to main content Machine Learning for Survey Consistency Fact Sheet | The Joint Commission

Machine Learning for Survey Consistency Fact Sheet

  • Launched in January 2021.
  • Uses artificial intelligence (AI) technology to produce results for the accurate and consistent standard or element of performance (EP) under which to score a survey or review finding.

AI is the ability of programs to learn and reason similar to humans. AI encompasses machine learning, which is based on the idea that systems can learn from data, identify patterns, and make recommendations and decisions with minimal human intervention.

The Machine Learning for Survey Consistency technology is being used by the Joint Commission’s surveyor and reviewer cadre across all Joint Commission accreditation and certification programs for all onsite, offsite and hybrid surveys and reviews. This new technology will help improve the consistency and accuracy of survey and review findings, while also creating time-saving efficiencies for surveyors and reviewers as they score findings.

With the new technology, surveyors and reviewers may input keywords or phrases to identify the correct standard or EP by mining previously scored findings. Previously, only a small fraction of data was leveraged as “searchable” to associate a finding with a standard or EP. With the new technology, all remaining vetted data is available and searchable – producing a result for the correct standard or EP that should be assigned to a finding to appear on the Top 5 choices more than 80% of the time. The efficacy level will continue to improve as the machine learning technology matures with continuous learning.

Staff from the Joint Commission’s Standards Interpretation Group (SIG) also are using machine learning technology via an interface called CLUES (Consistent Learning and Understanding of Event Scoring) that provides the same output. This is the first time that surveyors and reviewers as well as SIG staff have been able to work out of a database that takes into consideration an organization’s customizable services data.

Read More