Between 2004 and 2017, the use of electronic health records (EHR) in physicians’ offices in the USA rose from just over 20% to almost 86%, meaning that healthcare authorities now have more data on their patients than ever before. That includes records on what sort of medical issues people presented with, which treatments helped them, and how long they received medical care.
Now, thanks to predictive analytics, that data is helping to reform treatment strategies, cut medical costs, and, most importantly, save lives.
By analysing large data sets, computers are able to track patterns and extrapolate from them in order to perform predictive analysis. This becomes particularly useful when a number of patients present with an illness or condition which may have multiple treatment options.
Take depression, for example; the range of treatments runs the gamut from talking therapies to medication to (in rare cases) surgical intervention, so it’s difficult to manually keep track of what method works best for improving a patient’s condition. With predictive analytics, however, the software will be able to tell which treatment has the highest success rate within fractions of a second. This is especially useful when dealing with a range of demographics, too, as it’s very possible that factors such as race, gender, and age have an impact on the success of a treatment.
What’s more, with enough data, AI can even predict new treatments for medical ailments. By tracking the effects of drugs on patients, advanced predictive analytics software should be able to tell which components of existing medications are actually viable cures for other conditions. It might notice that patients who take one drug for a skin condition are less likely to develop cancer later on, for instance, or that individuals who rely on medications to help joint pain suffer strokes less frequently.
On top of that, AI models are able to simulate drugs trials in a controlled environment, and – if it analyses the effects of existing drugs – could shave years off the usual time it takes for a drug to go from the research to the distribution stage.
From an economic perspective, too, predictive analytics is having a positive impact on the medical industry.
A 2018 study from Duke University used big data analytics in order to predict patient no-shows at hospital and doctors’ appointments, and found that predictive models using clinic-level data could capture an additional 4800 patient no-shows per year. With this information, medical facilities could potentially avoid wasting working hours on patients who don’t turn up to appointments and improve efficiency by reallocating those slots to people who will turn up.
In addition to this, analytics software can also calculate the busiest times of day for medical facilities, and therefore help them ensure they have staff on-hand when they are needed. This, in turn, improves the patient experience, but also the staff’s, and is likely to have a positive impact on their job satisfaction, their performance, and their health.
The logistics of the supply chain can also be monitored and managed by AI and predictive analytics, and one survey found that implementing data-driven solutions could save the US healthcare system billions of dollars a year:
“Hospitals nationwide could reduce annual supply expenses by approximately $23 billion in aggregate through improvements in supply chain operations, processes, and product use, according to a Navigant analysis of more than 2,300 hospitals. This represents a 17.8 percent average total supply expense reduction opportunity or up to $9.9 million a year per hospital – an amount equivalent to the annual salaries of 150 registered nurses, or the cost of 4,000 cardiac defibrillators or five Da Vinci robots.”
But, as the old saying goes, ‘the prevention is better than the cure’ – and that’s where the power of predictive analytics really shines. Thanks to advanced machine learning techniques, predictive software is able to prevent medical issues before they happen.
A study published in the American Journal of Psychiatry last year found that data from EHR, when combined with the results of a standard depression questionnaire, was accurately able to identify individuals who were at a high risk of suicide. This sort of data is obviously invaluable to those working in psychiatric care, as it helps identify patients who need to be closely monitored.
In non-psychiatric cases, too, predictive analytics has proven to be a lifesaving tool. In 2017, the University of Pennsylvania developed a tool that was able to identify patients who were likely to undergo severe sepsis or septic shock a full 12 hours before the onset of the condition.
“We were hoping to identify severe sepsis or septic shock when it was early enough to intervene and before any deterioration started,” said senior author Craig Umscheid, MD, of the Hospital of the University of Pennsylvania. “The algorithm was able to do this. This is a breakthrough in showing that machine learning can accurately identify those at risk of severe sepsis and septic shock.”
Even when patients are initially treated successfully without AI intervention, predictive analytics can prevent further medical problems. Back in 2016, researchers from the University of Texas Southwestern found there were certain risk factors that would lead to an increased chance of a patient being readmitted within 30 days of their discharge. Using this statistical analysis, an AI may be able to predict the best course of action to care for a patient with these risk factors, and therefore prevent their readmittance.
Thanks to advanced analytics and predictive software, then, healthcare is changing for the better.