Interview with Larry Benjamin

Larry Benjamin was born in 1956, in Johannesburg.  His father, Alan Benjamin OBE, was British and his mother was South African. He explains how his father came to live in South Africa: “My father’s mother was quite ill and was sent to South Africa to warmer climes for her health, so my father and his father emigrated there to look after her.  Unfortunately she died before they got there, but he spent his formative years at university there, met my mother, had children, and then brought us back after the Sharpeville shootings in 1961, which was a time of political unrest.  I’ve been here ever since.”

Upon returning to the UK, the family lived with Larry’s grandparents before buying a house near Wembley, London.

 Larry says of his father: “He was always very interested in the application of technology. He probably was the most impractical man I’ve known, he couldn’t wire a plug successfully, but he was very interested in how other people could apply technology and the use it was put to. That was always an inspiration to me really.  He was always teasing and cajoling about how doctors would be replaced by robots and people would be walking around with mobile phones held to their ears in the future, which we all laughed at. He was quite prescient in a lot of his thinking and always encouraging about what was possible with IT, even though in the early days when we came back to the UK he studied at night school for a few weeks and learnt machine code.  He tells me he wrote the first program that marked the multiple choice in A level scripts; so I blame him for not getting straight As at A level. He was practical in computing for a short while, but essentially was an organiser using his great skills with people and organising things.”

Larry’s mother did some accounting and secretarial work but was mostly a stay-at-home mum looking after Larry his older sister and two younger brothers. Larry adds: “She didn’t go to university.  She married my father quite early on and looked after us kids really. We got a lot of home life skills from her.”

 Larry attended school in Kingsbury, formerly a grammar school which became a comprehensive school but retained many of its former grammar school teachers.

It was in his upper sixth year that the school brought in a base-8 computer which was travelling around the different schools in the area. Larry says: “It had a single row of digital readout across the top and we were just shown how it worked.  We were also sent to Middlesex Polytechnic to do some basic programming, and we did a bit of BASIC on a computer there.  It wasn’t until university that the Apple I computer and the Commodores came out. My brother had a Sinclair Spectrum, or a ZX81, and then I had a Sinclair Spectrum, with tape drives and various things.  So I had fairly early exposure to that technology, but not at school, at school it was all slide rules and log tables and that sort of thing.”

 Larry also had a visit to his Dad’s company called Systems Programming Ltd, (SPL). He says: “I remember seeing one of his early computers, which I seem to remember was the size of a three-bed semi house and used punch tapes.  I remember my mum lining up punch cards with knitting needles at home.  There was a lot of that sort of early technology around and I saw some of the computing which was lots of flashing lights and spinning tapes and things, but you wouldn’t let kids play with it in those days.”

Asked if he was aware of his father’s influence in IT, Larry says: “At times. I think he was the first director-general of the Computing Services Association. I knew he was fairly up on things at that point, knew a lot of people.  I don’t suppose I really got a true flavour of how well known he was until much later on. I remember when he brought home the first Apple, it was the pre-Macintosh one, the Lisa. I had a play with that and I tried to get the dot matrix printer working, but couldn’t quite.  So I was exposed to all that sort of chronic low-dose IT in all its forms.”

After school Larry went to the University College London to study medicine. He was inspired by his uncle who was a paediatrician, he adds: “He was very inspirational and  I couldn’t think of a lot else to do, really.  I very much enjoyed biology and chemistry, or zoology and chemistry I did at A level, and it just seemed the right thing to do. I don’t come from a long line of doctors in the family, so I didn’t know much about it, really, but it just seemed to cover all the bases, there was a lot of science involved.”

His exposure to IT continued at medical school where he was introduced to a programme written by one of the physiology professors about the respiratory system. He says: “It was all numbers, you put numbers in and you could simulate an overweight man going up in an aeroplane, saw what happened to his blood oxygen levels, and then crash diving to below sea level to see what happened to the oxygen levels.  All this was by typing numbers in.  I also wrote some early BASIC programs for the Siggaard Andersen nomogram, which was to do with bicarbonate and various other things in the blood in the ITU system. They had an Apple computer and I used to mess about on that and write small BASIC programs; they were very basic. I enjoyed that sort of chronic low-dose exposure to IT and could see the potential for applications in the future, but using computers wasn’t routine in medicine until much more recently.”

