Tuesday, 9 August, 2022
HomeTalking PointsRe-imagining medicine’s future with the help of digital technology

Re-imagining medicine’s future with the help of digital technology

Will artificial intelligence (AI) mean the end of some health professionals? Professor Tshilidzi Marwala, vice-chancellor and principal of the University of Johannesburg (UJ), writes in Daily Maverick that AI is merely a useful intelligent tool, just like magnetic resonance imaging. It will not replace radiologists, he says. Instead, it is spawning a new discipline that requires the understanding of both medicine and technology.

Prof Marwala writes:

In his book Deep Medicine, Eric Topol writes: “Eventually, doctors will adopt artificial intelligence and algorithms as their work partners. This levelling of the medical knowledge landscape will ultimately lead to a new premium: to find and train doctors who have the highest level of emotional intelligence.”

Recently, a cohort from UJ went on an exploratory partnership trip to the US. Intriguingly, Topol’s words spoke to the crux of the trip. For some time now, UJ has toyed with the idea of starting a medical school. Yet, we knew this offering had to differ vastly from other medical schools in South Africa.

I have long been an advocate of the use of digital technology in higher education. A few years ago, my colleague Bo Xing and I detailed the changes that higher education institutions needed to adapt to keep up with the Fourth Industrial Revolution (4IR). We wrote: “Higher education in the Fourth Industrial Revolution (HE 4.0) is a complex, dialectical and exciting opportunity which can potentially transform society for the better.”

We have certainly witnessed a hastening of this during the pandemic. Naturally, a new medical school would have to speak to this shift. The medical field is embracing AI with open arms, as are medical schools.

Our 10-day US trip, covering Case Western University, the University of Illinois at Urbana-Champaign and Thomas Jefferson University, indicated that the injection of digital technology into the curriculum is quite remarkable. For instance, manikins that simulate medical conditions are used to ensure students have practical experience from the start. Elsewhere, augmented reality (AR) and virtual reality (VR) provide an immersive experience for students.

The University of Illinois even launched its first AI in medicine certificate programme for hospital workers last year, showing the shift towards reskilling promoted by 4IR.

Much like the US, UJ intends to offer medicine to students who have completed a relevant undergraduate degree: a mix of stackable skills is key for the future of work.

As a university, we already encourage reskilling and upskilling to meet the demands of the 4IR. For example, doctors could benefit from courses in engineering to enable them to use AI technology.

Google clinical informatician and research scientist Martin Seneviratne asked at a conference in 2019 why, given the enormous amount of AI research, doctors are not currently using machine learning.

“Will AI mean the end of doctors? Most researchers couldn’t disagree more,” he said, adding that doctors using AI would replace doctors who do not use this technology.

But AI is merely a useful intelligent tool – like magnetic resonance imaging (MRI). It won’t replace radiologists; instead, it is spawning a new discipline requiring the understanding of both medicine and technology.

A study by The Lancet Digital Health suggested that the diagnostic performance of deep-learning models equals that of healthcare professionals. In 2018, a custom-built AI machine designed to diagnose brain tumours and predict haematoma expansion scored 2:0 against its human competitors, comprising 15 senior doctors from China’s premier hospitals. In 2019, researchers from the University of California at Berkeley and the University of California at San Francisco announced that they had created an algorithm that can detect brain haemorrhages with an accuracy higher than two out of four radiologists.

This week, Boston University School of Medicine researchers found that AI may be as accurate as clinicians in diagnosing dementia. In fact, the AI model surpassed clinicians at differentiating the type of dementia in patients who had been diagnosed.

This does not mean doctors are obsolete. Instead, AI takes over many of the time-consuming and tedious aspects of the profession while faster and earlier diagnoses give doctors scope for more patients. As Topol argues, AI’s significant opportunities include reducing errors and workloads, accurately diagnosing diseases such as cancers and fixing the precious and time-honoured connection and trust.

A 2019 study in the Annals of Family Medicine shows that primary care doctors spend, on average, two hours on administrative chores for every hour they spend in direct patient care. Physicians reported an average of between one and two hours of after-hours work per night, usually on administrative tasks. Then, of course, there are limitations to current AI technology, which is focused on reading pictorial scans like ultrasounds, X-rays and CT scans.

As The Lancet Digital Health says: “Scans are never interpreted on their own, they’re analysed alongside blood results, historical data, prescriptions from GPs and previous hospital admissions, referral letters, taking the patient’s history and then taking it again, what the nurse told you before they went on their lunch break, and any number of other sources of information.”

This, in part, also addresses concerns of a doctor shortage, particularly in the public sector. For example, the increased speed and accuracy of cancer diagnostics through analytics that can characterise tumours and predict therapies has not replaced doctors, but quickened their efforts and given them the space to attend to more patients.

