Preventive Healthcare
Smits’ interest in preventive healthcare began during her bachelor’s degree in Nanobiology in Delft. During a minor in Munich, she took courses on nutrition, pathology, and health behavior with great interest. ‘For my masters thesis in Computer Science, I wanted to combine that interest with data research.’ This led Smits to TNO. ‘TNO conducts a lot of research on preventive healthcare, so I hoped I could join that.’ She was lucky. At the time she sent her request to TNO, CBS and TNO had just decided to collaborate more on health. ‘My research perfectly aligned with that due to the combination with data.’Data Research
The masters student was able to work as an intern at both CBS and TNO. She had the opportunity to conduct her research and also joined other research groups at CBS and TNO. ‘That was an enormous luxury. If I encountered something in the data research, I could turn to a CBS colleague. For questions about the medical part of my thesis, I consulted a TNO researcher.’Similarities
For her thesis, Smits analyzed CBS data from 290,000 people over the age of 40 who participated in the Health Monitor in 2016. She looked at people who were prescribed medication for type 2 diabetes at that time. ‘I wanted to know what their social networks looked like. Something I could only investigate thanks to the unique CBS dataset on personal networks,’ says Smits. ‘But I was also curious about the dining facilities these people had nearby and whether they lived near exercise facilities, such as parks and sports clubs.’ Smits then examined whether there were similarities in these areas among people who used diabetes medication.
Surprising Outcome
An outcome Smits did not expect was the correlation between colleague networks, groups of people who work in the same place and often live in the same municipality, and having diabetes. ‘The data showed that people who take medication for type 2 diabetes often have colleagues who do the same,’ says Smits. ‘However, this only applies to colleagues of the same gender. This relationship is not seen among male and female employees.’ Possible explanations Smits gives for this connection are social influence and similar underlying personal characteristics. ‘Direct colleagues often have a similar background in terms of education and socioeconomic status. Additionally, it is plausible that people who spend a lot of time together daily and are similar in backgrounds adopt habits from each other. A similar clustering has also been found in research on other lifestyle diseases, such as obesity.’Exercise Facilities
The data on the living environment of diabetes patients yielded less than Smits initially thought. Jeroen Pronk, behavior researcher at TNO and research supervisor of Smits, shared this expectation. ‘We know that an unhealthy diet promotes the development of type 2 diabetes,’ says Pronk. However, Smits analysis showed that the proximity of exercise facilities only makes a significant difference when they are very close to home, about within 1 kilometer. The data Smits used did not provide clarity on the influence of healthy dining options nearby.Policy Recommendations
Pronk sees a valuable contribution in Smits research to improve prevention. ‘There is actually very little known about the influence of the social and physical environment on the development of type 2 diabetes,’ Pronk explains. ‘More insight can lead to better and more targeted prevention.’ Smits already has some ideas about what better prevention looks like. ‘My research shows that you can tailor preventive measures to specific neighborhoods and communities due to the clustering of type 2 diabetes. Or even to certain companies and colleague networks.’ For example, Smits would like to see companies legally required to encourage their employees to exercise enough during work hours. By, for example, freeing up time for this or offering courses on this topic. ‘In many company canteens, attention is already paid to healthy eating. A similar development in the field of physical activity could help reduce diabetes.’Various AI Methods
Following the research Smits conducted, CBS ensured a follow-up. ‘CBS conducted research on the distribution of the use of diabetes medication in the personal network,’ says Edwin de Jonge. He is a senior methodologist at CBS and was also a supervisor of Smits. ‘The results of this research will be published later this year. Additionally, we plan to publish a scientific article on the various AI methods Smits used. The CBS personal network offers opportunities to conduct research on diverse topics and with various methods. Because the network is still relatively new and globally unique, studies like Smits’ help to get a better picture of application areas and analysis possibilities.’Relevant links
- News - More obesity and diabetes among adults with poverty risk
- CBS Privacy Rules - Privacy