Asked if the impact of IT has been less than expected, Larry says: “The use of IT by clinicians has probably been less until fairly recently.  We’re getting better at it and there are good systems around now. I remember the early EMI scanner, as it was known at Northwick Park Hospital. We had one when I was a houseman there in 1980, it had two pixels, a black one and a white one, and it was very crude indeed. If you compare that with the spiral CT scanners of today that can do the same scan in a fraction of the time and give you fantastic detail with a fraction of the radiation, you just see that incremental change in technology all the time. That’s where it’s been useful and has actually helped medicine tremendously.  Another area is imaging systems; in ophthalmology we use a lot of imaging of the eye and other parts of the body, and they’ve changed dramatically.  There’s a thing called an OCT scanner, which is a machine which sends light into the eye and looks at the reflection of it off the retina, and that’s transformed our management of ocular conditions in the last fifteen years or so.  The traditional way to look at the back of the eye was to photograph it with a big camera.

My other passion in life is photography, so I was always into these imaging systems for the eye.  There’s now a fabulous wide angle retinal system.  I worked on a very early possibility of photographing the back of the eye with a wide-angle lens, very crude but quite useful, and there’s now a machine called Optos, which does it routinely and gives you a magnificent wide-angle view of the back of the whole eye, and it’s pretty much routine in all departments now.  But the OCT scan has changed the way we look at the retina in cross-section, so you can actually tell that something is now swollen, whereas previously you were looking at it en face, you couldn’t see whether it was swollen.  So it’s just transformed the way we manage some diseases, and indeed, Moorfields Eye Hospital in London’s recently done a project with Google DeepMind in King’s Cross where they’ve got a double neural network to analyse scans of these OCT scans of the eye, millions of them from patients in Moorfields, and they’ve now got the system to the level where the machine is as capable as a consultant with twenty years’ experience in getting the diagnosis right first time.  Routine scanning of the back of the eye can now be pretty much done by machine.  Those sort of things are coming along all the time and I think, although it’s not a big step change necessarily, it’s just gradual evolution and application of technologies that’s been so wonderful.”

 Asked if there was any resistance from clinicians, Larry adds: “There was absolutely, particularly for electronic data collection. We’re all used to doctors with bad handwriting and trying to analyse what they say, that’s almost a thing of the past now.” He continues to explain that rather than type, he now uses a voice recognition system which he feels is the way forward. He continues: “The use of big data to analyse outcomes in surgery is now becoming possible in a variety of specialties.  Most doctors, certainly the youngsters coming through, are absolutely au fait with the technology.  … We’re getting to the place in medicine where people just accepted it’s going to happen.

“The biggest difficulty for me has been seeing the poor implementation in some trusts and the resistance to implementation of IT systems and the very traditional ignoring the end user. Someone will write a system that looks very good superficially, but when you try and apply it to what it’s supposed to be applied to, it slows the user down because it’s the computer making the person do what it wants rather than the person using the computer to do what they want.  But I think we’re getting better even at that, I think there are now bespoke systems that are so specialised that it is making data collection much easier.”

On the subject of interconnectivity and being able to see different patient data across the treatment range, Larry adds: “Richard Granger tried to solve it with twelve billion pounds’ worth of our money and unfortunately failed with a single system.  The NHS Spine is useful for some things, you can look up most people’s allergies, but not everybody wants to be on it and so it’s not universal.  If you go to the optician to get an eye test, their systems are not allowed to be on the N3 network at the moment, so we can’t get access to their data.  They’re taking part in follow-up of post-op cataract patients, but we can’t get access to their data, so we don’t know what our outcomes are truly like. That’s being addressed slowly and I think people have realised that actually, interconnectivity is where it’s at.  GDPR and data protection is to blame for a lot of this, unfortunately, I think the rules are so strict that it’s very hard to convince people to let go of their data from one hospital to another or from a GP to a doctor, it’s incredibly hard.

“The rules have to be strict enough to protect people’s private data, and that’s really important. From the point of view of getting insurance, or getting a job, there are some data you might not want to be out there, but I do think that getting the right people to access the right information at the right time should be possible nowadays with fingerprint technology or scans of some sort of identification; it should be possible, it’s just very slow.  If I want a GP record on a patient, it’s virtually impossible for me to get hold of it. I have to write to the doctor and ask for specific enquiries, I can’t just get the record, and I think that’s silly, in a way.  It does make healthcare very un-joined up.”

On the subject of technology in healthcare around the world, Larry observes that many third world or developing countries have some up-to-date technology. He says: “Laser refractive surgery for curing or treating long and short sight is pretty universal round the world, most countries seem to have access to it somehow, for private patients mainly.  Lots of places are becoming much more adept at acquiring technology.  Orbis, who I work with, part of their mission is to get people up to date with technology for ophthalmology. We’re trying to install up-to-date machines, make sure they’re maintained, because part of the problem in the developing world is that if there’s not a maintenance technician or an engineer to look after the gear it just sits there with a blown fuse and nobody can use it.

