What is life in 52 seconds?

 

Sherlock in Health: How artificial intelligence may improve quality and efficiency, whilst reducing healthcare costs in Europe

Summary

Various megatrends are impacting healthcare systems in Europe, creating a scenario where artificial intelligence based technologies could be deployed for the benefit of all stakeholders. Converging of these trends, such as, aging of the population in the region, the resulting high cost of healthcare, coupled with patients becoming more demanding and value-focused, is creating a situation where technology can help in improving healthcare access, quality and affordability. It is not hard to imagine a future with intelligent technologies helping us diagnose diseases faster, and assisting doctors in treatment decisions, armed with evidence based analysis of likely outcomes.

AI is increasingly becoming a part of the healthcare ecosystem. Some of the applications developed, though in early stages, are having an impact across care pathways, starting at prevention, to diagnosis, treatment and recovery. At this stage, it is very important to analyse the demand and potential benefits from AI applications in healthcare in Europe. This will not only help us distinguish between realistic hopes and unrealistic hypes, but will also help in focusing our efforts in the right areas.

Analysing demand and potential benefits throughout healthcare is very complex and many factors come into play. We looked into three care pathways as a representation of both the benefits that might accrue from AI use in healthcare and the medical dilemma’s it will create. We estimate that large-scale AI use could yield benefits of the following magnitude.

•     For childhood obesity: AI use could yield cost saving of up to EUR 90 billion over the next ten years. This saving estimate includes benefits from lower medical costs, and reduced losses from lower productivity and sick days. AI could also help in increasing the efficiency of self-monitoring for preventing obesity.

•     For diagnosis of dementia: AI use could help save up to EUR 8 billion in diagnosis cost over the next ten years, largely driven by increased rate of diagnosis at primary care level. AI can help diagnose with up to 90 percent accuracy, bringing it to a large proportion of dementia patients who never receive a formal diagnosis.

•     For diagnosis and treatment of breast cancer: AI use for diagnosis and treatment of breast cancer could be very helpful in early detection, also helping in treatment decision making and reducing doctors’ direct engagement in potentially repetitive tasks. It could help save up to EUR 74 billon over the next ten years, if used on a large scale.

However, achieving these benefits will not be easy. Various challenges exist, such as lack of sufficient data, enabling data standards and regulations. According to various experts we interviewed, technical, legal and financial feasibility of adopting AI will be critical. Equally important, will be to assess the psychological feasibility – is the public actually ready and receptive to AI adoption in health? Based on our analysis

of constraints and interviews with industry experts, we make three recommendations to help improve technology-driven healthcare services in Europe.

•     Introducing a balanced scorecard in policy making will ensure that the focus is not restricted to any one policy area, such as: improving the quality of healthcare; containing the cost of care; or managing overall population health.

•     Moving quickly and consistently on regulations will ensure that the vision on AI within the healthcare industry is matched and supported by timely regulations.

•     Redefining reimbursements to support outcome based care will alleviate any payer related concerns for providers and patients, providing the much needed development boost to AI tools.

Introduction

Access to quality and affordable healthcare is a challenge that is growing every day. Increasing demand and increasing scarcity of health care personnel has put pressure on healthcare delivery, which is in constant need of optimisation. According to data collected from several EU nations, medical errors and healthcare related adverse events occur in eight to twelve percent of hospitalisations. Preventing such mistakes could help to prevent more than 3.2 million days of hospitalisation each year within the EU1.

Technological breakthroughs in artificial intelligence and the availability of big data present the promise of reducing such errors, whilst making healthcare more accessible and affordable. Artificial intelligence will not completely replace physicians and care workers, but it can play a key role in reducing the pressure on healthcare systems and be a decision supporting tool for physicians.

In this paper we assess the healthcare landscape2 in Europe3 and its readiness for artificial intelligence (AI) applications. We also estimate the probable benefits of using AI applications in healthcare, based on three different but interrelated dimensions: potential cost savings to patients; rise in efficiencies in healthcare services; and the increase in accessibility of healthcare services. Furthermore we analyse three conditions and associated care pathways:

•    Prevention of childhood obesity;

•    Diagnosis of dementia; and

•    Diagnosis and treatment of breast cancer

Based on the analysis, we identify some major challenges and the following steps needed, in order to begin moving to large-scale AI adoption and advanced healthcare.

1    http://www.euro.who.int/en/health-topics/Health-systems/patient-safety/data-and-statistics

2    AI will have an impact on all health related fields including healthcare services, pharmaceuticals, and life sciences. For this study, we are only looking at the potential impact on healthcare

3    The geographical scope of the study covers Europe. Interviews have been conducted in Austria, Germany and the Netherlands

Year in review: Not all bodies act their age

FOR THE AGES  People age at different rates, but the underlying drivers of aging likely remain the same, scientists suggested in 2015.

