NLP’s role in linking social determinants to heart disease
Even with significant medical advances and greater access to healthcare, heart disease still remains the leading cause of death globally. While this is cause for concern alone, the risks of heart disease and other chronic illnesses have been amplified due to the impact of the COVID-19 and its links to harmful effects on the cardiovascular system.
While conventional knowledge has led us to believe eating healthfully and adopting an active lifestyle can help safeguard against the propensity for conditions than lead to chronic illness, predicting and preventing disease isn’t quite so simple. These behaviours certainly contribute to overall health, but it’s only a piece of the puzzle. In order to get the full picture of a patients’ health, we need to start putting a bigger emphasis on both clinical factors coupled with a wide variety of social determinants to give them the best chance at a healthy, disease-free life.
Social determinants are classified as elements that directly impact a person’s health beyond diseases or drugs, such as access to healthy food, personal safety, housing, employment, literacy, family, employment, and personal freedom. These are often more important than clinical treatment when it comes to managing conditions like heart disease. But without knowing more than the results of a diagnostic test or one’s medical history, it’s impossible to understand the full story of how the patient got to this point and what could impact their ability to get better.
NLP and social determinants
It’s hard to argue against social determinants’ influence on health, but finding the connective tissue between a person’s life and the hard clinical data available to physicians doesn’t come without challenges. Social determinants can often only be read from free-text notes in a healthcare setting—not in structured data. Essentially, doctors will manually write out details of a patient’s social history, home environment, and similar types of health contributors. Structured data in electronic medical records (EMRs) only consists of lab results, billing codes, and what medications the patient is taking.
If a patient is experiencing substance abuse, unemployment, homelessness, or illiteracy, those will be in the free-text notes. As you can imagine, linking hand-written notes to medical records is not only challenging, but time-consuming. In order for medical professionals to realistically compile and use this information, they need technology to automate the process, and accurately. This is where natural language processing (NLP) comes in. NLP has the power to connect the dots between these disparate and siloed data sources to understand how these health events are related.
But merging free-text and structured data isn’t the only challenge. Sometimes medical professionals simply don’t know what they’re looking for. For example, let’s say heart disease patients who take a daily multivitamin and exercise regularly show alleviated symptoms. But how do researchers test what other behaviours could potentially improve health outcomes of this population?
NLP is the only viable way to correlate all potential variables—sleep, relationships, safety, employment, obesity, etc—to shed light on this in an effective and timely manner. Not to mention, important information also lives in diagnostic imaging reports, social media, and other modalities. You need software to connect the relationships between everything.
While cardiology is a field well-known for using data-centric governance models, data quality also needs to be a consideration when exploring social determinants of health, and data integration still presents another big challenge. In large research projects where information is collected from different entry points and data is available in different formats, it’s common for pertinent information to be missing or inaccurate. Once again, NLP is an excellent source for researchers working in the cardiology field to mitigate this issue. With existing datasets in this speciality, researchers and data scientists can more easily connect the dots.
Millions of people worldwide die of heart disease every year. While the outlook sounds grim, most cases of heart disease are preventable with lifestyle changes. That said, someone who is being physically abused or has a long history of drug addiction may not be in a place to prioritise eating more vegetables or adding a 30-minute workout to their routine.
Taking social determinants of health into account is often undercounted, but vitally important for patients’ overall health outcomes. Fortunately, technology like NLP has made it easier to start correlating social determinants to heart health, and has the potential to vastly improve prevention and treatment in the future.
How UiPath robots are helping with the NHS backlog
The COVID-19 pandemic has caused many hospitals to have logistical nightmares, as backlogs of surgeries built up as a result of cancellations. The BMJ has estimated it will take the UK's National Health Service (NHS) a year and a half to recover.
However software robots can help, by automating computer-based processes such as replenishing inventory, managing patient bookings, and digitising patient files. Mark O’Connor, Public Sector Director for Ireland at UiPath, tells us how they deployed robots at Mater Hospital in Dublin, saving clinicians valuable time.
When Did Mater Hospital implement the software robots - was it specifically to address the challenges of the pandemic?
The need for automation at Mater Hospital pre-existed the pandemic but it was the onset of COVID-19 that got the team to turn to the technology and start introducing software robots into the workflow of doctors and nurses.
The pandemic placed an increased administrative strain on the Infection Prevention and Control (IPC) department at Mater Hospital in Dublin. To combat the problem and ensure that nurses could spend more time with their patients and less time on admin, the IPC deployed its first software robots in March 2020.