New technology will decrease costs of healthcare worldwide. Almost two-thirds of healthcare costs come from non-communicable diseases, like cancer and heart failure, which, if caught early, can be treated more effectively, less expensively.

Amid these shifts and in our quest towards the creation of our own medical school, there are two key lessons to bear in mind.

First, in an ever-shifting context and with the proliferation of digital technology, we must adopt technology that prepares our students for the future of work but also gives them a more well-rounded higher education experience. We already have extensive experience with technology and blended learning models that can be replicated and improved on.

Second, we must emphasise the importance of partnerships.

The movement, and enrolment of staff and students across higher education institutions worldwide, challenge perspectives, allowing us to learn from each other. Economist Ricardo Hausmann says this is the key to skills-building and employment creation. The future of medicine and education alike is rapidly changing. As the pandemic has demonstrated, this is an opportunity we must grab to stay relevant, and as UJ’s tagline reminds us, constantly reimagine the future.

Study details

A comparison of deep learning performance against health-care professionals in detecting diseases from medical imaging: a systematic review and meta-analysis

Xiaoxuan Liu, Livia Faes, Aditya U Kale, Siegfried K Wagner, Dun Jack Fu, Alice Bruynseels, et al.

Published in The Lancet Digital Health on 25 September 2019

Summary

Background
Deep learning offers considerable promise for medical diagnostics. We aimed to evaluate the diagnostic accuracy of deep learning algorithms versus health-care professionals in classifying diseases using medical imaging.

Methods
In this systematic review and meta-analysis, we searched Ovid-MEDLINE, Embase, Science Citation Index, and Conference Proceedings Citation Index for studies published from Jan 1, 2012, to June 6, 2019. Studies comparing the diagnostic performance of deep learning models and health-care professionals based on medical imaging, for any disease, were included. We excluded studies that used medical waveform data graphics material or investigated the accuracy of image segmentation rather than disease classification. We extracted binary diagnostic accuracy data and constructed contingency tables to derive the outcomes of interest: sensitivity and specificity. Studies undertaking an out-of-sample external validation were included in a meta-analysis, using a unified hierarchical model. This study is registered with PROSPERO, CRD42018091176.

Findings
Our search identified 31 587 studies, of which 82 (describing 147 patient cohorts) were included. 69 studies provided enough data to construct contingency tables, enabling calculation of test accuracy, with sensitivity ranging from 9·7% to 100·0% (mean 79·1%, SD 0·2) and specificity ranging from 38·9% to 100·0% (mean 88·3%, SD 0·1). An out-of-sample external validation was done in 25 studies, of which 14 made the comparison between deep learning models and health-care professionals in the same sample. Comparison of the performance between health-care professionals in these 14 studies, when restricting the analysis to the contingency table for each study reporting the highest accuracy, found a pooled sensitivity of 87·0% (95% CI 83·0–90·2) for deep learning models and 86·4% (79·9–91·0) for health-care professionals, and a pooled specificity of 92·5% (95% CI 85·1–96·4) for deep learning models and 90·5% (80·6–95·7) for health-care professionals.

Interpretation
Our review found the diagnostic performance of deep learning models to be equivalent to that of health-care professionals. However, a major finding of the review is that few studies presented externally validated results or compared the performance of deep learning models and health-care professionals using the same sample. Additionally, poor reporting is prevalent in deep learning studies, which limits reliable interpretation of the reported diagnostic accuracy. New reporting standards that address specific challenges of deep learning could improve future studies, enabling greater confidence in the results of future evaluations of this promising technology.

 

Daily Maverick article – Reimagine the future of medicine and education in the 21st century of digital technology (Open access)

 

Boston University article on Artificial Intelligence (Open access)

 

The Lancet Digital Health article – A comparison of deep learning performance against health-care professionals in detecting diseases from medical imaging: a systematic review and meta-analysis (Open access)

 

Annals of Family Medicine article – Tethered to the HER (Open access)

 

See more from MedicalBrief archives:

 

OECD: How artificial intelligence could change the future of health

 

Artificial Intelligence can detect low-glucose levels via ECG without fingerprick test

 

AI arrhythmia predictor has ‘potential to transform clinical decision-making’

 

SAB Foundation: South African entrepreneur launches AI Solution for breast cancer identification

 

AI ‘at best’ on a par with human experts when making image-based diagnoses – review

 

AI outperforms humans in creating cancer treatments — but doctors balk

 

Google Health using AI to improve breast cancer screening

 

 

 

MedicalBrief — our free weekly e-newsletter

We'd appeciate as much information as possible, however only an email address is required.