BackgroundPoor monitoring and evaluation leads to a lack of evidence regarding what works in terms of improving school meal quality. Sweden has a long-standing policy of universal free school meals but updated legislation regarding nutritional quality came into effect in 2011. To coincide with this, we built a fully automated web-based self-audit tool for schools, called School Food Sweden, in order to i) establish a national database on overall meal quality, covering six domains, and ii) help schools and municipalities improve, based on the implementation strategy of audit and feedback. We have previously examined the pre- and post-legislation period. In this paper we describe changes in one domain – nutritional quality – in the years since the new legislation and examine if repeated use of the tool was associated with improvements.MethodsPrimary schools in Sweden self-selected to use the tool between March 2012 and July 2019 and audited a four-week lunch menu. Date of audit, number of audits per school (1-9), how often schools accessed their feedback (% of times) and school characteristics were noted. Factors associated with meeting nutritional criteria were examined using variance weighted least squares regression and logistic regression, using both repeated cross-sectional and longitudinal design.ResultsAlmost half of all ca 4800 primary schools in Sweden signed up to the tool. Nutritional quality was audited by 1500 schools at least once. Repeated cross-sectional analyses showed the proportion of all schools meeting the criteria increased significantly between 2012/13 and 2018/19 (from 11% to 34%). For all schools, each extra audit completed increased the odds of meeting nutritional criteria by 1.30 (CI 1.20-1.41), controlling for region and time since the introduction of legislation. In a longitudinal analysis (schools with repeat audits, n=774), both number of audits and the frequency of accessing previous feedback reports were predictors of meeting the nutritional criteria (OR 2.02, CI 1.23-3.31), even after adjusting for time since the introduction of legislation and number of days since the previous audit.ConclusionsBoth legislation and self-audit with automatic feedback appear effective in helping schools to improve school meal quality. Self-audit may be an essential complement to legislation, or a promising alternative in settings where regulation is not an option.