National Child Measurement Programme - effect of providing free lunches to infant school children.

Text removed is struck through; text added is underlined. Unchanged paragraphs are summarised rather than repeated.

Objective for processing

[1 paragraph unchanged] The objective of this study is for the data controller and processor (The University of Essex) to investigate the effect of providing free lunches to infant school children on weight (underweight/normal weight/overweight/obese) in reception class. The National Child Measurement Programme (NCMP) data is essential to the research question as it is the only source of national weight and Body Mass Index (BMI) data. Within the research program there will be separate analysis of the effect of free lunches on cognitive outcomes, but this is a completely independent research strand using administrative education records. There will be no linkage of data from that study with the NCMP data. No data will be processed outside England and Wales. All data is aggregated containing small numbers. The University of Essex is the sole data controller, who also processes the data for the purposes described in this Agreement. This project has been based on external funding from the Nuffield Foundation. This extension, however, is not covered by the grant and will be funded by The Research Centre on Micro-Social Change (ESRC-funded). Neither have access to any unsuppressed data and will not be able to influence or suppress the findings of the survey. The objective of this study is for The University of Essex to investigate the effect of providing free lunches to infant school children on weight (underweight/normal weight/overweight/obese) in reception class. The National Child Measurement Programme (NCMP) data is essential to the research question as it is the only source of national weight and Body Mass Index (BMI) data. Within the research program there will be separate analysis of the effect of free lunches on cognitive outcomes, but this is a completely independent research strand using administrative education records. There will be no linkage of data from that study with the NCMP data. No data will be processed outside England and Wales. All data is aggregated containing small number and will use pseudonymised school IDs. No suppression of small numbers will take place as this will suppress the obesity prevalence levels for around two-thirds of schools making the data not usable for this project. The research team have requested several years before and after the implementation of the free lunch program so that they can appropriately control for trends in weight. They will not match the data to any other data source. [3 paragraphs unchanged]

Processing activities

Under a previous agreement, NHS Digital transferred the NCMP data requested to the Institute for Social and Economic Research at the University of Essex by Secure Electronic File Transfer where it will be analysed by researchers at the University of Essex (“the research team”) only. All individuals with access to the data containing small numbers are substantive employees of the University of Essex. The data requested will be aggregated containing small numbers and will use pseudonymised school IDs. No suppression of small numbers will take place as this will suppress the obesity prevalence levels for around two-thirds of schools making the data not usable for this project. NHS Digital does not hold information on the proportion of children receiving Free School Meals (FSM). FSM take up rates, the deprivation quintile of school (using the Income Deprivation Affecting Children Score), and the take-up rates of school meals which are some of the contextual data items requested. The FSM value and take-up rates will also be grouped into quintiles (5 bands) to mitigate against re-identification. NHS Digital does not hold information on the proportion of children on Free School Meals (FSM). FSM take up rates and the deprivation quintile of school (using the Income Deprivation Affecting Children Score) and the take-up rates of school meals which are some of the contextual data items requested. The FSM value and take-up rates will also be grouped into quintiles (5 bands) to mitigate against re-identification. As such the University of Essex previously transferred a file of aggregated school level data containing this information to NHS digital which was merged with the NCMP data before it was passed back to the University via Secure Electronic File Transfer (SEFT). There will be no subsequent flows of data. The University of Essex transferred a file of aggragated school level data containing this information to NHS digital which was merged with the NCMP data before it is passed back to the University. The data will not be transferred to or shared with any other organisation. The data will be analysed using statistical methods such as regression analysis, before/after introduction comparisons, difference-in-difference statistical techniques. The research team is requesting a number of years before and after the implementation of the free lunch program so that they can appropriately control for trends in weight. They will not match the data to any other data source. The NCMP data requested will be analysed by the research team at the University of Essex. All individuals with access to the data containing small numbers are substantive employees of the University of Essex, all of whom have been sufficiently trained in data protection and confidentiality The data will not be transferred to or shared with any other organisation. The University of Essex will not match the data to any other data source. The data will be analysed using statistical methods such as regression analysis, before/after introduction comparisons, difference-in-difference statistical techniques. The research team have requested several years before and after the implementation of the free lunch program so that they can appropriately control for trends in weight. [8 paragraphs unchanged] This project has been based on external funding from the Nuffield Foundation. This extension, however, is not covered by the grant and will be funded by The Research Centre on Micro-Social Change (ESRC-funded). Neither have access to any unsuppressed data and will not be able to influence or suppress the findings of the survey.

