Read about the research we’ve funded
Detecting glaucoma using a combination of low-cost, portable and easy to perform tests
The 2019 IGA Ophthalmology Research Grant of £99,960 was awarded to Dr Victor Hu. We met him at his office in the London School of Hygiene and Tropical Medicine (LSHTM) where he works two days a week, spending the other three as consultant ophthalmologist at Leighton Hospital in Cheshire. By happy chance we also met up with Dr Hu’s collaborator Dr Pete Jones, Post-doctoral Research Fellow at the Crabb Lab at City, University of London.
Dr Hu’s team have a clear long-term goal: to develop a single device that is capable of performing all of the clinical tests, analysing the data, and producing a measure for the likelihood of glaucoma. Such a device would have the potential to aid glaucoma diagnosis and reduce the number of people presenting with advanced disease. The project should help determine whether such tests are feasible and accurate in a low-income setting and could in time provide better methods of glaucoma detection in Sub-Saharan Africa, a region with high glaucoma prevalence and a lack of conventional diagnostic equipment.
The global prevalence of glaucoma for the population aged 40-80 is estimated to be 3.5 per cent with over 60 million people affected. In the UK around two per cent of people over 40 are affected, but in low-income areas prevalence appears to be much higher and it is a far more common cause of blindness. Recent population-based surveys from Sub-Saharan Africa found a prevalence of four – eight per cent in those aged over 40, and a high proportion of severe vision loss in both eyes.
Dr Hu’s team will test the feasibility and reliability of three low-cost tests in different field conditions - in the UK and Tanzania - and use the results to develop an optimised algorithm for detecting glaucoma:
- Optic disc photographs - using a Remedio hand-held fundus camera (see glossary at end of page) with a smartphone attached
- Eyecatcher visual fields - the Eyecatcher performs buttonless visual field testing by combining an ordinary tablet or laptop computer with an inexpensive eye-tracking device. The aim is to provide a cheaper and more intuitive test. As discussed in IGA News Spring 2019, the IGA is currently funding final stage trials of using the Eyecatcher in home monitoring of visual fields
- Contrast sensitivity - performed using a mobile phone
As most IGA members will know, visual field examinations are a central part of diagnosing and monitoring glaucoma. The gold standard test uses the large and not-at-all-portable Humphrey Field Analyser (HFA), which costs around £20,000 - 25,000. But as Dr Hu told us “It also needs to be maintained, so the service agreement would often cost a lot, and in a place like Tanzania getting somebody to come out to service the machine or fix it might take months. And then once you have it of course you need a trained technician to run it. It’s an automated test but there are personnel costs involved too.”
The Kilimanjaro Christian Medical Centre (KCMC) in Moshi, Tanzania is one of the study sites. “KCMC is unusual for that region in that it does have a functioning Humphrey visual field machine but I don’t think there are many others, certainly not in government-run services.”
Sia Mbesi, who will work as research assistant for the next two years of the study, seen here in the eye clinic.
Dr Hu chose the KCMC as one of the trial sites as he did his PhD in Tanzania between 2008 and 2011. “That was on trachoma, caused by chlamydial infection in children. It’s spread by close contact; they get scarring of the eyelids and eventually as adults they get in-turned eyelids, and it’s quite a common cause of blindness in that part of the world. I know the hospital well so for me it’s a practical and useful place to do research.” The other test site is the Leighton Hospital in Cheshire, where Dr Hu is a consultant ophthalmologist.
We asked how the experience of being diagnosed and treated might differ in Tanzania, compared to the UK. “Patients tend to present a lot later in Africa - only when they’ve noticed a problem - so they often have either end-stage disease or a significant visual field defect. By contrast, in the UK a lot of patients get referred by the optometrist: they might have borderline high pressures or they might not have done the visual field test properly, so a lot of those will be false negatives or have very early disease. So the spectrum of disease is very different in Tanzania.”
Dr Jones added “We looked at all the data that they’ve been collecting in Moshi and the striking thing is there’s huge amounts of severe sight loss in relatively young people: working age people, and even children too. These are often people who are supporting entire family networks and are going to be living with their vision loss for 40-50 years. That’s really scary, and also unnecessary when you think that we have the means to detect glaucoma far earlier. The current levels of late presentation are not surprising, but they are shocking.”
