Precision Medicine News

Strategies for Using Predictive Analytics, AI to Improve Care

Bringing predictive analytics and artificial intelligence into the healthcare system can improve the quality of care for patients.

predictive analytics artificial intelligence

Source: Getty Images

By Erin McNemar, MPA

- As artificial intelligence (AI) technology becomes more interrogated into the medical community, mastering the element of predictive analytics will be a critical component in bettering the quality of care for patients.

In order to prove better care of individuals, providers must work efficiently, effectively, and accurately. Integrating predictive analytics into the healthcare system may accomplish those goals and could also provide some relief for medical professionals in the process. However, to meet these goals predictive analytic systems must be properly trained to avoid creating further care disparities.

How Predictive Analytics Works

By using machine learning and data mining technologies, predictive analytics works to assess risks and predict results. The system will look at historical data and statistical modeling to determine outcomes for individual patients.

As technology continues to advance, predictive analytics is showing potential in several medical areas. Through predictive analytics, scientists at James Cook University have made progress in developing strategies to keep premature babies alive.

By creating a Neonatal Artificial Intelligence Morality Score (NAIMS), doctors can assess the risks that each premature infant faces and give them a score. The risk score helps doctors determine what course of treatment will be best for each infant and what potential risk factors could arise along the way.

READ MORE: Predictive Analytics Model Forecasts Drug Overdose Deaths

“Preterm birth rates are increasing almost everywhere. In neonatal intensive care units, assessment of mortality risk assists in making difficult decisions regarding which treatments should be used and if and when treatments are working effectively," said JCU engineering lecturer Stephanie Baker in a press release.

For predictive analytics to be successful, however, ensuring that algorithms are trained correctly is a critical step. According to a playbook recently released by the Center for Applied AI at Chicago Booth, there are four steps providers can take to avoid algorithm bias.

These steps include understanding the goal of the algorithm, being specific in the target, being prepared to update or throw out the algorithm if it is not meeting the company goals, and continuing to conduct algorithm audits.

According to the Center for Applied AI at Chicago Booth, there are two missteps that organizations often take when it comes to algorithms. The first is the algorithm will have the correct target but is trained in non-diverse populations and misses large groups of people. The second is the algorithm will have the wrong target altogether.

However, according to the Center for Applied AI at Chicago Booth, these mistakes can be avoided by following the provided four steps.

The Benefits of Predictive Analytics

READ MORE: Predictive Analytics Tool Accurately Detects Patient Deterioration

There are several benefits to using predictive analytics in the medical field, the first being that it increases the quality of care. By using historical data, medical professionals can determine the best course of action and treatment plan for each individual patient. This can play an important role in chronic disease management

Not only can predictive analytic methods lead to a quicker diagnosis, but these methods can also efficiently evaluate the effectiveness of a treatment plan. Rather than spending months on a treatment plan that will end up not producing the desired results, predictive analytic methods can recognize trends sooner, saving the patient time and money for unnecessary treatment.

With predictive analytic methods, doctors can switch the treatment plan to something more effective and more likely to assist with chronic disease management.

In addition to the benefits of predictive analytics that are found in chronic disease management, there is also a push to make the analytic method more accessible, eliminating care disparities.

Researchers at the University of Michigan are working to develop an app using artificial intelligence that will scan speech and vocabulary patterns to catch possible early signs of Alzheimer’s disease. Using predictive analytics, the app then will give the user a score and indicate if they are in a high-risk category for developing the disease.

READ MORE: Artificial Intelligence Predicts Atrial Fibrillation, Stroke Risk

While the app is unable to make a formal diagnosis, it will encourage early intervention and patient engagement.

"You cannot replace that human interaction," Jiayu Zhou, an associate professor in MSU's College of Engineering, who is leading the project’s artificial intelligence development said in a press release. "The final assessment will be done by a patient's physician. But if you have doubts and the app says you're at a higher risk, you don't have to wait. You can visit a clinician and take the next steps."

Predictive Analytics and the Decision-Making Process

While predictive analytics shows promise at giving quicker diagnoses, developing treatment plans, and increasing accessibility, it can also provide benefits to patients in the medical decision-making process.

Along with predicting the effectiveness of treatment plans, predictive analytics can determine the mortality rates of patients that enter the intensive care unit (ICU). This information is critical to help doctors make decisions regarding what resources to allocate to individual patients.

In situations where decisions must be made quickly in order to save a patient’s life, predictive analytics serves as an important tool for medical professionals 

“When a patient first arrives, a series of decisions must be made, each committing further resources and subjecting the patient to increased risk of adverse events, incidental findings, costs, and potentially avoidable diagnostic testing and empiric treatment. Stronger predictions, using data gathered from the ED, allow clinicians to avoid marginal tests, procedures, and admissions,” a paper published in Annals of Emergency Medicine stated. “The utility of prediction depends both on accuracy and time to the clinical encounter. At initial presentation, the utility of prediction is low because insufficient patient data exist to drive accuracy.”

As new illnesses and diseases, such as what was seen with COVID-19, develop, providers should continuously collect data regarding the best courses of treatment. This data can then be worked into predictive analytics, effectively improving the quality of care patients receive.