Contact person
- andrew.wills@uia.no
Re-thinking the data landscape
Standard approaches for evidence-generation in public health nutrition such as surveys and randomised controlled trials often rely on primary data collection(s). These can be slow and expensive and often suffer from low participation with under-representation of groups most vulnerable to poorer dietary health.
Another issue is the logical misalignment between the complexity of many public health challenges and the conventional methodological tools used to address it. Dietary health emerges from a non-linear and constantly-evolving complex system that includes cultural, social, economic and environmental drivers – these shape individual dietary behaviour. By contrast, the conventional methods used to understand and intervene are slow-moving, non-iterative and linear, dealing with one or a small number of factors at a time. This has contributed to an evidence base where equivocal findings on narrow questions are the norm, findings fail to translate or scale, and little progress is made on pressing dietary-related challenges such as childhood obesity.
One opportunity for progress in this field is work that contributes to new data sources and a richer data infrastructure. Such data should operationalise and describe food system elements, be timely, accessible, repeatable and updatable (ie; with some degree of automation) and have enough scientific rigor to provide analytical insights that facilitate decision-making.
Reusing and analysing existing information, so-called secondary use, offers a promising way forward. In this regard, few countries match Norway’s administrative and health registry data, wealth of cohort studies and capacity to link individual population-level data. However, we can also think creatively and exploit advances in data science and artificial intelligence to identify other sources of data. Such data for example, may come from digital footprints or smart devices (eg, grocery transaction data and footfall), be scraped from the internet (food pricing, geospatial data on the food environment), or be mined from publicly available documentation (eg; local government planning, funding and policy making).
The EATDOC project
EATDOC is an initiative motivated by these principles.
In a nutshell, it aims to test and explore whether useful information about dietary health related interventions, actions and activities can be mined from publicly available documentation.
Our first project seeks to map direct food provision across Norway in early childhood education and care (ECEC) settings – namely kindergartens, primary schools and afterschool clubs, using publicly available documentation from municipalities.
Instead of the go to standard approach of survey and primary data collection we will implement data-science and artificial intelligence (AI) based methods to extract information about direct food provision from municipality documentation, and process and structure it into useable datasets.
While this is a proof of principle project, we also hope it will have substantive utility for public health nutrition. Expected outcomes include a visualisation of the scale and diversity of food provision in ECEC settings across regions and over time; linkable dataset(s) for knowledge sharing between municipalities, identification of data gaps, and code and knowledge for replication and application to other settings. Lastly, where variation exists between municipalities or ECEC settings, there is potential to design powerful quasi-experimental studies to generate evidence about the effects and efficacy of dietary-health related interventions and activities.
EATDOC extensions
The project can be extended in several ways. For example, identifying other dietary health related events and interventions in these settings such as standards/ restrictions governing what foods and beverages are allowed on-site and budgeting decisions; expanding the information sources to include websites or information from other sources such as food suppliers; implementing the framework to other settings such as secondary schools or the local food environment.
Privacy Notice
These documents are openly accessible to the public and contain information related to municipal policies, decisions, meeting agendas, reports, and other administrative records. Although some documents may contain unredacted personal information, EATDOC will not process this or link it with other data or information. EATDOC is focused on identifying, extracting and analysing information related to public health nutrition, food environments, school meals and other nutrition related interventions, actions and policy. It does not seek to profile, evaluate or draw any conclusions about any individual. Personal details such as names of committee members are not used for analysis and are not included in the research outputs.
Project group
- andrew.wills@uia.no
- anine.medin@uia.no
- Phone
- +47 38 14 14 28
- syed.s.ali@uia.no
- Phone
- +47 38 14 12 14
External members
