Product Promotion Network


DeepMind’s AI can detect over 50 eye diseases as accurately as a doctor

Step by step, condition by condition, AI systems are slowly learning to diagnose disease as well as any human doctor, and they could soon be working in a hospital near you. The latest example is from London, where researchers from Google’s DeepMind subsidiary, UCL, and Moorfields Eye Hospital have used deep learning to create software that identifies dozens of common eye diseases from 3D scans and then recommends the patient for treatment.

The work is the result of a multiyear collaboration between the three institutions. And while the software is not ready for clinical use, it could be deployed in hospitals in a matter of years.

Those involved in the research described is as “ground-breaking.” Mustafa Suleyman, head of DeepMind Health, said in a press statement that the project was “incredibly exciting” and could, in time, “transform the diagnosis, treatment, and management of patients with sight threatening eye conditions […] around the world.”

The software, described in a paper published in the journal Nature Medicine, is based on established principles of deep learning, which uses algorithms to identify common patterns in data. In this case, the data is 3D scans of patients’ eyes made using a technique known as optical coherence tomography, or OCT. Creating these scans takes around 10 minutes and involves bouncing near-infrared light off of the interior surfaces of the eye.

Doing so creates a 3D image of the tissue, which is a common way to assess eye health. OCT scans are a crucial medical tool, as early identification of eye disease often saves the patient’s sight.

The software was trained on nearly 15,000 OCT scans from some 7,500 patients. These individuals had all been treated at sites operated by Moorfields, which is the largest eye hospital in Europe and North America.

The system was fed their scans alongside diagnoses by human doctors. From this, it learned how to first identify the different anatomical elements of the eye (a process known as segmentation) and then recommend clinical action based on the various signs of diseases that the scans show.

In a test where the AI’s judgments were compared with diagnoses by a panel of eight doctors, the software made the same recommendation more than 94 percent of the time.

Whose call is it anyway?

Results like this are extremely encouraging, but experts in the medical community are still worried about how AI systems will be integrated into care practices. Luke Oakden-Rayner, a radiologist who’s written extensively on the subject, says advances in AI are fast pushing us toward a tipping point where software is no longer a tool that’s applied and interpreted by a doctor, but something that makes decisions on behalf of humans.

The first systems are just beginning to cross this line.

In April, the FDA approved the first AI-powered program that diagnoses disease without human oversight. As one of the program’s creators put it: “It makes the clinical decision on its own.” (Coincidentally, like today’s new algorithm, this software also analyzes eye scans. But it only looks for one disease, diabetic retinopathy, whereas DeepMind’s is sensitive to more than 50 conditions.)

This is the point at which the risk from medical AI becomes much greater.

Our inability to explain exactly how AI systems reach certain decisions is well-documented. And, as we’ve seen with self-driving car crashes, when humans take our hands off the wheel, there’s always a chance that a computer will make a fatal error in judgment.

The researchers from DeepMind, UCL, and Moorfields are aware of these issues, and their software contains a number of features designed to mitigate this sort of problem.

First, the software doesn’t rely on a single algorithm making the decision, but a group of them, and each is trained independently so that any freak error will be overruled by the majority. Second, the system doesn’t just spit out a single answer for each diagnosis.

Instead, it gives several possible explanations, alongside its confidence in each one. It also shows how it has labeled the parts of the patient’s eye, giving doctors an opportunity to spot faulty analysis.

But most importantly, the software isn’t a straightforward diagnostic tool. Instead, it’s designed to be used for triage, the process of deciding which patients need care first.

So while it does guess what conditions a patient might have, the actual recommendation it makes is how urgently the individual needs to be referred for treatment.

These features sound incidental, but each of them operates like a speed bump, slowing the algorithm down, and giving humans a chance to intervene. The real test, though, will come when this software is deployed and tested in a real clinical environment. When this might happen isn’t known, but DeepMind says it hopes to start the process “soon.”

Gold from the data mine

Along with its clinical possibilities, this research is also interesting as an example of how AI companies benefit from access to valuable datasets.

DeepMind, specifically, has been criticized in the past for how it has accessed data from patients treated by the UK’s publicly funded National Health Service (NHS). In 2017, the UK’s data watchdog even ruled that a deal the company struck in 2015 was illegal because it failed to properly notify patients about how their data was being used. (The deal has since been superseded.)

Today’s research would not have been possible without access to this same data. And while the information used in this research was anonymized and patients could opt out, the diagnostic software created from this data belongs solely to DeepMind.

The company says that if the software is approved for use in a clinical setting, it will be provided free of charge to Moorfields’ clinicians for a period of five years.

