select search filters
briefings
roundups & rapid reactions
factsheets & briefing notes
before the headlines
Fiona fox's blog

sedentary behaviour and diabetes

Publishing in the journal Diabetologia, a group of scientists have investigated the results from a study of activity and report that sedentary time was associated with an increased risk of type 2 diabetes.

 

Title, Date of Publication & Journal

Associations of total amount and patterns of sedentary behaviour with type 2 diabetes and the metabolic syndrome: The Maastricht Study

Diabetologia, 2016-Feb-02

 

Study’s main claims – and are they supported by the data

The paper does not show that an extra hour of daily sedentary time increases your risk of type 2 diabetes.

Sedentary time seems to be associated with diabetes. But that could mean being more sedate leads to a higher risk of diabetes; or it could mean having diabetes causes people to be more sedate. We cannot know which from this observational study.

The reported increased risk is 22% per extra daily hour of sedentary time. But the paper provides very little evidence beyond that first extra hour. So it would be premature to conclude that each extra hour is associated with the same additional risk.

 

Strengths and limitations

Limitations

The study is observational, so we can’t draw any robust causal links between behaviours and diabetes.  The authors acknowledge this, reporting all findings as ‘associations’ rather than confirmed causal links/mechanisms.

Without measurements over time, we cannot differentiate between:

Having diabetes causes people to be more sedate

Being more sedate leads to a higher risk of diabetes

So the reference to increased risk of developing diabetes is wrong – that suggests diabetes followed being sedentary, and we have no way of knowing that.

The reported 22% figure relates to how reported sedentary time can be used to predict who has type 2 diabetes. That is not the same as saying being sedentary causes type 2 diabetes.

22% is a very large relative-risk, but we are not provided with the absolute risks. So we cannot tell what the actual risk is before or after that extra hour of sedentary time – it’s hard to understand the impact of the findings

The analysis presents a 22% increase in risk for EACH additional hour (above the mean). However, there is really only a practical range of one hour between the least and most sedate individuals.

The authors say they have accounted for other factors (notably smoking, alcohol, CVD, BMI, education and mobility). But some of these are expected to have a big impact, the size of which are not provided. That makes it hard to assess the 22% risk in context – is that big compared to those other factors? It’s strange that information is not provided.  Sedentary time seems too simple a measure for predicting increased risk.

What is the public health message here? Should we be more active – or less sedate? In other words, do we need people to sit at their desk for less time during the day, or go for a jog in the evening? The paper does not provide the information to answer these questions, and the two are very different.

Strengths

The study size is very large, including a representative sample if individuals from the general population.

The authors include many potential confounders (but do not report their effect-sizes).

The authors stress the results are associations.

Note: Sedentary is defined as the total amount of time spent lying or sitting while awake, added up over the course of a whole day.

 

Additional information:

Clinical trials in Europe, Asia and the USA have previously shown that lifestyle modification in those at high risk reduces the risk of developing diabetes.

 

Any specific expertise relevant to studied paper (beyond statistical)?

None.

 

Before The Headlines is a service provided to the SMC by volunteer statisticians: members of the Royal Statistical Society (RSS), Statisticians in the Pharmaceutical Industry (PSI) and experienced statisticians in academia and research. A list of contributors, including affiliations, is available here.

in this section

filter Headlines by year

search by tag