Although demand for healthcare is increasing in many parts of the world, it isn’t always being met. Whether it’s health services struggling to meet the needs of ageing populations or rural patients without access to local services, there’s often a shortfall in supply. Patients also expect better outcomes than health services may be able to deliver.
Artificial intelligence (AI) may be part of the solution. AI technologies could help increase access to care, improve outcomes, and make better use of a physician’s time. But there are many barriers to adoption of AI in the healthcare industry. These include the difficulty of assembling the digital data AI needs, as well as a lack of AI workers across the world.
Patchy digital data
A lack of patient data in digital format is just one of the barriers to AI use in healthcare. Some AI technology sifts through data from groups of patients to determine the best approach for individual patient care.
One example is IBM’s Watson technology, presently capable of using patient data to craft better outcomes for individual cancer patients. Another example is the use of machine learning and advanced analytics to triage patients in intensive-care situations and identify those most at risk.
This kind of technology relies on the right kind and right volume of patient data being available.
Healthcare technology systems and the data-gathering practices of providers don’t always provide the right data in the right format for these types of AI use. While healthcare technologies can be extremely innovative when it comes to diagnosis and treatment, hospitals and clinics also depend on much more outdated technology to manage patient data.
In some healthcare systems, patient data is still being stored in paper files. In others, it’s stored in a multitude of old systems that aren’t great at sharing information with one other or with new systems.
AI is most effective when it’s able to work with a complete set of patient health records. Getting all this data together is quite a challenge. Of course, there are also many privacy concerns and regulations around sharing patient data in any format, which also makes it harder to marshal the data that AI needs.
Skill shortages and costs
There’s a lot of excitement surrounding the potential of AI for many industries, but a shortage of experienced and skilled AI talent may mean this potential can’t be realized everywhere. It will likely take years for the labor force to emerge with the skills healthcare needs to develop and apply AI broadly.
In the meantime, the industry will probably be dependent on third parties for support with AI technology, which is a costly way to apply technology. Skill shortages are likely to be a major limiting factor for applying and developing AI in any sector, including the healthcare industry.
Lack of AI workers isn’t the only factor pushing costs up. The sheer cost of developing any new technology means that it can be prohibitively expensive for most patients and may not provide the outcomes being sought.
Although there’s been much hype about IBM’s Watson computing system, which was promoted as a solution to improving cancer treatment, recent analysis shows that what the coverage missed was a caveat about the lack of evidence that the technology improved patient outcomes, lowered costs, or provided some other benefit. In fact, following a University of Texas audit, M.D. Anderson Cancer Center found that Watson did not work with their new electronic medical records system, and the cancer center is now seeking bids for a new contractor.
Culture and language
One surprising barrier to the adoption of AI in healthcare is cultural differences between markets. Although IBM Watson is being marketed worldwide, much of the data it relies on is relevant to its home market in the U.S.
This means some of the treatments it recommends are not available in all markets, or not available within that market’s insurance system. It’s a barrier to adoption if not all the recommendations it makes are applicable locally.
But culture also means that AI is received and valued differently in the various markets of the world. China has a serious problem of conflict between patients and doctors. There’s a high rate of violence against workers in the healthcare industry and a general lack of respect for the profession.
AI is seen as a potential intermediary between angry patients and workers in the healthcare industry that could help protect practitioners. In a market where doctors’ opinions are often mistrusted, AI intervention in medical plans could help reassure patients that their physician is giving the correct advice.
Despite the very real barriers to AI adoption in this industry, it’s highly likely that healthcare will see huge advances and widespread use of AI technologies in the next few years – and at the right price.
A report by Accenture predicts huge and rapid growth and wide-scale investment into AI startups in the sector. It’s thought that developments may help replace a shortage of skilled health workers by offering services such as virtual nursing and also help improve outcomes with precision surgery carried out by robots.
AI can potentially also help fight against problems such as prescription errors and fraud within the system.
The key driving factor is the potential for cost reduction that it offers the industry. Demand for healthcare cannot always be met at an affordable price for patients and it’s likely that, in the long term, we’ll see AI play a role in making healthcare more affordable.
As the world population ages and healthcare demand rises, this is going to be a welcome solution to meeting needs.