Expected output

The main outputs of the project will be a project an end-of-project report and a research brief. It is expected that The University of Essex will submit two academic papers will be submitted to leading international economics journals, which will also be present at national and international academic conferences as well as non-academic outputs (see yielded benefits). As a paper has now been drafted and presented, it is expected that academic reviewers will ask for additional checks and analysis, making access to the data for a longer time-period necessary. The results of the project are of interest to a variety of [26 words unchanged] interested in child obesity, the media and the general public. The findings have been, and will be be, disseminated widely through dedicated web pages, Institute for Social and Economic Research (ISER) newsletter, a dissemination event aimed at policy makers, practitioners and academics, academics in November 2020, press release and subsequent media engagement to ensure the policy audience is reached. The results included in these outputs will be descriptive statistics such as means and standard deviations, and regression coefficients from statistical models; small numbers will never be included. Outputs will be aggregated with small numbers suppressed. All outputs will comply with the ICO Guide for Anonymisation. All outputs will be cleared with NHS Digital before publication. Presentation of results took place from the second half of 2018 and throughout first half of 2019, with publication of the a report, research brief and dissemination event for academic and non-academic audiences (including press release) completed. Submission of the academic paper to an economics journal is also planned for late 2019/ beginning of 2020. It is expected that academic reviewers will ask for additional checks and analysis, making access to the data for a longer time-period necessary. All publications based on this dataset will be shared with Public Health England via NHS Digital before they are published. PHE will have no access to unsuppressed data and will not be able to influence or suppress the findings of the study. A paper was drafted and presented the results at the Royal Economic Society conference in April 2019, the European Association of Labour Economists conference in September 2019 and at a seminar at NTNU, Norway. A draft report was published in early 2020, and the results were covered in a piece in the Observer. The results were also discussed at an advisory group meeting in 2020. Based on the feedback received the study is currently performing final checks before publication of final report and submission of academic papers. The University of Essex spoke about findings at a panel event on the future of universal free school meals at the Labour Party conference fringe in September 2019. The researchers had a meeting with the Cabinet Office, DfE and DHSC to discuss the results. The results have fed into the discussions around the Summer Food Fund and a piece will be published in the CPAG magazine Poverty in October. An article in Poverty, the Child Poverty Action Group magazine will be published in October 2020. There will be another end-of-project report which includes results based on education and dissemination event in November 2020. Submission of the first academic paper to an economics journal was planned for October 2020 and the second shortly thereafter. It is expected that academic reviewers will ask for additional checks and analysis, making access to the data for a longer time-period necessary. All publications based on this dataset have been and will be shared with Public Health England via NHS Digital before they are published. All outputs have been and will be cleared with NHS Digital before publication. Outputs will be aggregated with small numbers suppressed in line with HES Analysis Guidance. All outputs will comply with the ICO Guide for Anonymisation. PHE will have no access to unsuppressed data and will not be able to influence or suppress the findings of the study.

Expected measurable benefits

The main benefit to health and social care is to assess if the result that the Universal Infants Free School Meals program leads to fewer children being [11 words unchanged] are more likely to become obese adults and develop associated health conditions. If there is no The reduction in the number of children being overweight or obese then it calls into question whether can indicate (under some assumptions) that the Universal Infants Free Schools Meals program is an efficient use of public funds. Over a fifth In 2015/16 more than 20% of Reception children in reception year and 33% of Year 6 children were overweight or obese in 2015/16 and this figure was a third in year 6. obese. Overweight and obese children are more likely to be ill, be absent [95 words unchanged] estimated to be £6.3 billion in 2015 (see Public Health England, https://www.noo.org.uk/LA/impact/economic). [1 paragraph unchanged]

Benefits reported

A paper was drafted and presented the results at the Royal Economic Society conference in April 2019, the European Association of Labour Economists conference in September 2019 and at a seminar at NTNU, Norway. Based on the feedback received the study is currently performing further checks and refining the analysis more to get a better picture of the mechanisms that drive the beneficial results found. The results of this study have and will feed into the policy around free meals at school. This is how it is hoped the research will benefit health of children. The findings are summarised as follows: Spoke about findings at a panel event on the future of universal free school meals at the Labour Party conference fringe in September 2019. A parallel event scheduled for the Conservative conference was cancelled as no parliamentarians were prepared to attend. A meeting has been scheduled with the Cabinet Office, DfE and DHSC to discuss the results. Since September 2014 school lunches – previously free for very low income students - are available to all children in England in their first three years in school free of charge. Researchers draw on data from the National Child Measurement Program (NCMP) to evaluate the effect of switching from targeted to universal free lunch provision on the body weight outcomes of children aged 4-5, showing how the treatment effect evolves over the school year as the cumulative dosage of exposure to free meals increases. By the end of the school year, on average a child exposed to free lunches is 1.2 percentage point more likely to be of ‘healthy weight’ and 0.7 percentage points less likely to be obese, and has body mass index (BMI) that is 4.3% of a standard deviation lower than one who is not. These findings serve as evidence that can be used to influence policy. Ultimately, through this research, the hope is that the results will feed into the policy around free meals at school. This is how it is hoped the research will benefit health of children. The findings to date are summarised in the draft abstract as follows: "The Impact of Universal Infant Free School Meals on Child body weight outcomes Since September 2014 school lunches – previously free for very low income students - are available to all children in England in their first three years in school free of charge. We draw on data from the National Child Measurement Program (NCMP) to evaluate the effect of switching from targeted to universal free lunch provision on the body weight outcomes of children aged 4-5, showing how the treatment effect evolves over the school year as the cumulative dosage of exposure to free meals increases. By the end of the school year, on average a child exposed to free lunches is 1 percentage point more likely to be of ‘healthy weight’ and 0.5 percentage points less likely to be obese, and has body mass index (BMI) that is 4.2% of a standard deviation lower than one who is not."