Diabetes developments – by Simon O’Neill
In a regular blog series, Simon O’Neill, Diabetes UK’s Director of Health Intelligence and Professional Liaison, rounds up the latest diabetes news.
Big Brother is watching you…
For those of you who remember the supercomputer HAL, in 2001, A Space Odyssey, the idea of supercomputers may come as a rather worrying trend. But with IBM Watson crunching numbers and predicting hypos hours ahead of time, Google’s DeepMind Health Division is now working with the Moorfields Eye Hospital NHS Foundation Trust to try and improve the diagnosis of eye conditions – diabetic retinopathy and age-related macular degeneration (AMD) – two of the main causes of sight loss in the UK.
One particular scan of use in the diagnosis of these conditions is optical coherence tomography (OCT), but these are very complex scans and require specialists to analyze them. With more than 3,000 OCTs being carried out a week at Moorfields alone, although analysis for an individual patient is possible, help is needed to analyze these scans as quickly and accurately as possible. And with the brain power of a supercomputer, the mass of data that’s available could help us identify whether there are trends in the development of these eye problems which could help improve treatment options.
This is where DeepMind Health’s platform comes in. The hope is that by training DeepMind on a million of these images, it can help with faster and more accurate analysis to determine the best treatment option for each patient. At this stage the project is only a research programme but the potential for improvements in eye health outcomes is huge.
DeepMind has already found itself embroiled in a patient data controversy when it entered into a data-sharing agreement with the Royal Free London NHS Foundation Trust. The agreement meant that DeepMind had access to highly sensitive patient data without having consent from individuals. Moorfields are obviously acutely aware of such a sensitive issue and have agreed only to share approximately one million anonymised digital eye scans.
As long as they don’t have the same problem as the astronauts in the film when they tried to turn HAL off – “I’m afraid that’s something I cannot allow to happen”.
Yes to telehealth
A new study from the US asked 1,700 adults how they would like to communicate with their primary care team and whether technology should play more of a role.
Currently the vast majority of people still use traditional channels to communicate, with 26% regularly making appointments in person and 76% using the phone. However 59% said that they would now choose a primary care team who offered a mobile app for communication rather than a team who don’t. For millennials (those born between the 1980s and early 2000s) this rose to 70%, indicating how much this generation relies on technologies in all aspects of their daily lives.
62% of those questioned also indicated that they would be open to having telemedicine consultations – either using phone or Skype type technology rather than having face to face meetings. This seems to be growing as more people are using this in their daily lives as a main method of communicating.
Most interestingly was the desire for data from wearable devices, such as FitBits or CGM systems, to be more easily shared with their HCPs. 78% of people thought that their doctors should have access to this data, both to keep their medical record more up to date but also so that HCPs could provide more tailored and personalized care.
Perhaps less surprising, in an insurance based health system, around 66% of millennials thought they would be more likely to use a health tracking device, provided by their insurer, if that led to lower insurance premiums based on data from the device. Although less relevant to the NHS, other similar incentives are already being used in the UK. Aviva Drive is a free app that monitors your driving skills. Once you’ve driven 200 miles, you get an individual driving score out of 10. Safer drivers scoring 7.1 or more could save an average of £150 on Aviva comprehensive car insurance. Private health insurers in the UK are also investigating these possibilities.
Having said that, a smaller survey of 500 people found that the main functions they wanted from telemedicine were not access to health trackers, but the ability to view their medical record and make appointments online and the ability to access after care information following a procedure
Scientists at Tufts University in Massachusetts are developing ‘smart’ stitches for wounds that are able to detect how well a wound is healing and whether there is infection developing. Currently wound assessment is done visually, looking for signs of infection or breakdown but with these stitches, there is the ability to send an automatic alert to staff to indicate that something is wrong, hopefully at a much earlier stage where further intervention may protect the wound.
The scientists integrated nano-sized sensors into threads and then dipped them in chemical sensing compounds before attaching wireless electronic circuitry. The stitches were then used in the normal way in animal studies. This allowed the research team to monitor the pressure, stress and strain a wound was under, how well it was healing, and whether the body’s chemistry was out of balance and toxicity was developing. The smart stitches can also monitor glucose levels, which could be potentially very useful in people with diabetes, where wound healing can be slow.
