Background: Health eTools for Schools was developed to assist school nurses with routine entries, including height and weight, on student health records, thus providing a readily accessible data base. Data-mining techniques were applied to this database to determine if clinically signi?cant results could be generated. Methods: Body mass index (BMI) data collected and entered in eTools by school nurses from 657,068 students attending 1156 schools in 49 of 67 Pennsylvania counties during 2005-2009 were analyzed. Students in each BMI category were sorted; regression was used to model mean and percentage trends. A chi-square test of individually matched BMI percentages was computed and migration across normal, overweight, and obese states determined. Results: The highest percentage of obese students occurred in middle school. The mean trends for obesity and overweight had increasing slopes of 0.189 and 0.227, respectively; with regression slope for overweight greater than 59%. Within groups, substantial percentages of individually matched BMIs changed signi?cantly (p less than 0.0001) over 2 years, migrating between normal weight, overweight, and obese. A comparison of 2009 measured BMI for grades 9-12 from eTools with 2009 Pennsylvania Youth Risk Behavior Survey self-reported BMI yielded substantial differences. Conclusion: A pattern of increasing BMI for elementary students with a corresponding decrease among middle and high school students emerged. The means trends for both overweight and obesity were greater in 2009 than in 2005, increasing steadily to 2008 and slightly declining to 2009. The dominant overall pattern ?ows from overweight to obese. If continued unabated, percentage of students who are obese will dominate over time. (Contains 4 figures and 2 tables.)