Diabetes developments – Simon O’Neill

Simon O'NeillIn a regular blog series, Simon O’Neill, Diabetes UK’s Director of Health Intelligence and Professional Liaison, rounds up the latest diabetes news.

This week Simon updates on NICE guidelines and data protection.

New NICE guidance for Type 2 diabetes

This month NICE released the update to their guidance on the management of T2D. Following the launch of the updated guidance in 2015, NICE agreed to set up a rolling committee to ensure that new evidence was incorporated into guidance more rapidly. This is the first update using that approach.

The main change is the advice on SGLT-2 inhibitors. In the original draft NICE guidance these were largely overlooked, being added into the second revision. In this May update, SGLT-2s have been promoted up the chain. Metformin is still the first treatment of choice in those who can tolerate it, with a choice of dual therapy with DPP-4 inhibitors, pioglitazone, sulfonylureas or SGLT-2 inhibitors if HbA1c rises to 58 mmol/mol. In those who can’t tolerate metformin, SGLT-2s can now be used as a first line treatment, if a sulfonylurea or pioglitazone is not appropriate. This change appears to come as a result of the very positive cardiovascular outcomes demonstrated with the use of various SGLT-2s in trials such as EMPA-REG, CVD-REAL and DECLARE.

The guidance now also includes more information on the use of SGLT-2s, including the risk of DKA and the need to test for ketones in people with symptoms of DKA even if their blood glucose levels are normal. It also clarifies that any of the three (empagliflozin, dapagliflozin or canagliflozin) can be used as mono, dual or triple therapy or with insulin, though does stress, as with all drugs, that if there is more than one in a class, the cheapest option should be prescribed.

Interestingly the guidance doesn’t mention the increased risk of minor amputations that has been seen with this class of drug and is now included in the patient information leaflet as a potential risk. This was highlighted by the EMA back in February.

Data protection

Google’s DeepMind artificial intelligence programme and its relationship with the Royal Free London NHS Trust has come under investigation by the Information Commissioner’s Office (ICO) as a result of potentially sharing 1.6 million patient records without the necessary consent.

The partnership, which began in July 2015, was to develop an app that could warn doctors if a patient is at risk of developing acute kidney injury. But in April 2016 questions were raised about whether DeepMind obtained proper patient consent for tests of the app. The partnership had implied patient consent to share their data with DeepMind because the information was being used to improve patient treatment, a legal basis known as “direct care.”

However, Dame Fiona Caldicott, the National Data Guardian, felt that during the pilot test of the app, called Streams, the main goal was to make sure that the app functioned well, not to improve patient outcomes, and therefore consent should have been sought. She stressed that, though she appreciated the great benefits that new technologies can offer to patients, patient data should always be shared in a transparent and secure manner in order to build public trust.

Google have been quick to stress that the data was not shared with any other part of their business or used for any commercial purpose and is held only within the DeepMind service.

The ICO have been working with DeepMind, the Royal Free NHS Trust and the National Data Guardian’s office to finalise their investigation, which is expected soon.

This highlights an issue for new technology which can access big data to investigate potential patterns that haven’t been visible before. For such an app to work, it needs to be tested to ensure that it is giving safe and reliable advice but, with a reluctance for patients to share data, how do you generate enough data to test. Hopefully the ICO will make recommendations to enable the Government to provide clear guidance on the development of such big data technologies.

You might also like