The IPC at Mater plans to continue using robots to manage data around drug resistant microbes such as MRSA once the COVID-19 crisis subsides.
What tasks do they perform?
In the IPC at Mater Hospital, software robots have taken the task of reporting COVID-19 test results. Pre-automation, the process created during the 2003 SARS outbreak required a clinician to log into the laboratory system, extract a disease code and then manually enter the results into a data platform. This was hugely time consuming, taking up to three hours of a nurse’s day.
UiPath software robots are now responsible for this task. They process the data in a fraction of the time, distributing patient results in minutes and consequently freeing up to 18 hours of each IPC nurse’s time each week, and up to 936 hours over the course of a year. As a result, the healthcare professionals can spend more time caring for their patients and less time on repetitive tasks and admin work.
Is there any possibility of error with software robots, compared to humans?
By nature, humans are prone to make mistakes, especially when working under pressure, under strict deadlines and while handling a large volume of data while performing repetitive tasks.
Once taught the process, software robots, on the other hand, will follow the same steps every time without the risk of the inevitable human error. Simply speaking, robots can perform data-intensive tasks more quickly and accurately than humans can.
Which members of staff benefit the most, and what can they do with the time saved?
In the case of Mater Hospital, the IPC unit has adopted a robot for every nurse approach. This means that every nurse in the department has access to a robot to help reduce the burden of their admin work. Rather than spending time entering test results, they can focus on the work that requires their human ingenuity, empathy and skill – taking care of their patients.
In other sectors, the story is no different. Every job will have some repetitive nature to it. Whether that be a finance department processing thousands of invoices a day or simply having to send one daily email. If a task is repetitive and data-intensive, the chances are that a software robot can help. Just like with the nurses in the IPC, these employees can then focus on handling exceptions and on work that requires decision making or creativity - the work that people enjoy doing.
How can software robots most benefit healthcare providers both during a pandemic and beyond?
When the COVID-19 outbreak hit, software robots were deployed to lessen the administrative strain healthcare professionals were facing and give them more time to care for an increased number of patients. With hospitals around the world at capacity, every moment with a patient counted.
Now, the NHS and other healthcare providers face a huge backlog of routine surgeries and procedures following cancellations during the pandemic. In the UK alone, 5 million people are waiting for treatment and it’s estimated that this could cause 6,400 excess deaths by the end of next year if the problem isn’t rectified.
Many healthcare organisations have now acquired the skills needed to deploy automation, therefore it will be easier for them to build more robots to respond to the backlog going forwards. Software robots that had been processing registrations at COVID test sites, for example, could now be taught how to schedule procedures, process patient details or even manage procurement and recruitment to help streamline the processes associated with the backlog. The possibilities are vast.
The technology, however, should not be considered a short-term, tactical and reactive solution that can be deployed in times of crisis. Automation has the power to solve systematic problems that healthcare providers face year-round. Hospital managers should consider the wider challenge of dealing with endless repetitive work that saps the energy of professionals and turns attention away from patient care and discuss how investing in a long-term automation project could help alleviate these issues.
How widely adopted is this technology in healthcare at the moment?
Automation was being used in healthcare around the world before the pandemic, but the COVID-19 outbreak has certainly accelerated the trend.
Automation’s reach is wide. From the NHS Shared Business Service in the UK to the Cleveland Clinic in the US and healthcare organisations in the likes of Norway, India and Canada, we see a huge range of healthcare providers deploying automation technology.
Many healthcare providers, however, are still in the early stages of their journeys or are just discovering automation’s potential because of the pandemic. I expect to see the deployment of software robots in healthcare grow over the coming years as its benefits continue to be realised globally.
How do you see this technology evolving in the future?
If one thing is certain, it’s that the technology will continue to evolve and grow over time – and I believe there will come a point in time when all processes that can be automated, will be automated. This is known as the fully automated enterprise.
By joining all automation projects into one enterprise-wide effort, the healthcare industry can tap into the full benefits of the technology. This will involve software robots becoming increasingly intelligent in order to reach and improve more processes. Integrating the capabilities of Artificial Intelligence and Machine Learning into automation, for example, will allow providers to reach non-rule-based processes too.
We are already seeing steps towards this being taken by NHS Shared Business Service, for example. The organisation, which provides non-clinical services to around two-thirds of all NHS provider trusts and every clinical commissioning organisation in the UK, is working to create an entire eco-system of robots. It believes that no automation should be looked at in isolation, but rather the technology should stretch across departments and functions. As such, inefficiencies in the care pathway can be significantly reduced, saving healthcare providers a substantial amount of time and money.