IN A NUTSHELL Author's Note.. In the next few years, we should seriously consider using AI to reduce health inequalities at the level of the individual. The development of cloud storage allows vast amounts of data to be stored and manipulated, and its use in conjunction with AI, permits data analysis on a scale not previously possible ..
By Dr. Brian Johnston
Strategic Intelligence Analyst
London, United Kingdom
The Power of AI for Health Inequalities
George Orwell made his famous quote “All animals are equal, but some animals are more equal than others,” in the novella Animal Farm, and in a sentence described an ancient and pervasive mindset. Why share a resource fairly and equitably, when we can keep more for ourselves and use long-lasting legal, political and social structures to restrict access to its benefits?
Health inequalities represent a potent example of this mindset in action, where affluent people often live longer, healthier and more productive lives, that are less impacted by the ravages of chronic illness and long term conditions, than their more deprived contemporaries. The wealthy have a freedom of choice that is denied to others; when times are hard, they can move quickly to a place of safety, and if required, access the best of medical care at short notice, avoiding long waiting lists, delayed operations, and crowded hospitals.
This is not to say that we should be either jealous or resentful of rich people who receive excellent healthcare in a timely manner. Instead, we should desire such high quality health and wellbeing for as large a proportion of society as possible, and do everything in our power to reduce differentials in accessing these resources. The best possible health and wellbeing, based on good quality healthcare, should be a universal aspiration, and one that key stakeholders and leaders should dedicate their lives to achieving.
Unfortunately the causes of health inequality are many and varied, and some, like poverty, poor education and unemployment, are often deeply rooted in societies, affecting successive generations. In many parts of the world, poverty has proven to be both pernicious and persistent over decades, resistant to effective intervention and a blight on large sections of the population. Whether it is the effects of war, corruption, political repression, criminal activity, incompetence, megalomania, greed or inertia on the part of those supposedly in control, many of the drivers of health inequalities are created by humans. From this perspective, they should therefore (in theory) also be reversible, by humans.
Another potentially reversible factor, climate change, is believed to differentially impact the populations of low and middle income countries (LMICs) due to their increased vulnerability from low socioeconomic status, poor health infrastructure and geographic location. A recent review published in the British Medical Journal (BMJ), looked at the projected effects of climate change on human health in LMICs, and has predicted that it may substantially increase the burden of communicable and non-communicable disease in LMICs.
However, whilst climate change does create health inequalities, no-one entirely escapes its negative effects. For example, it has been associated with the spread of mosquito borne diseases into areas of Europe where mosquitos did not previously thrive, due to recent hotter and wetter summers. These weather conditions are likely to have contributed towards an increase in locally acquired cases of West Nile Virus (WNV) within Europe. In addition, they have also allowed the Asian Tiger mosquito to expand its range and become established in a number of European countries. This type of mosquito (Aedes albopictus) spreads the “tropical diseases” dengue and chikungunya, and locally transmitted cases of dengue have been reported in both Italy and France this summer, with a locally transmitted case of chikungunya also detected in France.
Whilst climate change reminds us that no-one, not even in affluent countries, is immune from certain negative impacts on health, the reduction of health inequalities remains a major challenge for every country. A recent investigation by Lord Darzi looked into the state of the health service in the United Kingdom and how it could better address health inequalities. It recognised the disproportionate investment in hospital care, shortcomings in social care, and the need to shift focus towards prevention, through the enhancement of primary and community care. Whilst infrastructure, leadership and capital investment changes are important drivers for effective change, they are by no means the entire picture.
The shift towards prevention is a vitally important one, with the responsibility for a person’s health and wellbeing resting with everyone – we all bear a joint responsibility for making and keeping this world a better, healthier place to live. Health services and related structures should be there to enhance and support this endeavour by providing treatments, advice and expertise. In this respect, changes to infrastructure and the way that health and wellbeing are delivered, certainly play their role, but, as the Marmot Review clearly demonstrates, health inequalities can and do worsen over time, even when we know what should be done at a structural level to reduce them.
A wide variety of factors contribute to the worsening of health inequalities and these include deprivation, unemployment, housing, transport, personal relationships and our environment. These factors interact in complex ways and an understanding of this interplay is essential for the delivery of effective healthcare services aimed at reducing health inequalities. In the years to come, AI will undoubtedly shed light on patterns within the data relating to these factors, which have proven elusive or wholly undetectable by human beings.
The integration of primary and secondary care datasets at local, regional and national levels will also help in this endeavour, permitting the development of treatments and producing efficiencies in service delivery, that would not otherwise have been possible. However, this data integration has historically proven to be a difficult task, due to the variety of hardware and software used across different healthcare organisations, data quality issues, governance considerations and a plethora of other practical considerations.
In the next few years, we should seriously consider using AI to reduce health inequalities at the level of the individual. The development of cloud storage allows vast amounts of data to be stored and manipulated, and its use in conjunction with AI, permits data analysis on a scale not previously possible. For example, at an individual level, apps and wearable devices could allow large proportions of the population to monitor their bodily systems in real time, if cheap, durable, reliable and mass produced kit could be manufactured in a cost effective way.
By linking these devices to AI algorithms, with direct access to the person’s medical history and test results held in a data lake, medical red flags and emergencies could generate immediate warnings. In addition, the individual could be directed to the nearest appropriate healthcare facilities relative to their current position, which would have been notified of their coming and have received a summary of all the relevant information. Algorithms could also be used to triage patients and channel them towards facilities that provide specialist care in their condition, by having hospital computer systems talk to each other in real time, across secure connections.
Of course, the benefits would not be limited to acute events, as AI systems could significantly improve prevention, by highlighting negative patterns within the individual’s health data and even suggesting behaviours to bring the person back into line with a “healthier” set of results.
Much of the hardware necessary for these AI initiatives is currently available or in development, but making it cheaper and more accessible to larger numbers of people on low incomes, or living in deprived areas may prove a serious challenge. However, where possible, this approach could be adopted in LMICs and within poor communities everywhere, as a future proofed, practical, scalable and efficient way of addressing health inequalities.
Such projects, designed to catalyse the power of AI and cloud computing to produce real change in impoverished communities, may not solve the problem of health inequalities, but they could seriously impact it, and where one human life is saved as a result, they will have done good in the world.
By the same Author on PEAH
Scared New World
Shifting Sands – Health in a Changing World
How to Combat Future Pandemics
The New Abnormal
Living with COVID in a Transformed World
Death in the Time of COVID
Unleashing the True Potential of Data – COVID-19 and Beyond