis outbreaks in school using moving average method

Objective To establish an early warning model of tuberculosis (TB) for the early detection of epidemic of TB in schools in Shaoxing. Methods The semimonthly reported TB cases in schools in Shaoxing from 2010 to 2015 were collected from Tuberculosis Information Management System. The semimonthly serial number of TB cases in schools from 2010 to 2015 (1-144) were used as independent variable and the semimonthly reported cases of TB were used as dependent variable, then the regression equation was established. According to this regression equation, the semimonthly simulated cases, residual, average moving residual, standard deviation of residual were calculated. According to the characteristics of infectious diseases, x +1.0s, x +1.6s, x +2.0s of warning values were used as the alert thresholds respectively for grade Ⅲ TB cluster (case number 2 within 3 months), gradeⅡTB cluster (case number 6 within 3 months) and TB outbreak(case number 10 within a semester). Results The established regression equation was y=-0.014x+6.34. The average moving residual of semimonthly simulated cases from 2010 to 2015 were 3.26, 2.11, 2.13, 2.31, 3.88, 4.22, 4.02, 3.84, 3.90, 3.03, 2.34, 2.21, 2.07, 2.19, 2.08, 2.21, 2.32, 2.43, 2.63, 2.41, 3.11, 3.19, 3.46 and 3.23 respectively. Fourteen TB cluster Ⅲ grade, 7 TB cluster Ⅱ grade and 3TB outbreak occurred in school in shaoxing from 2010 to 2015. The sensitivity and specificity of early warning model for TB outbreak,TB cluster Ⅱgrade and TB cluster Ⅲ grade were 100%, 98.61%; 100%, 99.30%; 71.43%, 99.30% respectively. Conclusion The early warning models established by use moving average method is helpful for the early detection of TB epidemic in schools.