Dr. Naila Siddiqui Kamal, senior gynaecologist and senior lecturer, Imperial School of Medicine London, reviews how AI has tackled the challenges of the pandemic
There can be no doubt we have witnessed an unprecedented uptake of technology-enabled health care models during the last twelve months.
With its impact on human suffering, the COVID-19 pandemic has totally changed the shape of delivery of health care. Technology has taken centre stage amongst the many disruptions that we have witnessed.
Of these, the ones that emerged as the anchor for continuation of delivery of health care are telehealth technology-based models. In the US, telehealth was used by 45% of consumers in the COVID-19 era, compared to 11% of consumers in 2019 that called on this method to replace health care trips to the doctor’s office.
Those who were already on the periphery of adoption of these models, had a golden opportunity to fully immerse their local health care model as a telehealth model while the sceptics had no other option to engage with it and so it’s no surprise that we find that the consensus changing. We still find patchy understanding amongst the wider diaspora about what exactly entails telehealth and what specific skills and competencies are required for safe and effective telehealth practice of medicine.
The technology used in telehealth can be as simple as a text-based model using an android phone to highly sophisticated models using artificial intelligence, augmented reality and holoportation. We also see a number of symptom checkers for COVID-19 infection risk assessment that have been further developed using complex neural networks.
So it’s an apt time to review how AI has powered some of the solutions in tackling the covid challenges. One of the leading examples that I feel really took advantage of AI technology is the case study of The Medical University of South Carolina (MUSC).
This case study defines how artificial intelligence expertise was applied to data coming from MUSC’s virtual urgent care platform (Zipnosis, Minneapolis, MN), which while rich in clinical information, was locked in the “free text” record.
Their researchers developed natural language processing tools to extract symptoms, risk factors, and medication treatments from notes to add to the COVID-19 registry. A team applied deep learning neural networks to notes and developed models to predict which patients would eventually have positive SARS-CoV-2 results.
Deep learning neural network algorithms offered substantial improvements over the rule-based algorithms and subsequently were used to prioritize scheduling of testing. Their telehealth and information solutions systems were assessed for functionality and readiness to serve as data collection and monitoring systems.
The main achievements were not only to remotely monitor patients but also to save thousands of dollars in PPE, which was in short supply, and was saved for instances where they were most needed.
Recent designs of telepresence robots are designed to be able to autonomously move around hallways and rooms by being remote-controlled using a software interface connecting the user to the robot through a Wi-Fi connection. This concept has been recently developed by adding the use of the combination of artificial intelligence and vision systems for navigation and obstacle detection.
Another example is applying AI in systemizing the retrieval and analysis of data which addresses the difficulties encountered in hospital procedures. A study by the American College of Physicians states that doctors spend 50% of their time on EHR and desk work, so they came up with Remedy. This is a system that manages patient records effectively by replacing the manual process of examining the patient’s vitals with a chat interface questionnaire, recording it in storage, and sending the results to the doctor. This vastly speeds up the process, allowing faster sending of prescriptions and other information directly to the patient as they can send photos or videos directly for examination.
So it’s fair to conclude that telehealth and AI have played a pivotal role in combating the challenges from the COVID-19 pandemic and continue to present innovative ways according to the emerging contexts.
But it is essential that our focus remains on the global picture and not just the geographic locations where technology is flourishing in its solutions. The vast populations who have fared worse in the pandemic are in low socio-economic sections of the world. Many charities and NGOs as well as grants from WHO are made available to tackle the sustainable development goals in these populations. The wireless connectivity often is present and only the planning and implementation of a bespoke ecosystem is needed.
I also played a part in engaging in developing and delivering an educational resource called TeleOBGYN virtual health program (www.medret-academy.com) This is a 12 week virtual course aimed at delivering a blended curriculum on health informatics with a focus on telehealth and its integrated models (including AI solutions) and OBGYN topics relevant in the current pandemic crisis for frontline HCAs, particularly those working in poor countries.
As we bear the brunt of another wave of COVID-19 across the globe, training and education in these technologies is proving ever more crucial – enabling clinicians to continue the fightback