Around 400 study participants will be recruited over the two sites, including a range from advanced glaucoma to healthy eyes. The three tests will be performed in addition to patient’s usual tests. Dr Hu explained the process:
“The patient will come in for their normal glaucoma clinic and they’ll be approached while they’re waiting for their appointment. If they’re happy, they’ll be recruited to the trial, consent will be taken and then they’ll be tested straight away. They’ll have the Eyecatcher visual fields done first in a side room where it’s a bit quieter. Contrast sensitivity will then be measured, which again is a very quick electronic test, maybe a couple of minutes, and then the fundus photograph. Ideally we’ll ask the patient to repeat each test because we want to get an idea of comparability, but all of the tests are very quick so we’re hoping patients won’t mind doing that. They’re all completely non-invasive tests.”
So how will the researchers know if the tests work effectively?...
“The Eyecatcher visual field test is currently being validated in clinical and laboratory trials. The second test is digital photography of the optic disc and this study will use expert graders to grade the optic discs for glaucomatous damage. But the team anticipate that in the future machine learning will help to automate this process. Artificial intelligence-based automated software for detecting diabetic retinopathy from retinal images has reached an advanced stage, and similar work has begun on glaucoma detection.”
…. and how will they develop an algorithm for detecting glaucoma?
“It’s quite straightforward really. You’ll have a group of people whom expert clinician says have glaucoma, and you’ve got a group who the clinician decides are healthy (‘false positive referrals’). You then train the algorithm to replicate the clinician’s decisions by statistically combining all the available data. In doing so, the computer might discover that it’s better to base its decisions more on some information (for example the fundus image and the visual field test) than on others (for example contrast sensitivity measurement). The algorithm might even determine that some tests aren’t required at all, in which case we could dispense with them in future. Whatever the outcome though, you’ll have an objective way to synthesise all the clinical data into an instantaneous decision: in this case, whether or not to flag the patient as ‘high risk’.”
The trial is due to start in September and will run for two and a half years. “I’m in the process of getting though all the ethical permissions which is quite complex. In the UK we’ve learnt that we need permission from the Medical Healthcare Regulations Authority. This was quite a surprise because our previous studies using these technologies didn’t need to go through MHRA. But now the regulator’s opinion is that because there’s some software that interacts with the patient, all our tests need to be treated as ‘medical devices’ and so need to be regulated accordingly. So that’s another quite time-consuming and expensive step unfortunately. In Tanzania we also need permission from KCMC hospital and from the National Institute of Medical Research. So it’s quite complex and time consuming. Then I aim to start recruiting in January, for one year in Crewe and for 18 months to two years in Tanzania.”
The new tests obviously have the potential to revolutionise glaucoma detection in low-income regions and save the sight of many people. But one day, could they replace the tests used in our own eye clinics?
“Whether they could work in a clinic to replace other tests – maybe, but that certainly isn’t our immediate aim. There’s a huge number of people that never set foot in a clinic.
Many adults don’t go for regular eye tests. And some vulnerable groups simply can’t attend traditional clinics, such as stroke patients and children with special needs. Similarly, you’ve got increasing numbers of cases where very elderly people have a spectrum of disorders, and it’s often not suitable for them to attend regular glaucoma clinics.
“I think at optometry level it could definitely come in and I could see it working in primary care for GPs with an interest in glaucoma, and you’re going to save the taxpayer money by having fewer false referrals. If it works in the rural outskirts of Moshi, then we could also give it to people to take home. Then you can be testing more often, getting more accurate data, fewer trips into the hospital; there’s lots of different benefits that come from the same technology.”
Dr Hu thinks that eventually all three tests could be carried out by one single piece of portable equipment. Dr Jones adds “If there were the willpower and the money you could get all those three running on a mobile phone today. It is possible, and I very much hope it will happen. It’s just about getting the right people together in the same room, with the money, the resources and the will to make a difference.”