“Orbis tries to apply the money that they raise in such a way that they can implement technological change, but also teaching and training and maintenance of systems and equipment.  One of my great missions with Orbis is to try and implement a data-driven electronic patient record and we’ve now got one on board the aeroplane, which is very good, but we don’t yet implement it in each of our projects around the world.  So it’s all paper based, it all relies on people retrieving data and returning it centrally and analysing it by hand.  There’s a lot of room for manoeuvre in the developing world and we’re trying hard with Orbis to implement some of that, but it’s a big problem.”

He highlights the potential of telephone technology which is prevalent across the globe as one valuable option towards bringing up to date technology in healthcare to some countries, adding: “We have a project where a colleague of mine has developed a system where the telephone could take a picture of the retina for diabetic retinopathy, you can GPS locate that patient because the phone is with them at the time and you know where to go back to next year to find him or her again.”

Larry has been recognised with a number of awards for his work, including the Kodak Photographic Bursary for investigating retinal photography, plus the . 

Oxford Region Support Grant for Developing of Computerised Strabismus. Larry explains this latter project which targeted correcting squints in children.

He says: “A lot of children are born without the ability to use both eyes together, for one reason or another, so they don’t have stereo vision.  Socially it can be quite a barrier to development if the eyes are not looking in the right direction.  Some of what I used to do with Orbis was straightening squints.  When I worked in Oxford we had a craniofacial unit which dealt with children with inherited disease to do with the formation of the skull and the bones of the skull whereby the actual orbits of the eyes were rotated abnormally and so they had very peculiar eye movements.

“I had the idea of trying to computerise these eye movements to analyse what was going on and I enlisted the help of a nuclear physicist who was working at the Joint European Torus Project in Oxford. He was trying to learn C++ at the time, and he offered to write this program in C++, initially free, and then eventually we got some funding and we paid him to do it properly.

“Now, we have a model which runs on a PC which draws two eyes on the screen with all six muscles around each eye and you can move the eyes in any direction to see what’s happening in each of the muscles, and then you can operate on each muscle by moving it appropriately and see what happens to the position of the eyes.  You can also rotate the orbits to see what’s happening in these children to produce the strange eye movements. That became a clinically available product, we gave it away free.  It’s  about fifteen years old, I haven’t updated it, but I was thinking about trying to get some more funding to get it updated, to give it to partners with Orbis around the world, because it’s a very useful learning tool for surgeons trying to anticipate what to do in these operations.”

Larry says that lasers in ophthalmology are fairly ubiquitous and are used for treating various parts of the eye for such conditions as diabetes, glaucoma, and short-sightedness. Larry explains: “Most people have heard of laser refractive surgery which means having your short-sightedness treated by a laser that reshapes the front of the cornea.  The cornea is the transparent film on the very front of the eye that does about two-thirds of the light bending for the eye, and if you change the shape of it you can change how the eye focusses. That’s a very popular technique, although it is not available on the NHS because it’s considered to be a cosmetic procedure, but the technology’s brilliant and is getting better and better.”

Larry joined the WCIT, which his father founded, because of his interest in applying IT to medicine. He says of his reason for joining: “I thought that I could somehow be useful on the health panel.  One of the initial things we did was to try and give trusts advice about implementing IT systems, because that was very haphazard in the early days. There were all sorts of systems put in which either didn’t work or didn’t work well together, they certainly didn’t work between trusts.”

Another reason was to try to counter what Larry sees as short-sightedness when it comes to some decisions regarding technology and it’s use. He explains: “I helped implement a clinical database quite early on for our diabetic population. Through it, we collected a lot of data for every patient that came to the hospital covering fourteen years.” He goes on to say that through annual analysis, the hospital was able to predict how many more clinics were needed, how many patients they had, and what future demand for services might be. He continues: “One day one of the managers in the trust told me that they were going to switch off that system because it was considered too old and they couldn’t afford to keep it running. I asked for a server to keep it on because it was valuable data. However, that was considered too much money and the system was switched off, so we now can no longer analyse our diabetic population.  That sort of short-sightedness really made me cross, and it was to try and counter that sort of thing that I thought I’d get involved with WCIT.”

Larry says of his management style, that, like his father, he is a people person: “My kids say that when I get cross I go quiet.  I’m not a table thumper and I don’t shout. I think working with people is the critical thing, you’ve got to know your stuff, you’ve got to understand them and their problems and try and help them. Being a people person I think is really important.  You can’t help everybody with everything, but I think if you take an interest in people and they’re part of the team, that’s very powerful.”

While Larry says that data collection is now being done well, he would like to see it done for all procedures in the future. He adds: “I’d like to see that for all procedures so that we can keep standards up and it’s good for informing training and teaching. Funding is also an important area. We need funding for keeping up with the technology.  Having the technology at your disposal is so useful and so vital now in managing people’s conditions and expectations and the application of that technology is very important, trying to maintain those modern standards. But the most pressing thing is that systems to collect data are still a bit of a weak spot and we need to get that a bit more national and routine.