Although these threads have been used simply as sutures, the researchers see many other possibilities. Because thread is cheap, thin and flexible it could be developed for other functions such as smart bandages to monitor wound healing where sutures are not needed.
The research is still in an early stage, for example the wireless technology could only communicate with a mobile phone that was up to 10 meters away, but further work is planned and the team are looking for volunteers to trial the smart stitching at skin level rather than on more complex and deeper wounds. Further studies will then be needed to ensure the sutures are safe and not likely to damage wounded tissue.
Type 1 vaccination
As previously discussed, Type 1 diabetes (T1D) is now being seen as something that is potentially preventable. A speaker at the American Diabetes Association conference described waiting for Stage 3 (in the very helpful Diabetes TrialNet diagram below, when T1D is actually diagnosed), as rather like waiting for someone to have a stroke and then managing it from then on. If you know someone is at risk of stroke, you do whatever you can to reduce that risk and prevent the stroke, such as lowering blood pressure and cholesterol. The problem in T1D is knowing what to do in the progression towards T1D that can actually have any benefit in preventing it from developing.
A lot of work is being done to investigate the immune activation/response phase, to try and prevent the body’s own immune system from destroying beta cells and thereby either preventing, or delaying, the onset of T1D. One approach is through vaccination – trying to teach the immune system that beta cells aren’t the bad guys – and there are several researchers trying different methods.
Research at UCL is injecting fragments of lab produced beta cells into people with newly diagnosed T1D (who probably still have about 50% of their beta cells remaining) to see if these fragments can familiarize the immune system to beta cells and stop the attack process – rather like desensitizing someone who has an allergy. The reason that the trial is with people newly diagnosed with T1D was in case the injections of beta cell fragments had the opposite effect and actually over activated the immune response. Early signs are promising and, if proven to be effective, this could be a treatment given to those who already have the genetic tendency to T1D and some of the antibodies.
Another approach, from Boston US, is looking at Tumour Necrosis Factor (TNF) which appears to be involved in increasing the production of ‘good’ T cells and reducing the number of ‘bad’ T cells which are normally involved in autoimmune attacks. Interestingly the BCG vaccination (which everyone in Europe and the US used to get to protect against tuberculosis) appears to increase the amount of TNF, and therefore, it’s hoped, the number of ‘good’ protective T cells and removing the ‘bad’ T cells linked to the destruction of beta cells. Higher levels of TNF also seem to protect remaining beta cells. Again being trialed in people already diagnosed with T1D, the use of BCG seems to preserve some beta cell function which has resulted in more stable blood glucose levels. Again, if it works, this could be used on people at high risk of T1D.
A third approach, from Dresden, is looking at the use of oral insulin to prevent the immune response from happening in the first place. The theory behind this is ‘oral tolerance’ which believes that the immune system won’t respond to things that it first encounters in the mouth or digestive system – which is why very few of us are allergic to foods. Insulin taken orally has no effect on blood glucose levels, as it is broken down and digested, so there is no risk of hypoglycaemia regardless of the dose given. A similar approach was trialed before with no overall results, but in this new trial, the researchers are giving 10 times the amount of insulin to very young children at very high risk of T1D. It is too early to say whether this is going to be any more successful than the previous trial, but the researchers are hopeful that this might be an easy treatment option for preventing T1D.
But why do T cells in the immune system attack beta cells in the first place. A team in Denver have been investigating that for many years and seem to have found the answer, at least in mice. The immune system is designed to destroy things which shouldn’t be inside the body – such as viruses and bacteria. So why do they fix on the beta cell? Beta cells modify two compounds which bind together to form insulin. The team wondered if occasionally these compound strands (peptides) might bind wrongly and produce something which they called hybrid insulin peptides.
When these were exposed to T cells, there was a strong immune reaction – so though the individual peptides are quite acceptable to the immune system, the hybrids aren’t recognized and are seen as a threat. They also tested human T cells, which also had a strong reaction to the mice hybrid insulin peptides, which supports the theory that this is also happening in humans. This has potentially opened up a whole new area of research to prevent the immune attack, either by stopping production of the hybrids or teaching the immune system that they are safe.