Please could you give a brief introduction to Type 2 diabetes?
Type 2 diabetes is a condition in which the body does not make sufficient insulin or the insulin it makes doesn’t work sufficiently well to maintain a normal level of blood glucose and so the blood glucose rises. The problem with this is that it leads to an increased risk of complications such as heart disease, stroke, amputations, kidney failure, blindness and so on affecting the large and small blood vessels in the body, leading to huge costs for individuals, the health system and wider society.
Type 2 diabetes is increasingly common. Perhaps up to 4.5 million people will have it in the UK by 2020. It is more common in people who have a family history, people who are overweight, and people who have other related conditions like high blood pressure. It is more common in different ethnic groups like Asians and Afro-Caribbeans.
How many people are known to have type 2 diabetes?
366 million people have type 2 diabetes worldwide, 80% of whom live in low and middle income countries. In the UK it is probably around 5% of the adult population. Historically, it has been believed that a similar proportion has diabetes but don’t know it yet.
The prevalence of type 2 diabetes varies according to different countries in the world. In some countries, such as some pacific islands, up to 50% of the population develop diabetes. In most high income countries you’d expect the prevalence to be single figure percentages.
What are the main problems of having undiagnosed diabetes?
Most people that have undiagnosed diabetes don’t feel any symptoms. The issue is the harm that the raised blood glucose causes to the large and small blood vessels. This begins to happen before diagnosis, such that when people are diagnosed with diabetes many of them have complications already present. Up to 50% of people have some signs of tissue damage at the point at which they are diagnosed.
How did your research into screening for type 2 diabetes develop?
That is a long story. It started off while I was doing a study when I was in Southampton, which was focused on trying to improve the care of people with newly diagnosed diabetes. What I noticed was that all the people with newly diagnosed diabetes that the practices recruited had certain characteristics about them that made me think that it would be quite easy to find these people earlier because they were clearly different from the age/sex matched population in the practices. For example, they were more overweight, they were more likely to be on blood pressure drugs and so on.
So I developed a score to see if that would help in identifying people who had diabetes. The question then was if we can find these people, should we? That led to a whole program of work trying to decide whether it was worthwhile trying to find people with diabetes earlier and if we did then what treatment should they be offered.
The study that we published is one example of the studies that led from that initial interest back in the late 1990’s.
What did your research involve?
We recruited general practices in and around Cambridge and surrounding counties, and, with their permission, we helped them to look in their medical records with a search using the score I just mentioned. We identified the people aged 40-69 who were at highest risk of having undiagnosed diabetes. We took the top 25% of those at highest risk of undiagnosed diabetes and randomly allocated the practices to either screen or not screen.
In the practices that were screening, we gave the list of people at high risk to the practice and we asked them to invite those people for finger prick testing. So, they wrote to all of those people offering them an appointment and if they didn’t attend that appointment to arrange a time for a follow-up one. Around 73% of those people at high risk attended in the screening practices. We also had a comparison group where we didn’t tell the GPs who the high risk patients were, so they didn’t systematically invite people for screening.
What were the results of your research?
What we found 10 years later was that there was no difference in mortality in the high risk populations in the screening practices and the high risk populations in the control no-screening practices. This suggests that population screening for type 2 diabetes in the short term won’t lead to reduction in premature mortality rates in the wider population.
Why do you think there wasn’t a reduction in mortality rate? Is this because you only looked over 10 years?
It could be that if we look over a longer period then we might see a reduction. But, I think one of the main things is that if only 2.9% of the population screened were found to have diabetes, then you have to have quite a big impact in that small group to make a difference to overall population mortality.
What we didn’t do was have any kind of intervention or specific care for the 97% of people who didn’t have diabetes. What we have shown is that those people weren’t falsely reassured and it doesn’t look like they went away and had a worse diet and increased their glucose and risk, but it is unlikely they derived much benefit from the screening.
