As the federal government tries to hone the future of its “actionable medical knowledge” database, the Texas Medical Association is urging officials to ensure the tool offers meaningful, evidence-based information to physicians, including with the evolution of artificial intelligence (AI).
The CDS Connect project, started in 2016, is a federal database of clinical decision support (CDS) items – such as clinical practice guidelines, peer-reviewed articles, local best practices, and clinical quality measures – in varying forms and across a variety of clinical topics. As part of an effort to enhance health care quality and safety, CDS items in the freely available web-based platform are integrated into electronic health records (EHRs).
“Physicians should have confidence that the CDS tools used by their EHR vendor reflect current care standards and have been properly peer reviewed by national specialty societies,” Manish Naik, MD, Chair of TMA’s Committee on Health Information Technology, told Texas Medicine Today.
This was one of the top recommendations TMA offered up in a July 23 letter to federal officials to ensure CDS Connect continues to evolve to help Texas physicians’ medical decisions. The missive was in response to a request by the Agency for Healthcare Research and Quality (AHRQ) seeking feedback regarding a “sustainment model” for the database.
A variety of organizations, including other federal agencies, contribute information to the repository. To ensure the database’s credibility and usability in the future, TMA “recommends partnering organizations be credible and nationally recognized in providing clinical guidance with evidence-based care standards leading to the best possible outcomes for patients treated,” TMA President G. Ray Callas, MD, wrote. “National medical specialty societies are good examples of these types of organizations. Physician informaticists also would be ideal reviewers.”
The letter also states CDS Connect information should:
- Display a date of last review;
- List reviewing organizations;
- Be reviewed at least once a year;
- Be easily identified on whether that information is active or retired;
- Spell out if information is a draft or “experimental;”
- Give users a way to offer feedback when they see incorrect or outdated material in the system.
TMA’s letter also points out CDS Connect could have the potential to be aided by AI in searches but cautions the database should be peer-reviewed and protected from any inaccuracies as a result. In that regard, TMA encourages AHRQ to offer a platform for feedback on AI searches and recommends AI-generated summaries refer back to the source materials to ensure verification.
Last Updated On
August 08, 2024
Originally Published On
August 07, 2024