<img alt="" src="https://secure.lazy8krti.com/218831.png" style="display:none;">

Article: Does AI offer fresh hope or false hype for healthcare in emerging markets?

Rachel Howard, February 2021

Artificial intelligence shutterstock_644724364- RobotArtificial intelligence (AI) and machine learning (ML) offer exciting possibilities to uncover new healthcare insights from big data generated by individuals and aggregated from the collective experience of populations. This has huge potential to improve and even transform patient care.

Nowhere is more in need of harnessing the potential of AI than low and middle-income countries. Imbalances of healthcare resources and a shortage of per capita healthcare professionals (HCPs), in many cases exacerbated by ‘brain drain’ to high-income countries, often translate into long waiting times and ultimately inadequate public healthcare provision.

Emerging markets have been identified as fertile grounds for innovation in healthcare digitalisation, with the potential to ‘leapfrog’ mature markets due to:

  • high unmet need to relieve pressure on overburdened healthcare systems
  • being less impeded by legacy digital systems, such as electronic medical records (EMRs) that use outdated and fragmented technology platforms – they can skip the server and move straight to the cloud

To begin our digital special edition of New Dawn I bring these themes together, taking a critical view of the potential of emerging markets to leverage the power of AI in healthcare. I’ll review the progress of AI in healthcare to date and share case studies of where AI is being implemented in emerging markets, attempting to filter out the jargon and read between the lines of the bold claims of venture capitalists seeking to attract investment. To what extent will AI improve healthcare for underserved communities in countries who arguably need it the most, and what challenges have still to be overcome?

Promise vs practice – the unpredictable trajectory of predictable medicine
AI’s potential in healthcare is polarizing. Its strongest proponents have gone as far as to suggest that it will one day turn doctors into dinosaurs. The omnipresent ‘digital doctor’ will lurk on our phones and within our wearables, passively checking our vital signs to alert us we are getting sick before we know it. After being invited to take part in a trial with my Fitbit last year that suggested the device could identify signs of COVID-19 before symptoms start, this doesn’t actually sound that outlandish.

Critics argue that ‘serious’ applications for AI, which truly have the potential to replace doctors, remain far from becoming mainstream. Regulators are still scrambling to establish clear pathways for the proliferation of AI-powered medical devices coming to market. IBM Watson, once hailed as the supercomputer that would revolutionise healthcare after beating two trivia champions on an episode of Jeopardy! back in 2011, has in recent years come under fire for failing to make much of an impact, with tools adding less value to clinical practice than originally hoped and practical applications largely confined to niche areas such as genomics. Issues of data protection and security have yet to be resolved, and the sheer complexity of the decision space in many aspects of medicine means we are still a long way from a point where we can rely on algorithms to adequately determine the patient care pathway. The limitations of health technology infrastructure and incompatibility between IT systems have resulted in slower than anticipated progress in ‘joining up’ the various aspects of patient care to break data out of isolated silos and provide a holistic view.  

However incremental progress within specific fields of medicine is undeniable, with medical imaging at the forefront. Deep learning algorithms have matched the performance of experts when it comes to early detection of cancers and diagnosis of eye diseases. The number of approved and CE-marked AI/ML-based medical devices has increased substantially since 2015, particularly in the field of radiology. A 2016 article in the New England Journal of Medicine predicted that machine learning will “soon exceed human accuracy” and displace a large proportion of the work of radiologists and pathologists.  

Emerging market applications
When evaluating its impact, rather than considering all AI applications together, it’s perhaps more instructive to consider them within a taxonomy of the different functions they serve within different areas of healthcare: hospitals, pharmaceuticals, medical equipment and supplies, medical insurance, diagnostics, and telemedicine.

The most notable applications in emerging markets fall into the latter two categories. Both of which are key areas in which the COVID-19 pandemic has provided further impetus for the virtual delivery of healthcare, accelerating the digitalisation of healthcare and the diffusion of AI into health – meaning recent examples abound.

In October 2020, health tech company Ada announced a partnership with Saudi investment group Obeikan to deploy digital health technologies across the MENA region. Their CE-marked symptom-checking app, Ada Health, uses a chatbot to provide users with healthcare recommendations, with tailored recommendations of conditions with a higher incidence in these markets based on localised disease models.

Similarly, digital health provider Babylon recently signed a 10-year deal with the Government of Rwanda to roll out its AI-driven symptom-checking services to almost the entire population. While it claims this will make Rwanda the world’s most advanced country for digital health, China and India are the major EM innovators in this area, with their vast populations providing huge potential data sources and scalability.

The Chinese government aims to make China the world’s leading AI innovation centre by 2030, and investment in the sector is booming. Ping An Good Doctor, a medical consultation platform with over 350 million registered users, ‘strives to provide every family with a family doctor, every person with an e-profile and everyone with a healthcare management plan’. Equipped with knowledge of about 3,000 diseases, its AI ‘diagnosis system’ triages patients for consultation based on their self-reported symptoms.

In India, which has a large rural population who are particularly underserved by skilled healthcare professionals, AI offers unparalleled opportunity to extend medical services. For example, the start-up Sattva MedTech attempts to tackle the country’s high neonatal mortality rate by enabling lower-skilled healthcare workers to diagnose foetal distress by using electrocardiogram-based monitoring systems powered by AI. India’s Minister for Health has spoken positively of the need for AI in healthcare, with these technologies offering a means of reducing systemic inequalities in access.

An open verdict – Not a solution but a step forward
How successful the AI applications outlined above have been in practice remains to be seen, with many implemented so recently that objective monitoring and evaluation is lacking. The value of apps designed to improve healthcare delivery are arguably only as good as the level of integration they have with the broader healthcare ecosystem, and the human services to which they can ultimately refer patients. Questions of data security, privacy and ownership have yet to be fully resolved, and regulation is even looser and more limited in emerging markets than in countries such as the US and the EU where regulators are still grappling with how to handle this proliferation of new technologies.

Despite these challenges, it’s clear that while AI isn’t a panacea, nor is it a fad that will be going away any time soon. For the foreseeable future, these do not present an adequate substitute for the fundamentally human aspects of medicine – empathy, and the need to make complex decisions no machine has yet managed to master. However emerging markets, with their HCP shortages, are perfectly positioned to benefit from its potential to triage telemedicine services, and its preventative and diagnostic applications in particular, as the most routine tasks can be undertaken by intelligent machines. Where HCPs are at their most overstretched, this alone makes them infinitely better than nothing, as they have the potential to improve efficiencies, freeing up HCPs’ time for the tasks where they can add most value. And where AI initiatives are able to positively impact the patient pathway by enabling faster access to relevant healthcare services and more accurate diagnoses in emerging markets, pharmaceutical companies also stand to benefit, as this in turn has the potential to widen access for their products.

Market insights are essential to maximise the potential benefits AI health tech can deliver: we can conduct help our clients to identify drivers and barriers to adoption of these new technologies through market research, and our user experience division can provide direction to inform how their integration into the healthcare system can be optimised.

Case studies:

Virtual reality shutterstock_759145798

Tomography

How we evaluated the opportunity for a mobile optical coherence tomography

Learn More

shutterstock_1404433022-education-hologram 3818

AI dosing technology

How we assessed AI dosing technology in ophthalmology

Learn More

Sign up to receive Rapport

Rapport is our e-newsletter and online resource for sharing our expertise and experience in global healthcare market research.

Sign up here