What we have shown, in a related publication, is in that 2.9%, those people that were found earlier, they do seem to have done better than patients that are clinically diagnosed with diabetes. And we have also demonstrated that intensive treatment in that group does seem to reduce risk of heart attacks and premature deaths. It is just that that group represents a relatively small number out of the total population of people at high risk. This is why I think it didn’t impact on population mortality.
But, it was important for us to quantify that, because in most screening tests when you evaluate them, people just focus on the people with the disease and not the overall effects of the screening program. In breast cancer screening there is some concern about over-diagnosis, over-treatment and over-investigation. By comparing total mortality in the group that is invited for screening and the group that is not you get some assessment of whether you are doing overall harm or overall benefit.
Was your research limited in any way?
There were some limitations – as there are with all research. One of the key ones was where we are based and where we did the research is probably a more affluent area compared to the rest of the country and other parts of the world. It is also an area with a largely white population. Consequently, we can’t say that screening in other areas might not have a bigger benefit. Even if you screen in areas where the prevalence is higher, in this country you are still only going to identify perhaps 6% of the population with diabetes, rather than 3%. And so, it is still questionable whether this would affect overall population mortality.
The other issue is that if you do any screening programs in more deprived areas the attendance rate tends to be lower, less than the 73% that we found, which reduces the population benefits of screening.
What impact do you think your research will have?
I think what would be nice to happen is that we don’t rush into doing a whole screening program for diabetes and perhaps use our research to suggest ways to adjust current procedures. The NHS is currently running health checks. Rather than inviting everyone aged 40-75, maybe they could be more targeted to those at higher risk. I think our study and others that we have done point in this direction.
What plans do you have for further research into this area?
We will follow up this group of patients over a longer period. We have also sent a sample in the screening arm and control arm questionnaires to identify whether they are more or less depressed, whether they have had more or less heart attacks, whether they are on more or less medication and so on.
We focused on mortality in this paper but we are also looking at morbidity and quality of life and so on in another bit of analysis that we submitted for publication.
Would you like to make any further comments?
I think one of the key issues for me is that screening is not quite the same as healthcare. I am a family doctor, and if someone comes to see me because they have a problem, I try to do the best I can for them given the limitations of my ability and knowledge and the limitations of medical evidence.
Screening is different, because in screening you go out and actively call people in who are healthy and well and minding their own business. You call them in to do tests on them. And so, I think the level of evidence that we need should be at least as high, and probably higher than for treatment, where we are responding to someone’s expressed need.
So, I think it is important that we do randomized trials of screening before we roll screening programmes out into a policy. I think it is important that we did this sort of work to demonstrate at the very least that screening doesn’t cause harm, in this case mortality. This is one of the messages from the paper: it doesn’t look like we’ve increased mortality by screening. And we have already demonstrated that it doesn’t cause false reassurance or lots of anxiety and does appear to benefit the small minority whose diabetes is detected earlier.
Where can readers find more information?
http://www.thelancet.com/journals/lancet/article/PIIS0140-6736%2812%2961422-6/abstract
About Dr Simon Griffin
Simon Griffin is Assistant Director of the Medical Research Council Epidemiology Unit (www.mrc-epid.cam.ac.uk), Deputy Director of CEDAR (the UKCRC Public Health Centre of Excellence for Diet and Activity Research, http://www.cedar.iph.cam.ac.uk/), an Honorary Consultant at Addenbrooke’s Hospital and NHS Cambridgeshire, primary care lead for the Eastern Region Diabetes Research Network and an assistant General Practitioner at Lensfield Medical Practice.
He qualified from the London Hospital Medical College in 1986 and trained in Clinical Epidemiology and Public Health at the University of Southampton and the London School of Hygiene and Tropical Medicine prior to his appointment to the University of Cambridge.
He leads a research programme which contributes to efforts aimed at preventing the growing burden of diabetes, obesity and related metabolic disorders by translating epidemiological knowledge into preventive action, and evaluating the effectiveness of different preventive approaches in clinical trials.
He has been awarded over £36M in extramural research grants and authored over 170 publications. Away from work Simon plays soccer and surfs.
Maternal diabetes linked to a slight increase in ADHD risk in children