It is well-known that obesity is related to other health problems, but few researchers have been able to measure the relationship in monetary terms. Even fewer have examined the issue from a population perspective.
In the study by Edwards et al. that appears in this issue of the Journal of the Norwegian Medical Association, the authors use data from the Nord Trøndelag Health Study (HUNT) to show that the economic costs of specialist healthcare services increase with body mass index (BMI) (1). While the greatest costs are attributable to BMI over 35, this represents a relatively small group of people. Considerable costs are associated with the group with less severe obesity, which is a much larger group. From this perspective, it is equally important for society to combat overweight as to combat morbid obesity.
Perhaps the finding is not surprising, but it is a useful reminder of the need to adopt an overall perspective. At the individual level, many phenomena show increasing harm attributable to a single variable but there are few people at the extremes. Alcohol consumption is a good example. Some researchers claim that there has been too much focus on the few with the highest consumption (2). Ultimately, the correct balance will depend on the benefit and costs of specific interventions, not just the number of individuals.
Indirectly, the study by Edwards et al. also illustrates the importance of large population studies. Although Norway has a large body of good register data, this study could not have been based on register data alone. Population studies such as HUNT and the Tromsø Study include information about weight, height, diet, smoking habits and alcohol consumption that does not appear in the registers. The registers identify those receiving various diagnoses and the interventions carried out. Population studies can provide information about the possible underlying causes of diseases. Thus registers do not eliminate the need for population studies. They are useful both separately and, not least, in combination.
Norway has a big advantage in this respect with both good register data and large population studies. But the existence of data is not sufficient. Unfortunately, access to the figures, and especially to links between different registers and studies, is still difficult, and requires considerable effort in the form of applications, time and money. The process often costs hundreds of thousands of Norwegian kroner and can take several years.
Researchers should have broad access to standardised datafiles with key variables from both registers and population studies
The health sector should learn from Statistics Norway. They offer a service – microdata.no – whereby you can log on and analyse income and social security data at an individual level while data about individuals remain anonymised and invisible, and stay on the platform at all times. Users only see the results of the analyses they run. The service is not perfect and does not cover all areas of use, but if we had such a system for health data, it would make it far easier, faster and cheaper to identify many key correlations.
In addition to providing simpler access to health data, an infrastructure should be created that would make it easier to perform analyses quickly. For example, it often takes a long time to obtain, clean and link data. In my opinion, instead of each of us sitting in our own office compiling datasets and algorithms, the researchers should have broad access to standardised datafiles with key variables from both registers and population studies.
It appears that at central level, there is a move in this direction. The first version of the national health analytics platform from the Directorate of e-health foundered because of legal problems. However, the new strategy involves the re-use of existing solutions and stands a better chance of success. If we continue to invest in this type of infrastructure, it may result in many more commendable research articles. This will probably also be useful in a population perspective.