Hospital Testing and Auditing Processes Lag Behind in Artificial Intelligence Usage

While artificial intelligence integration in healthcare organizations is rapidly increasing, a report has revealed that auditing infrastructure and testing processes cannot keep up with this pace.

◉ 0 views
Hospitals Are All In on AI, but Testing and Oversight Haven't Caught Up - MedCity News

As healthcare systems show intense interest in artificial intelligence technologies, a new report revealed that artificial intelligence deployment in hospitals has left testing and governance infrastructure behind.

Artificial Intelligence Adoption is Increasing Rapidly

While healthcare systems increase their commitment to artificial intelligence, the majority of hospitals use third-party tools for clinical and administrative workflows.

Testing and Auditing Infrastructure is Inadequate

According to a report published by the UPMC Center and KLAS Research, enthusiasm for artificial intelligence has outpaced the development of governance and infrastructure.

Validation Methods and Strategies

It was observed that fewer than half of the hospitals participating in the research have a dedicated environment where they can test artificial intelligence tools before patient care integration.

Sixty-three percent of healthcare systems describe their artificial intelligence strategies as developing or ad hoc, with validation methods ranging from formal testing to flexible pilot implementations.

Challenges and Process Costs

While time, resource, and capital constraints cause this gap, hospitals can spend six months or longer on applications by turning to standard IT processes that may not deliver the expected value.

Real-World Data Platforms and Solutions

UPMC uses a real-world data platform called Ahavi, which allows hospitals to validate third-party tools against anonymized patient data prior to deployment.

Additionally, UPMC tests vendor algorithms against its own patient population to catch bias and model drift.

Need for Continuous Monitoring and Training

It is stated that the use of artificial intelligence by clinical teams has increased exponentially in recent years, making it difficult to maintain evaluation standards.

While the healthcare industry continues to lack a consistent approach to evaluate these tools, clinical teams need to be trained on bias and equity issues.