Tupã Fit: Brasil2030 - an intelligent system for multisectoral use in promoting the nutritional health of children
Short video:
Long video:
Team: Tupã Fit
Track: Healthcare - Prevent Childhood Obesity
Brasil2030: an intelligent system for multisectoral use in promoting the nutritional health of children
Members:
- Luís Roberto Almeida Gabriel Filho (gabriel.filho@unesp.br)
- Camila Pires Cremasco (camila.cremasco@unesp.br)
- Stephen Kunihiro (stephen@nkodontologia.com.br)
- Vitor Gomes Ribeiro (vitor@nkodontologia.com.br)
- Paulo Eduardo e Silva Barbosa (paulo.barbosa@nutes.uepb.edu.br)
- Adalcino Júnior Araújo Tavares (adalcino.junior@nutes.uepb.edu.br)
- Rodrigo Brito do Nascimento (rodrigo.brito@nutes.uepb.edu.br)
Prioritizing needs
The increase in childhood obesity is alarming the world, and there is already a consensus that we are facing a global syndemic complex and multidimensional causes. Parents, educators and caregivers, managers, and even market initiatives for this segment of the population do not have at their disposal an environment capable of connecting different dimensions of knowledge and actors in this process in order to stop the progress of obesity, especially in children under the age of six. An information management mechanism capable of processing complex data and producing syntheses for decision-making and guiding the work of multiple actors is needed. At the same time, such information management mechanism needs to feed back information systems to make more robust the evidence needed to guide the transformation of social practices and the different environments to guarantee the protection of children in favorable environments for their growth and development.
In Brazil, 31.8% of children aged 0 to 5 years are overweight, which has encouraged the prioritization of the topic on the agenda of public policies and initiatives of society in general, following the Ministry of Health guidelines. However, a major challenge has been to promote coordinated action within the public and private sectors, also connecting existing initiatives within society to favor greater organicity, visibility, and resolvability of actions in different scenarios. Actions are needed to effectively provide comprehensiveness and quality of education, excellence and cost-effectiveness in care service, sensitiveness of public places and markets to children's needs and to encourage the adoption of healthier lifestyles.
Encouraged by the Swedish initiative "SWELIFE: zero childhood obesity at school start by 2030" and taking into account the Brazilian context, a group of researchers from the Centro de Ciência e Tecnologia em Soberania e Segurança Alimentar e Nutricional (INTERSSAN) accepted the challenge of pointing out the needs to generate the transformation of society for the prevention of childhood obesity with the "Twin Cities" strategy as its horizon. For this challenge, it was necessary to mobilize a Brazilian city to develop a technology capable of promoting the engagement of families and different sectors and actors of society in a network action for the protection of children and the promotion of favorable environments for their development and well-being, involving the family nucleus.
In this sense, it developed a panel of experts to generate the challenge to developers who are configured as follows: Given the assumptions of the twin cities and bearing in mind: (1) the favorable and formal acceptance of the proposal by the public management of the city of Tupã, in the State of São Paulo; (2) access to schools, community services, and other living environments for children and their families; (3) access to data from information systems from education, health and social assistance for children under the age of six in the municipality; (4) access to national databases that contribute to socioeconomic, health, education, availability and access to food indicators, disaggregated at the municipality level; (5) access to databases of interest produced within the municipality; (6) Brazilian initiatives related to public policies that relate to the well-being of the population; (7) civil society initiatives that converge towards citizen participation in the mobilization of social agents in educational processes and policy monitoring; the following challenge arises: what is the technological solution capable of creating a management model that can integrate the databases of public policies implemented in the territory, providing services and information about schools, family environments and social networks for the protection and promotion of health and well-being of the population?
Such a technological solution shall ensure the health of children and sustainable and inclusive food systems, with the main reference element being the 0 to 6-year-old child in their territory and the effective engagement of families and communities.
