Introducing Approval Controls: Best Practice for Safety & Quality in Clinical Coding

Striking the right balance between working fast, and working safely and accurately is a particular challenge for practices implementing AI tools. The temptation to complete documents as quickly as possible must be balanced against having guardrails that ensure that the human-in-the-loop is performing proper checks of AI suggested coding. At BetterLetter, data quality has always been our highest priority, and for this reason we designed the application to ensure that every AI suggestion must be individually and explicitly approved by a human before being written to a clinical system. 

Many of our customers have expressed that they would like to have a risk-based approach to approving AI-suggestions, and this is reflected in our recently released feature, Approval Controls. 

Approval Controls allow practice administrators to assign different approval profiles across users and teams, tailoring the level of AI interaction a coder will experience based on their coding proficiency.

Less experienced coders can be required to approve every AI-suggestion individually, while power-users are able to complete letters with fewer AI prompts and approvals. As coders gain experience, their approval settings can be adjusted to match their skill level, and the system will adapt as their proficiency grows. This helps to ensure coders at every level can work safely with appropriate guardrails in place, and that their efficiency can improve in line with their evolving skillset.

You can read a full technical overview of this feature here.

At BetterLetter, our approach is validated by real-world success across NHS teams, and shaped by the experiences and realities of the practices we work with. 

“For me, the first thing that struck me [about Betterletter] was the safety element. BetterLetter keeps the human factor in the coding process. BetterLetter does things the right way, they don’t cut corners. They follow protocol and good governance. You can see they’ve dotted every I and crossed every T.” - John, Practice Manager at Blackwood Health Centre

To learn more, email us at hello@betterletter.ai