Researchers at Lero are discovering how software engineering can help healthcare catch up to the digital age.
Our world is one where healthcare revolutions come in many forms. Penicillin, insulin, vaccination, the discovery of DNA – the list goes on. All of them have shaped the way we conceptualise and treat our bodies, and the way clinicians treat their patients. The field of medicine is rapidly advancing, and yesterday’s science fictions are now in early trials.
What, then, is the role of Lero, the Science Foundation Ireland research centre for software? Hosted by the University of Limerick, Lero’s research covers everything from driverless cars and artificial intelligence to cybersecurity, fintech, gov-tech, smart communities, agtech and health-tech.
But what can software engineering and computer programming offer healthcare? Connection, digitisation and integration are just some of the benefits.
‘I talk about technology supporting, not replacing’
– PROF ITA RICHARDSON
Lero researchers have a plethora of health problems they would like to solve with software. A recent example involved a public health nurse arriving to check on an elderly patient. Of the 15 minutes assigned to the visit, thirteen minutes were spent recording the details required for a new mattress. After taking down this information, later to be transcribed to an online system, the nurse did what she could for the patient in the short time left.
The following day, the nurse was shopping and noticed a shipment of bread being delivered. The barcodes had all of the information needed for the process and were quickly scanned with a handheld device. Why did this similar process take less than 10 seconds, while the health system had no such ability? Why not scan the mattress, and send this information to the system?
For all of our amazing advances, there is a disconnect that hampers our healthcare. Technology is of no help if the user doesn’t have sufficient access. A profile of health problems is futile if it is stored on an inaccessible paper file. A treatment plan for a brain disorder lacks efficacy when it is misplaced or, worse still, never recorded for a patient. A doctor can’t analyse information they never received. And this is all too often the world we live in.
Our healthcare system is handicapped not by a lack of understanding or ability, but by inadequate communication systems that can’t cope with the volume or sophistication that must be accommodated. For Lero, software is the answer.
An ear for surgery
Our world is one where sound is abundant with information. Auditory analysis can map the seabed. The depth of a turkey’s ‘gobble’ might infer how many kilograms it weighs. And, most importantly to mechanical and medical engineer Dr Daniel Riordan, sound can tell a surgeon when to stop scooping bone marrow from a femur.
When a hip is being replaced, the ball of the ball and socket joint is sawn off. Its replacement is put in, and a stem is inserted into the femur to attach it. If the stem is inserted directly into the soft marrow, there will be some wobble and an instability in the hip. As a result, it is necessary to scrape out the marrow until the stem can be secured. If too little marrow is taken out, the hip won’t be secure. Too much, and the bone will crack.
Currently, surgeons performing a hip surgery replacement know when to stop by the sound and feel of the instrument inside the bone. There’s no precise measurement or feedback system. No exact way to know when the surgery is finished. There is just a sense from the tactile and auditory feedback, honed from years of experience.
New medical students practise with fake limbs before levelling up to pig bones (the closest animal analogue) and then onto human cadavers before finally taking the reins in the operating theatre.
Dr John Rice is an orthopaedic surgeon with years of experience in the process, and Lero’s team hope to replicate his ear in software. By recording auditory information as well as the level of success of the surgery, Riordan hopes to replicate this feedback system and empower other surgeons with Rice’s finely tuned experience.
‘The real research is the AI to make this work, to understand every bone and every person’
– DR DANIEL RIORDAN
Capturing Rice’s inherent knowledge of hip surgeries has the potential to save years of surgery training. But giving software this auditory intuition isn’t easy. Hips vary. People vary. Some are taller, some are wider, some are older. All of these factors affect the acoustics of the bone as if it was a musical instrument. Capturing the variation requires as many surgeries as possible. Riordan reckons it would take 100 participants for the software to be effective but the more, the better.
“The real research is the artificial intelligence to make this work, to understand every bone and every person. This project has a duration of four years, and if we find positive results, we would be looking at contacting teams that make surgical robots,” said Riordan.
Paging Dr Google
Our world is one where the knowledge gap between doctor and patient hinders both sides. Patients will try to help a doctor understand their situation in as much detail as they know to provide, but doctors are often left playing medical detective, picking up what they can in an effort to provide cure, care or both.
The more driven patients will take to Google to self-diagnose and speed the process along, as well as advocate for themselves. Practitioners, in response, will fall into one of two camps: those who dismiss the search results as unreliable, and those who parse through the information and evaluate through a medically trained lens.
From a doctor’s perspective, the problem with Google is multi-layered. Its search algorithm is not optimised for accuracy or balance. What’s more, many of the articles will be inaccessible due to medical jargon.
Lero is working to close this patient-doctor information gap. Dr Marco Alfano’s objective is to design an intelligent platform that will empower patients to act as their own advocate. This involves translating existing medical texts so they can be understood, and filtering existing web search results with an algorithm that reflects a patient’s needs, rather than a commercial driving force. It also involves fostering a medical understanding, as well as patient-to-patient relationships, and connecting communities.