As a primary hypothesis, it is proposed that it is possible to build a digital tool that uses artificial intelligence to interpret and connect data and people, providing synthesis and application of knowledge and broad engagement of those directly or indirectly involved. As a secondary outcome, the existing data integrated with new data generated of different types will provide new hypotheses, allowing the purposeful assessment of public policies and other actions in different society sectors. Besides, it supports the development of knowledge related to big data management using machine learning to search for practical solutions to the problems identified.
Choosing a test scenario: the city of Tupã
The city of Tupã was chosen for convenience due to the link between the researchers and the local management and the favorable reception of the mayor. Tupã is a city in the São Paulo countryside, with an estimated population of 65,570 inhabitants, of whom 4,700 are children under six years old. Among 526 children aged 0 to 5 years (data available in the health service's information system), 25.38% are overweight, of which 4.56% are classified as obese. It is a city of tourist interest due to its cultural heritage of indigenous traditions and European immigration. The basis of the economy is agriculture and livestock. In general, its social indicators are located in the median or the third quartile of the Brazilian population.
Creating a network of actors to develop the technological solution
The mentors: researchers linked to the INTERSSAN linked to the Special Interest Group (SIG) of Obesity and the Rede Nacional de Pesquisa (RNP-Brazil), stimulated by Sweden's partners. The group is composed of Nutritionists, Doctors, Physical Educators, Geographers, Biologists, Advertisers.
The invited specialists: specialists from Brazil, invited to participate in a panel of experts to define the problem and monitor the development of the solution, in which Nutritionists, Pediatric Doctors, and Health Managers participate.
The developers: invited by RNP and INTERSSAN to meet the needs of the technological solution. This group is made up of researchers from São Paulo State University/UNESP (Luís Roberto Almeida Gabriel Filho, Camila Pires Cremasco), researchers from the company NK Health Information System (Stephen Kunihiro, Vitor Gomes Ribeiro) and developers from the Strategic Health Technology Center of Paraíba State University/UEPB (Paulo Eduardo Silva Barbosa, Adalcino Júnior Araújo Tavares, Rodrigo Brito do Nascimento).
Public managers from Tupã: managers of educational policies, health, social assistance, media, sports/leisure, with the mayor's formal approval.
Parents and guardians: mothers or guardians with anthropometric measurements recorded in the system before the COVID-19 pandemic. During the Hackathon period, these parents were approached by health agents' telephone to complement data in the computerized health system. Parents who had children in schools participated in the Hackathon product evaluation survey.
Electing users of the technological solution:
To meet the scope of the project, it was decided that the application should provide an open environment for alerts and information (later interaction) that are in the collective interest (managers, suppliers, associations, public policy councils, and managers in general) and a restricted environment to parents, educators and caregivers.
Selecting indicators and other inputs for the digital tool
Public data from the Health, Education, Social Assistance, and Planning Secretariat information systems were used. These information systems are nationwide. Data of interest were selected for the identification of children and their families. As an outcome variable, the classification of nutritional status was used according to the body mass index (BMI/age) generated by the Sistema Nacional de Vigilância Alimentar e Nutricional (SISVAN). Based on evidence from the literature, variables that influence nutritional status were included: the markers of food consumption (9 to 16 variables according to age group), information about the family (3 variables), information about the school environment (15 variables), information about the neighborhood where the families live (3 variables, expected to expand). Anthropometric and food consumption data for parents and pregnant women are expected to be included. This was not done because there was no data inserted in the municipality's health system. How these variables influence the children's nutritional status and the weight given to them was attributed based on consensus among specialists (with the tool's use, this information matrix can be validated).
Transforming inputs into solution
With this information, we started modeling the data. We used SAS to handle this process, employing fuzzy logic procedures to attribute numeric values, for each one of the following topics: food consumption, school environment, family structure and neighborhood environment. This kind of reasoning method can operate with multiple values of the same variable. This way, it was possible to create a scoring system to analyze complex data and display summarized information in the form of linguistic variables with gradual changes. For example, a child’s weight can be expressed as far below, below, adequate, above or far above, in relation to body mass index. These variations are also color-coded in the interface. The rules for evaluating each of those categories were combined and turned into algorithms. Python programming language was then used to validate the mathematical model and build the prototype.
