Express front page headline on diabetes is based on flawed research
24th Sep 2019
One in four hospital beds will be occupied by diabetics by 2030.
It’s difficult to know how many hospital beds will be occupied by diabetics in 2030, but the way in which this estimate was calculated is seriously flawed.
“Diabetes: the forecast to shock Britain: One in four hospital beds will be occupied by diabetics in 11 years”
Daily Express front page, 18 September 2019
A recent Daily Express front page claimed one in four hospital beds will be occupied by a diabetic in eleven years. This is based on a piece of research by Diabetes UK which has a number of problems, and should not be used as an accurate prediction or forecast about the future.
That said, it’s likely that the number of people with diabetes will increase over the next decade, as the population ages, and that will impact on hospital bed use. But we just don’t know by how much exactly. It’s also important to note that most people with diabetes in hospital aren't there to be treated for their diabetes.
How many people and hospital patients have diabetes?
An estimated 7% of patients aged 17+ at GP practices in England had diabetes in 2017/18. The figure rises to nearly 9% of the adult population using survey data which tries to estimate undiagnosed cases as well. Similar estimates for Wales suggest nearly 10% of adults have diabetes – diagnosed or otherwise.
People with diabetes are at increased risk of medical conditions like heart attacks and strokes, and so in general are more likely to be admitted to hospital.
In 2017, it was estimated that one in six (17.6%) NHS hospital beds in England and Wales was taken by someone with the condition, up from 15% in 2011. Those figures are a snapshot of the situation in NHS hospitals on a particular day.
The vast majority of these people aren’t there to be treated for diabetes.
Drawing a straight line into the future
The Express wrongly states that the one in four figure comes from an official estimate from the National Diabetes Inpatient Audit. This is the source of the underlying figures, not the 2030 prediction.
The figure comes from unpublished estimates from Diabetes UK, based on assuming that this growing trend in hospital beds being occupied by diabetics continues at the same rate for the next decade.
In other words, if you essentially draw a straight line between the rate of beds occupied by diabetics in 2010 (Diabetes UK gave us a figure of 14.9% for this year) and the rate in 2017, and continue that straight line into the future, it hits 22.6% in 2030, which is generously rounded up to “one in four”.
This has little value as a prediction or forecast about the future and shouldn’t be described as one.
Firstly, we’ve seen no evidence to suggest why the rate of increase between 2010 and 2017 might stay the same over time. The rate of increase between 2010 and 2017 wasn’t constant, growing more slowly between 2016 and 2017 than in previous years—though there’s no way to know whether this is the new ‘normal’ or if it’s just a blip.
Secondly, this sort of calculation just doesn’t hold up. When you just draw a straight upward line based on past data, eventually it reaches 100%—and then passes it. So by the logic of this method, eventually 100% of patients in hospitals will have diabetes.
Diabetes UK told us it made a separate estimate using a diabetes prevalence model produced by Public Health England which supported the findings.
However when we tried to replicate its calculations, we found that current projections estimate a 10% rise in the proportion of adults with the condition between 2017 and 2030. If you apply that increase to hospital attendances, the proportion of hospital beds occupied by diabetic patients would rise from under 17.6% in 2017 to roughly 19.3% in 2030. This doesn’t support the one in four estimate.
In response to our concerns, Diabetes UK said:
“Our report contained an illustrative prediction based on an extrapolation of existing data – designed to highlight the scale of the issue. It was not intended to represent an extensive model or forecast. Any predictions we make are always based on the most current information we have available to us.”
So how much more common will diabetes be in 2030?
Even if these estimates don’t give a reliable picture, the underlying story is sound: diabetes is expected to become more common in future both in the general population and in terms of hospital admissions.
Professor in diabetes and endocrinology at the University of Southampton, Richard Holt, told us it was likely the numbers of people with diabetes would increase:
“Over the last few decades, the number of new cases has been increasing dramatically as a result of the obesity epidemic, low levels of physical activity and changes in dietary patterns. Other factors, such as urbanisation, also contribute.
“Type 2 diabetes is a condition which occurs more commonly in older people and so the ageing population has also increased the numbers of new cases. There is some evidence, however, in Western countries that the rate of increase is slowing down.”
He also pointed out that better screening programmes in countries like the UK have led to earlier diagnoses of the condition, and that would likely lead to increases in the prevalence of the condition.
As for the effect on hospital admissions, Professor Holt agreed that while people with diabetes have increased health risks, “the relative likelihood of admission will probably not change dramatically over the next few decades and so as the numbers of people with diabetes rise, it is likely that the numbers in hospital with diabetes will increase.”
Dr Mayank Patel, an adult diabetes specialist working as a consultant at University Hospital Southampton, told us he thought the one in four prediction would turn out to be right.
“Of course things could change, but these are realistic projections. UK trusts currently have 15-25% diabetes prevalence. Patients are living longer with diabetes and more patients are being diagnosed with type 2, so the prevalence rises. It must also be remembered that the vast majority of hospital inpatients are not there for a diabetes reason, they happen to have it. For various reasons, having diabetes can increase the length of hospital stay and therefore put pressure on hospital capacity.”