“I think part of it has been resistance to implementing IT and part of it is cost. We’re getting better now at bespoke systems. During the era of Richard Granger, I remember writing a letter which was co-signed by a number of eminent people, about the implementation of bespoke systems and how that we all knew that collecting data at a local level by specialists with specialist knowledge was the important thing and that from the IT perspective, interoperability was what it was all about.  It didn’t matter that you had little systems collecting data all over the place, you had to make them talk to each other.  The concept of one system to do everything was never going to work and indeed didn’t.  I still think this interoperability is something we need to fund much more of, to give people the systems that work for them and allow them to connect to each other. … All of that will come with time, the technology changes and improves and it’ll happen.”

I don’t know that I’ve personally made a big mistake. I’ve been very frustrated in terms of not being able to get funding to do what I’ve wanted to do.  I suppose that’s been my error; not fighting hard enough for the funding.

“How we raise funds and spend the money is often, to me, not very transparent. I slightly regretted not having better business skills, that was one of my weaknesses, knowing how to make those sort of changes in a way that’s effective and I wish I was better at that.

“One of the lessons I’ve learned over the years is to work with management rather than against them, and when you get a good manager to make use of them.  One of the other regrets I had in my career was I thought of trying to implement a course called ‘Medicine for Managers’.  There’s an awful lot of managing for medicine, we all go on courses of how to be managers, but none of the managers learn about medicine. If you could have a three-month residential course with managers sitting in clinic and doing on-call, staying overnight, and talking to Mrs Smith and explaining why she can’t have that drug because it’s too expensive, they would understand a lot better what the pressures are from our end and I think it would make the team approach much more realistic.  I did talk to someone at one of the big teaching units near us and he said it was a good idea, but that’s as far as it got.”

Having retired in 2019, Larry returned to work on the cataract backlog caused by COVID 19. Asked how he found the NHS upon his return, he says that in some ways it was an improvement. He says: “For years we’d been saying, can we work from home please, can I access my emails from home and being told that it can’t be done because security was too much of a problem. When COVID came in, working from home happened overnight.  So there have been huge improvements in that respect, things just got done.  We do a lot of telephone follow-ups now.  We’ve implemented a robotic system for patient follow-up for cataract surgery, which is well received, saves a lot of nursing time and patient travel time; so real world costs are saved. So, I think there have been huge improvements from COVID which have been good to see and which we’re trying to keep going, but at the same time, you can’t really do doctoring without being face to face for some of the time and we’ve been a bit slow to re-implement some of those clinics.”

Having completed his clinical work during COVID 19, Larry has retired again but teaches and runs training courses at the Royal College of Ophthalmologists in London. He is still working Orbis and will be doing a simulation training programme on the Flying Eye Hospital in Vietnam in May in 2023. He explains: “We can teach and train on the aeroplane, but also broadcast the teaching and training around the world with various simulation techniques. I hope to carry on with that as long as they’ll have me.”

 Speaking about what he’d like Orbis to do next, he adds: “The implementation of IT is critical.  If we could gather data from around the world in developing countries, that would be a massive boost to our understanding and treatment of those populations.  It’s not until you start looking at what you’re doing properly that you can understand the outcomes, and if we could implement that around the world with Orbis in terms of data collection, we would have a repository of data that was unparalleled.  It would be good for drug companies, for learning, for teaching, and for implementing change.  That’s got to be our next big project; carrying on doing what we’re doing in terms of teaching and training people to treat eye disease and prevent and treat blindness, but on top of that making it more refined and give us access to the huge amount of data that we generate each year.”

On the subject of AI, Larry says: “I’ve always argued that it’s not really artificial intelligence if people are programming these machines to do what they want, it’s not really the machine making the decision, it’s the machine doing what it’s asked to do.  Whether we get to a true sentient machine in some time in the future remains to be seen.  It’s probably not impossible, whether it’s desirable or not, I don’t know.

“I’m not sure we want there to be a sentient machine, but I do see the incredible application of properly programmed machines and I think to try and understand how a brain thinks is something we’re nowhere near at the moment, and to try and emulate that with electronics is even further away, but there’s no question that the power of the fast calculation in a computer is what we want to harness, and that’s going to be incredibly helpful.

“The project with Google DeepMind analysing scans is using AI and has been fabulous.  So its application to routine analysis of data which you don’t want a human to have to sit and do is going to be big.  In terms of algorithms and decision making; it can certainly help people to make better decisions.”

Interviewed by Richard Sharpe

Transcribed by Susan Nicholls

Abstracted by Lynda Feeley