But that doesn’t stop DeepMind from selling the software to other hospitals in the UK or other countries. DeepMind says this sort of deal is standard practice for the industry, and it tells The Verge it “invested significantly” in this research to create the algorithm. It also notes that the data it helped corral is now available for public use and non-commercial medical research.

Despite efforts like this, skepticism about the firm remain.

A recent independent panel set up by DeepMind to scrutinize its own business practices suggested that the company needs to be more transparent about its business model and its relationship with Google, which bought the firm in 2014. As DeepMind gets closer to producing commercial products using publicly funded NHS data, this sort of scrutiny will likely become increasingly pointed.

An eye on the future

Regardless of these issues, it’s clear that algorithms like this could be incredibly beneficial. Some 285 million people around the world are estimated to live with a form of sight loss, and eye disease is the biggest cause of this condition.

OCT scans are a great tool for spotting eye disease (5.35 million were performed in the US alone in 2014), but interpreting this data takes time, creating a bottleneck in the diagnostic process.

If algorithms can help triage patients by directing doctors to those most in need of care, it could be incredibly beneficial.

As Dr. Pearse Keane, a consultant ophthalmologist at Moorfields who was involved in the research, said in a press statement: “The number of eye scans we’re performing is growing at a pace much faster than human experts are able to interpret them. There is a risk that this may cause delays in the diagnosis and treatment of sight-threatening diseases.

“If we can diagnose and treat eye conditions early, it gives us the best chance of saving people’s sight.

With further research it could lead to greater consistency and quality of care for patients with eye problems in the future.”

Has train travel actually improved for passengers in the last decade?

The rail industry has had plenty of bad press in recent weeks, after timetable chaos led to mass frustration across the UK. It may come as no surprise, then, that new Which? analysis has found passenger satisfaction has stagnated over the past 10 years. We analysed official Transport Focus data, which revealed that satisfaction with punctuality and reliability of train services has dropped from 79% to 73% over the past decade.

These findings, alongside plummeting trust and soaring fares, paint a bleak picture. We dig into more of the detail of our broad analysis of Transport Focus data, below. Sick of all the train chaos?

Help us improve the rail system and support our call for a better rail service.[1]

How have passengers rated train services over the past 10 years?

As you can see from the graph below, there’s been little change in passenger satisfaction for several key aspects of rail over the last decade.

Less than half of passengers have been satisfied with value for money and handling of complaints across the past 10 years, according to our analysis. Satisfaction with value for money is stuck at an average of 46% in the year to spring 2018, having increased by just three percentage points across the decade for all passengers. The situation is even worse for commuters, with a marginal rise from 30% to 31%.

Satisfaction with handling of delays has remained stubbornly low, with only a four percentage point increase in satisfaction (year to spring 2008: 34%; year to spring 2018: 38%).

Falling trust and rising fares

Trust in the train industry is approaching its lowest point of the past six years, according to Which?’s consumer insight tracker. In July 2018, only 23% of people told us that they trust train travel. This represents a six percentage-point drop compared with July 2017.

In fact, this makes train travel one of the least-trusted consumer industries, beaten to last place only by car dealers. At the same time, rail fares have increased by 40% since 2008, which is more than one-and-a-half times higher than CPI inflation (26%) over the same period.

What does Which? think about all of this?

Reflecting on our recent findings, Peter Vicary-Smith, Which? chief executive, says: ‘With persistent poor service, delays, cancellations and the hassle of getting compensation for journeys, it’s unsurprising that trust in the rail industry has been consistently low and only getting worse. ‘Passengers expect increased satisfaction to come with the hike in their ticket prices, not a decade of disappointment and unprecedented disruption like many faced this year.

‘If the rail system is to start working for passengers, the government must step in to ensure they are automatically compensated for delays and cancellations, and the new Rail Ombudsman must be up and running as soon as possible so passenger complaints don’t continue to go unheard.’

Which are the best and worst train companies for commuters?

Grand Central was the best-rated British train company in our survey of 2,865 commuters in October and November last year. And it may come as no surprise that Southern Railway found itself at the bottom of the rankings.

When we survey train passengers, we ask them to reflect on their overall experience rather than a single journey, unlike the NRPS. But, like the NRPS, we question passengers on all the important aspects of their journeys, from punctuality and value for money to the cleanliness of on-board toilets.

We then assess all responses to generate star ratings and a league table of train companies. And as well as asking commuters, we also questioned 11,135 leisure rail users about their experiences. Head straight to best and worst UK train companies to find out how your train service stacks up against the rest.[2]

Your consumer rights with train pain

Here at Which?, we know how frustrating it is when trains don’t run on time, or when they’re cancelled.