Benefits
Let us give an effective response to childhood obesity with which all stakeholders will benefit from:
- Mothers or guardians will receive information about their children's nutritional status and how they should deal with the problems detected. They will have information about the school environment and the neighborhood where they live, facilitating, for example, their shopping.
- Educators will have information to monitor the child's development at school, provide care and guide mothers in adopting healthy lifestyles.
- Caregivers of health and social care services will have quick information on children's families' nutritional status and characteristics and the recommended procedures according to the diagnosis detected.
- Members of public policy councils, associations and society organizations, and interested citizens will be able to monitor the nutrition indicators of children in schools and neighborhoods and school and neighborhood environments.
- Food suppliers in nature will have information about the market for these products in the city, such as the spatial distribution of food outlets. Other food suppliers may be made aware of the type of food to be made available. Furthermore, the tool may, in the future, be a bridge between producers and final consumers, with an emphasis on the school community.
- Managers will have a quick and practical tool to support management and decision-making. They will be able to obtain improvements in children and family members' health and nutrition indicators, optimizing the use of public resources and public management.
- An indirect benefit may be the increase in data coverage and its use for decision-making in the health system, considering that a large part of the information currently generated is not used to evaluate and improve public policies.
- The tool can evolve as a laboratory to generate new hypotheses and evidence, allowing integration with researchers from different regions of Brazil and countries like Sweden and technological development oriented to the theme.
Market viability
Brasil2030 is a tool of interest for municipal public management, and it can be adapted to other market needs, such as, for example, databases produced in institutions and health plans, both private and public, scalable to the reality of other countries. Within the municipality, resources for operation and maintenance with resources from the Sistema Único de Saúde (SUS). However, this tool has great potential to attract interest from large funders, considering its ease of use and cost-benefit ratio. A manager will acquire the tool because it will provide him with an overview of the child's health situation linked to nutrition and healthy lifestyles and provide an overview of the community's resources to provide greater well-being for families.
It will facilitate decision-making in municipal departments while supporting service users' guidance and will increase directly (increased frequency of assistance in health units) and indirectly the coverage of health care and health promotion services. It will facilitate the work of educators and technicians and support mothers and guardians in childcare. It will contribute to their image as an efficient and innovative manager.
Another positive externality of the tool will be the empowerment of citizenship concerning their surroundings since it will enable direct communication between the citizen and the public manager through a channel of citizen denunciation (e.g., an indication of the impossibility of using public space due to lack of lighting at night, detection of vandalism in public facilities for the practice of children's games, lack of an open market in the sector, lack of green areas).
Challenges
A major challenge will be to print on the product its characteristic of a digital tool capable of facilitating decision making for healthier lifestyles in those who are part of the nucleus closest to children as mothers and other family members, educators, and caregivers. Similarly, for users of synthesis data, the challenge of printing the characteristic of agility, ease, and integration is presented in the face of a complex reality full of data and information that most of the time is not used for the evaluation and improvement of processes. Another major challenge will be to expand the tool's functionality, integrating more interactive processes capable of processing qualitative information, making the digital city closer to the real city.
Survey Results
From March 26 to March 31 we applied a survey with parents from Tupã city that used the Brasil2030 tool in a controlled pilot managed by public managers. We had 61 respondents in a universe of 180 parents. The answers are summarized in the figure below.
We can see that in all the eight questions the results were positive, showing great possibilities of high adoption of the tool. Further developments and investigations will be performed in the next days, showing that the hackathon was very beneficial for the fight of childhood obesity in the city of Tupã and the Brazilian scenario. Our proposal may contribute to slowing down the advance of childhood obesity in the region.
