Predictive Analytics & AI in Health Data Management
“Health information professionals are at the core of all information within a health record, making them the ideal professional to be involved in predictive analytics.”
Tammy Combs, Senior Director of Clinical Documentation Integrity (CDI) Programs at the American Health Information Management Association (AHIMA)
Modern healthcare runs on data. In 2018, the healthcare industry generated approximately 30 percent of the world’s data volume (RBC Capital). That share has likely only increased in recent years, with the growing adoption of wearable tech, remote patient monitoring, and electronic health records. But raw data is only as valuable as the insights one can derive from it.
Today’s health information professionals leverage predictive analytics to sift through the data and identify patterns that can translate to improved outcomes for patients, providers, and the bottom line. In the near future, they’ll also utilize AI-powered tools to similar ends. But these potent and novel applications are not fail-safe, and each requires a unique skill set and careful stewardship.
Read on to learn more about the opportunities, challenges, and future of predictive analytics in AI and health data management.
Using Predictive Analytics and AI in Health Data Management
“Predictive analytics is in a continual state of growth as we see more healthcare professionals focus on this type of data analysis,” says Tammy Combs, senior director of clinical documentation integrity (CDI) programs at the American Health Information Management Association (AHIMA). “Health information professionals are at the core of all information within a health record, making them the ideal professional to be involved in predictive analytics.”
Predictive analytics remains more common in the healthcare industry today than AI-powered tools. The differences between them are as numerous as their similarities. Both derive insights from large data sets in distinct ways.
AI-powered tools are largely autonomous, with their algorithmic processes taking place under the hood. Predictive analytics, by comparison, requires a more hands-on approach from health information professionals. Both, however, still require the careful guidance of a human overseer who can strategize and implement the lessons learned from the patterns derived.
“Organizations using predictive analytics may identify a pattern of patients who continually miss appointments and choose to double-book those appointment times,” Combs says. “This information then branches into care management. When specific trends are identified, the care management team is flagged to discover why these patients struggle to make appointments and set up supportive services.”
The possibilities for AI in healthcare are as unbounded as one’s imagination. But its practical applications are likely to be implemented with a safe and steady hand. Reducing administrative burden on patient care providers is a simple first step. But AI will need to follow strict compliance standards around privacy and safety when using information in the health record, and health information professionals have an important role in ensuring that it does.
Challenges Involving Predictive Analytics & AI in Health Data Management
The stakes are different in healthcare than they are in other industries. Mistakes can cost much more than money, and incorrect or incomplete data can be more harmful than no data at all. Reliability is paramount in healthcare and especially in health data management. Skills in data collection, maintenance, securitization, and policy are already highly valued in the healthcare world but will likely become even more important going forward.
“Reliability is one of the biggest challenges,” Combs says. “There is now so much data available that organizations must ensure the information added to a health record is accurate to result in correct predictions. Maintaining high-quality clinical documentation within the health record is vital to accurate predictive analysis.”
AI has all the same concerns around data reliability and many more around its autonomy, which can save time and resources, unlocking huge potential, but not without significant risk. Healthcare providers have already called for caution and regulation regarding AI (BMJ 2023). Many patients are skittish, too (Pew 2023). Regulation is coming, and health information professionals must stay up-to-date on how those impact the industry as a whole.
“With Congress and the Executive Branch following the developments of AI closely, their ability and interest in regulating these technologies will shape how the health system views, has access to, and interacts with these technologies,” Combs says.
The Future of Predictive Analytics and AI in Health Data Management
Healthcare is constantly changing, and so are predictive analytics and AI. Health information professionals sit at the nexus of all three. Their role is massively important and likely to undergo its own disruptive changes in the coming years.
“As with many healthcare roles, there will be an evolution in the skills and competencies needed to support predictive analysis,” Combs says. “These roles may include data collection, report building, analytics, and auditing the data for reliability. Organizations may need to investigate the need for advanced skills and competencies and consider providing training as health information roles evolve.”
Done right, predictive analytics and AI can be a revolutionary force within healthcare. AI tools may become as essential to providers as the scalpel is to the surgeon. But these tools will remain tools, not replacements. And while the full future of predictive analytics and AI in healthcare remains to be seen, health information professionals must play a prominent role in shaping it.
“AI does not replace critical thinking, which is essential in this process, making the need for health information professionals vital to building the logic running AI,” Combs says. “AI should be a resource to assist, not replace, healthcare professionals.”