We have several articles to help you get to grips with your compensation rights, to make sure you’re not left out of pocket, including:[3]


  1. ^ call for a better rail service (
  2. ^ best and worst UK train companies (
  3. ^ compensation rights (

Best mortgage rates not always offered to existing customers

Homeowners who accept a product transfer from their mortgage provider could be paying GBP600 more a year than they need to, new data shows. Some lenders are excluding their best deals from the product transfers they offer to customers at the end of their fixed-rate period, according to the The Telegraph. In the worst cases, this is costing homeowners GBP600 more per year than if they’d been transferred to the best deal available from that lender.

  • For advice on product transfers and help finding the best deal for your personal situation, call Which?

    Mortgage Advisers[1] on 0800 197 8461.

How do product transfers work?

When you’re nearing the end of your mortgage’s introductory deal period, your mortgage lender will usually offer you a ‘product transfer’. They will send you a list of available deals based on the equity you hold in your property. For example, if you’ve paid off 25% of the original price of the property, you’ll be offered a 75% loan-to-value (LTV) mortgages.

These may be more competitive than the deals available when you first took out the loan. There are two main benefits to accepting a product transfer. First, it will almost invariably be cheaper than being moved onto your lender’s standard variable rate[2], which is what will happen if you don’t accept a transfer or remortgage.

Second, it’s a lot quicker and simpler than remortgaging. You won’t be subjected to a credit check or affordability assessments, a property valuation won’t be carried out and there’s no legal work involved.

Do product transfers give you the best mortgage rates?

Some lenders will offer you exactly the same products as they would to remortgage customers.

If this happens, you will be able to take up the best rate available from your lender, although of course you might be able to get a better deal by remortgaging with a different provider. However, some mortgage providers offer a restricted list of deals, and this is where the problem lies. Researchers from The Telegraph[3] looked at two-year fixed-rate deals being offered as product transfers and found that Leeds Building Society, NatWest and Bank of Ireland UK didn’t offer existing customers their best rates.

When the researchers repeated the exercise for five-year fixed-rate deals, Leeds Building Society and NatWest again didn’t offer their best rates as product transfers. The worst scenario they found would have resulted in an existing customer paying GBP600 more per year than a new customer with the same provider. However, the researchers also found cases where existing customers were offered better deals than new customers.

Bank of Ireland, Platform Home Loans, Accord and Nationwide all offered cheaper five-year fixed-rate deals as product transfers.

What do the mortgage lenders say?

When The Telegraph put its findings to the lenders, it received the following responses:

  • Leeds Building Society: products cannot be directly compared as their features, including fees and incentives, vary.
  • Bank of Ireland: ‘some differences can occur from time to time due, for example, to timing or where some products have different features.’
  • NatWest: new and existing customers are currently offered the same rates if they’re applying directly, but some lower rates are available via brokers.

Should you remortgage or take a product transfer?

Usually, when you’re offered a product transfer, your lender will send you a list without any advice on which deal would be best for you. It can be very difficult to work out which is the best option, as arrangement fees often mean that the deal with the lowest interest rate isn’t cheapest overall. Comparing the product transfers with deals from other lenders can be even more confusing.

For this reason, it’s well worth taking advice from a whole-of-market broker such as Which? Mortgage Advisers[4]. Which?

Mortgage Advisers’ David Blake says: ‘It often makes sense to accept a product transfer, but this isn’t always the case. ‘A good broker will look at all your options and make an informed recommendation on whether there’s any point in remortgaging away from your existing lender as opposed to conducting a product transfer.’

How to find the best mortgage lender

When you take out a mortgage, you should research the lender’s customer service as well as its products. This will help you understand how well you’re likely to be treated towards the end of your deal.

At Which? we regularly survey thousands of homeowners about their mortgages. We use the results to create unique scores for all the major lenders, based on a combination of customer feedback and our expert analysis of how competitive their deals are. You can find out the best and worst providers in our mortgage lender reviews[5].

For expert advice on the right mortgage for your personal circumstances, call Which? Mortgage Advisers[6] on 0800 197 8461 or fill out the form below and they’ll call you back.

Your home may be repossessed if you do not keep up repayments on your mortgage. Which?

Limited is an Introducer Appointed Representative of Which? Financial Services Limited, which is authorised and regulated by the Financial Conduct Authority (FRN 527029). Which?

Mortgage Advisers and Which? Money Compare are trading names of Which? Financial Services Limited.


  1. ^ Which?

    Mortgage Advisers (

  2. ^ standard variable rate (
  3. ^ The Telegraph (
  4. ^ broker such as Which?

    Mortgage Advisers (

  5. ^ mortgage lender reviews (
  6. ^ Which?

    Mortgage Advisers (

1 2 3 741