Diabetes Screening - What's the use?

Given the high incidence of diabetes, could population-based diabetes screening make sense? This question was addressed by Prof. Dr. Oliver Kuß in a symposium at the DDG Congress 2021.

An overview of feasibility, effectiveness and appropriateness.

Given the high incidence of diabetes, could population-based diabetes screening make sense? This question was addressed in a symposium at the DDG Congress 2021. Prof. Dr. Oliver Kuß from the Institute of Biometry and Epidemiology (Institut für Biometrie und Epidemiologie) explained the criteria for the feasibility, effectiveness and appropriateness of population-based diabetes screening and summarised the evidence, which mainly comes from large Scandinavian registries.

Initially, Prof. Oliver Kuß narrowed down the topic with definitions.

What is screening anyway?

• Classification of asymptomatic persons with regard to a probability of disease or the identification of those who fulfil diagnostic criteria with the aim of treating them and preventing disease and death.
• It is part of secondary prevention as distinct from a preventive medical check-up on the one hand and diagnosis on the other hand.
• It manifests itself in two forms: Population-based screening - directly targeting the general population, and opportunistic screening - screening people who come into contact with the health system, regardless of the target disease.

Criteria for screening

• Of course, the disease should be sufficiently relevant in terms of incidence and prevalence, as well as in terms of causing suffering, e.g. reducing quality of life.
• It should be detected when it is not yet clinical, i.e. it should have a sufficiently long preclinical phase.
• The diagnostic procedure should be simple and safe, with high diagnostic quality in terms of sensitivity, specificity and predictive values.
• In diabetes screening, there is a two-step approach. First, all persons are tested with a test that is as sensitive as possible, but non-specific, possibly only with a risk score. If the result is positive, a specific clarification is carried out.
• If the result is positive, there should be effective treatment, with evidence that early treatment is better than late treatment.

The screening programme

• should have shown effectiveness in randomised trials
• the benefit should be greater than the harm
• must make sense from a health economic point of view

The current state of research

As far as type 2 diabetes is concerned, these points have been the subject of international debate for some time. There are several pro- and contra-reviews, especially in England and the USA.

Prof. Kuß gave two examples of randomised studies.

The ADDITION-Cambridge study: It was conducted in three countries - the UK, Denmark and the Netherlands. It was compared whether people were invited to a one-time screening or not. The outcome (all-cause mortality, cardiovascular mortality and cancer mortality) showed no relevant improvement after screening. In terms of mortality, the screening group was even slightly ahead of the comparison group that was not invited for screening. The one-time screening for diabetes did not result in a reduction in mortality.

The Ely study: Ely is a small town near Cambridge. Here, a randomised study was carried out among 15,000 inhabitants with two rounds of screening. Conclusion: Again, screening had no effect on mortality.

Another, non-randomised study from Denmark asked: Do people who are invited for screening have a better outcome than people who are not invited? Again, they saw a perfect null result - with 2 million observations. There were no advantages for the screening group.

A net effect?

But one can also ask the question differently: Is there perhaps some kind of net effect? Does screening benefit those whose disease is detected? This has been investigated in Denmark. Here, a distinction has to be made between those who were discovered during screening and those who only later showed symptoms of diabetes. 150,000 people were invited. The group diagnosed with diabetes was treated accordingly. The comparison group was only treated after they had become clinically conspicuous. There were clear differences. There was a significant survival advantage for the people who were detected at screening before they became clinically conspicuous. A Swedish study came to the same conclusion.

Conclusion

• In countries with well-functioning opportunistic screening, population-based screening programmes are unnecessary.
• The financial resources for such programmes are better spent on the treatment of clinically diagnosed people.
• There is little evidence on the effectiveness of population-based screening programmes for type 2 diabetes, even though the criteria for screening are fulfilled.

Reference:
DDG-Kongress 13.5.2021, O. Kuß, "Populationsbasiertes Screening für Typ-2-Diabetes: Evidenz aus Beobachtungsstudien"