Cells offer new clues to the mystery of growing old

BY 

MEGHAN ROSEN

Age was all the rage this year, as headlines about a provocative study blared what many people already suspected: People grow old at vastly different rates.

The study, out of Duke University, analyzed the health of nearly one thousand 38-year-olds and found that some resembled people a decade older while others appeared years younger (SN: 8/8/15, p. 10). Researchers determined this “biological age” based on health indicators such as body mass index, blood pressure and cholesterol level. The finding tapped into a mystery that has long captivated scientists and the public alike — “why some people can live to 120 with no disease, and others are already in bad shape at age 70,” says molecular biologist Martin Hetzer of the Salk Institute for Biological Studies in La Jolla, Calif.

A handful of recent studies have offered some tantalizing clues. Molecular mayhem within cells, scientists suggested this year, may lie at the root of aging. After examining populations of proteins in the brains and livers of rats, Hetzer and colleagues reported in Cell Systems that long-lived brain proteins appear to become damaged over time. “These are proteins involved in essential cellular functions,” Hetzer says. Unlike liver cells, brain cells rely on proteins that can survive for an animal’s lifetime. Some of those proteins help control messages that pass between brain cells, for example, while others help keep the cell organized. The breakdown of these proteins could be a key driver of aging, Hetzer says.

True age

Though they were all 38 years old, New Zealanders participating in a study reported this year differed in their biological age, which researchers determined based on various health indicators.

SOURCE: D. BELSKY ET AL/PNAS 2015

As could the unraveling of cells’ DNA, a study in Science suggested. In healthy young people, long stretches of DNA pack tightly together in neat bundles called heterochromatin. These bundles aren’t packaged quite so well in old people, researchers comparing DNA from the teeth of young and old people found (SN: 5/30/15, p. 13). The researchers originally identified the disorganization of DNA bundles in people with a premature-aging disorder known as Werner syndrome. The change in the 3-D architecture of the genome gives proteins easy access to stretches of DNA that are supposed to be tucked away.

Beyond the deterioration of proteins and DNA packaging, the body may age when essential barriers break down. When young stem cells in the brain divide, they build a wall that sequesters junky proteins into daughter cells. (This junk mucks with cellular machinery, a problem for stem cells churning out new brain cells.) But old stem cells aren’t so tidy, scientists reported this year in Science (SN: 10/17/15, p. 10). Old stem cells in the brain had worn-out walls inside an organelle called the endoplasmic reticulum. These weakened walls let cellular junk seep back in during division, cutting down on the stem cells’ ability to produce new cells.

Another protective wall, the blood-brain barrier, might also waste away with time. Usually this barrier guards the brain from dangerous toxins in the blood. But MRI scans measuring barrier permeability in living human brains found that old people had leaky walls around the hippocampus, a structure involved in learning and memory (SN: 2/21/15, p. 8). Researchers linked the leaky walls to damage to pericytes, cells found within the blood-brain barrier and crucial to its formation.

In normal human stem cells (left), DNA packs into tight bundles. But those bundles unravel in cells genetically altered to age prematurely (arrow, right), leading to enlarged nuclei.

W. ZHANG ET AL/SCIENCE 2015, REPRINTED WITH PERMISSION FROM AAAS

Differences in barrier integrity also showed up among old people. Those with learning and memory problems had more pericyte damage than healthy individuals of the same age, Berislav Zlokovic of the University of Southern California in Los Angeles and colleagues reported in Neuron. Breakdown of this brain barrier may kick off some of the cognitive troubles of old age, the authors proposed.

All in all, scientists continue to chip away at the mysteries of growing old. The molecular events identified this year add to previously discovered signs of aging — the shortening of telomeres, the protective caps on chromosomes, for example, and damage to mitochondria, the energy factories of cells. But the picture isn’t complete, or simple. Big questions remain, including how to distinguish healthy aging from disease, as well as the roles of genetics versus environment.

The next major challenge, Hetzer says, will be figuring out which cellular changes are merely a sign of aging and which changes are driving it.  Understanding these details could help scientists design antiaging drugs, says geroscientist Felipe Sierra of the National Institute on Aging in Bethesda, Md. “How we age is something we can control,” he says. “Trying to convince people to exercise or eat better has not worked, so we need an alternative.”

One day, doctors might prescribe drugs that fight aging, just like prescribing drugs for diabetes or high cholesterol. Now that sounds like a story that would make some headlines.