As Alfano points out, very few of these needs even require innovation. What’s needed is a coherent design that can be given to patients. The technology is already there, if only it can be integrated.
He and his research team have been working on a prototype website. He types ‘diabetes’ into the custom search engine and the results are ranked on various degrees of accuracy and quality, with the logic of their ranking explained. He then takes this information and feeds it into the team’s text translator. Difficult words are highlighted and a brief explanation of tricky medical terminology is included. Medical files, textbook definitions and doctor’s comments suddenly make sense. He brings up a diagram of the human body, where he can point and click and assign an ailment in an effort to diagnose.
The technology is there. The knowledge is there. And integration could mean diagnosis and treatment is within reach.
Sometimes the practice of healthcare is older than the technologies and techniques that comprise the system. Digital imaging files and results from advanced analytical techniques are often kept on paper. Physical space constraints can mean that the file is all too often out of reach when needed. Patients’ voices, if heard at all, are lost to mountains of paperwork and a system not yet digitised.
Lero’s work, however, is distinctly patient-driven. This philosophy is no more prominent than in the research conducted by Prof Ita Richardson and Dr Jim Buckley.
Richardson works on software processes for groups in need. These have included diabetics, people with mild intellectual disabilities, and the ageing population. By talking to these cohorts alongside their medical professionals, Richardson hopes to build apps that reflect what people need, rather than what software producers think they need. She acknowledges that many patient-generated apps have been amazing, but often fail to comply with EU regulations surrounding data safety and accessibility. And as these accessibility guidelines are more than 100 pages long, it is easy to understand why.
Once data-compliant, patient-led apps have been designed and implemented, Richardson said the benefits are almost immediate. In the case of an app designed to measure a diabetic person’s weight, blood pressure and glucose levels, patient outcomes improved purely through the process of recording data and visualising it in graphs. By seeing how their blood sugars improved after a walk or with a healthier meal, patients implemented their own interventions.
Even the simple act of recording the data is valuable. And Richardson said the goal would be to feed this data back to the doctors to reduce the visits necessary and relieve stress on both patient and doctor contact hours.
“I talk about technology supporting, not replacing,” said Richardson. “Diabetes patients need to meet with their doctor for medical assessment. That’s different than needing to have their blood pressure checked every six weeks. They need to work out medically how often they need to meet.”
Reaching out to the relevant groups, fostering relationships and getting real-life patient feedback is key to what researchers call PPI: public and patient involvement. Through this process, needs can be understood, accounted for and integrated. It can also take researchers down unconventional routes.
Richardson told me of one colleague working with mastectomy patients who joined the Shannon Dragons, a group of Dragon Boat rowers. Rowing is particularly beneficial to those who have undergone mastectomies as the upper body and arm workout is essential to rehabilitation.
‘One of the big battles is adoption’
– DR JIM BUCKLEY
Buckley, on the other hand, is involved in the Covigilant project to evaluate the Irish public’s attitude to the Ireland’s Covid-19 tracking app. He and his team have been seeking out the public’s opinion on different aspects of the app. What works well and what doesn’t? What are their worries? What would they like to see implemented? They have no hypothesis or point to prove – their work simply aims to find out how people genuinely feel about the app on their phone.
The Covid Tracker Ireland app is likely the healthcare app people have become most familiar with. Developed by Waterford company Nearform for the Irish Government’s Health Service Executive (HSE), there is no weightier example of an app coming from the top down.
“One of the big battles is adoption. You want to get people adopting it, and people using it,” said Buckley. “My job isn’t to coerce people to use it, but to evaluate how it is perceived and see best international practice. And to feed that information back to the HSE so that they can consider the insights for future evolutions of the app.”
The first results of the Covigilant study found that 82pc were supportive of the app and intended to download it. The study also revealed fears, largely around data protection and privacy. However, after deployment, user reviews on the Apple and Google Play stores indicated that the public’s perception of the app, in terms of data protection, was favourable.
Lero researchers continued to seek people’s opinions in order to configure the app moving forward, often with interesting insights.
The Covid-19 tracker app isn’t without its flaws. An earlier iteration saw it draining the battery of thousands of Android phones, which was a big blow to adoption. A fix was applied within days, however, and the real strength of integrated health came to the fore.
While old systems can exist for decades with delays and issues, Lero hopes to put patient needs first and fully utilise every opportunity on offer. They hope to revolutionise healthcare using what is already there – the software that can unlock its full potential.
By Sam Cox
Sam Cox was named the science and technology winner in the 2020 National Student Media Awards (Smedias). This award category is sponsored by Science Foundation Ireland and includes a €1,000 bursary to support and encourage up-and-coming science and technology journalism.
The 2021 Smedias are now open for entries. The deadline for applications is 15 April 2021.