Tapping into the 5,000 hours – how feasible is home monitoring of visual fields
Our £25,000 2018 optometry research grant was awarded to Dr Tamsin Callaghan, a dispensing optician and optometrist prior to completing her PhD at Cardiff University, and now a lecturer at City, University of London. Tamsin’s research is entitled Tapping into the 5,000 hours – how feasible is home monitoring of visual fields
“People with glaucoma and those at risk of glaucoma require lifelong hospital review to monitor their condition. This includes regular visual field (VF) tests in outpatient clinics which is time-consuming, expensive and can be unpopular with patients.
“More effective management of these patients may delay disease progression and reduce the economic burden of this disease. However this is challenging within the current system and alternatives to monitoring need investigation.
“Patients spend only a few hours each year in eye clinics but spend more than 5,000 waking hours each year doing everything else. We wondered whether glaucoma patients could self-monitor their visual field at home between clinic visits, and it’s this question that is explored in this study, the first to investigate the monitoring of VFs in the home.
“Home monitoring is increasingly popular in healthcare, and it’s already commonly used for blood pressure and diabetes monitoring. Advances in portable technology allow vision testing in unconventional settings and our new method of assessing VFs - the Eyecatcher visual field test - uses an inexpensive tablet device to allow people to carry out VFs in their own homes. It is the first home-monitoring programme to use eye and head movement, designed to assess and encourage steady eye positioning when patients are not being supervised in a hospital or research setting.
“20 people with moderate glaucoma in at least one eye will be first assessed at City, University of London, using the standard visual field test: the Humphrey Field Analyser test. All participants will receive instructions on how to use the Eyecatcher test and half of the group will also have an additional training session. All participants will then be issued with the Eyecatcher tablet and asked to monitor their VFs on both eyes at home once a month for six months.
“Using the front facing camera of the device, participants’ faces will be photographed at set times during each field test, with the data used to assess the Eyecatcher’s reliability and participants’ movements.
“When home-monitoring is finished, participants will return for a final assessment at City University, and carry out a final Humphrey Field Analyser test. The Eyecatcher VF results will be compared to the Humphrey results.”
We asked what level of technical ability the participants would need in order to take part – would the trial be limited to those people who are confident using computers and tablets? “Hopefully no, although it’ll be interesting to test this. It’s very easy; it’s a case of turning it on and pressing a button then following the instructions. So you don’t need a huge degree of digital literacy to do it, but we understand this may be a perceived barrier, and it’ll be interesting to see if that’s the case. Most of our patients are increasingly using tablets and iPads at home so they are maybe more used to the technology than we realise.
“This trial will use just one type of device and participants will be given a simple tablet to do the tests on, but the eventual aim is to develop a model that can be used on many different types of device, the sort of thing patients might have at home.”
We know from IGA member feedback that many of you have struggled taking VF tests in optician and hospital premises because of distractions like noise and interruption, and we wondered whether the trial had ways of accounting for any unexpected variables in peoples’ home environment that might affect the results. “If for example someone didn’t perform the test well because they were interrupted by the phone ringing, the extra data and photos provided by the Eyecatcher should help the researchers understand that. And because we’re doing a series of tests we can average the results, so if there are four good measures and two not so good this would be interesting too. Patients normally take just one Humphrey test a year, and there’s always chance factors like tiredness and background distractions could affect that annual test."
We asked whether there was a chance people might become ‘hooked’ and do the test more often, introducing a learning effect which could improve performance? “A previous literature review of research on home testing in diabetes and blood pressure didn’t reveal any information to suggest this happened, and discussions with glaucoma patients led researchers to think this wouldn’t be a major factor, but they appreciate it is a possibility. Feedback will be collected from participants on their home-monitoring experience to identify any difficulties, potential improvements, and to compare the home VF test experience with the Humphrey. The effect of the training support package on participants’ compliance will be studied and all feedback will be analysed to identify any trends or themes.”
If home VF monitoring for glaucoma is feasible, this could have significant, far-reaching impact for those with glaucoma and clinicians. Hospital glaucoma clinics would benefit from shorter eye clinic appointments, reducing costs and waiting lists and allowing at-risk patients to be seen more quickly. Most importantly, people with glaucoma would benefit from increased VF monitoring, potentially detecting VF progression at an earlier stage.
Remedio hand-held fundus camera
Fundus photography involves photographing the back of the eye, also known as the fundus. The camera consists of a microscope attached to a flash enabled camera and is used to take pictures of the optic disc, central and peripheral retina, and macula.