Wiki Calculating coding quality

[sd4210]

Do any of you perform quality checks on drug administration?
We perform routine quality checks on our coders. We code drug administration for facility outpatients but there is a debate over how the drug admin should be counted. At 1 time we counted every drug admin we coded, however if a coder forgot to charge the drug admin the missed drugs could effect the coder's quality percentage for the entire month. We changed to just counting each different drug admin CPT code. For example if there were 2 sequential pushes that would only count as 1 coded correctly or missed. I'd appreciate any input you could offer.

 

[thomas7331]

Some organizations that I've worked for classify errors in two ways - 'financial' and 'statistical'. A financial error is one that causes, or has the potential to cause, an incorrect payment (overpayment or underpayment) or denial. This type of error carries more weight because of the risk of monetary impact that it can have on the organization. A statistical error is one that just involves incorrect reporting, such as an incorrect choice of diagnosis code or other incorrect information which does not necessarily result in a material change to the claim or the payment. I've not heard of counting errors per code, only per encounter. I've usually seen quality measured as the ratio of accurately coded encounters to total encounters coded, with statistical and financial quality tracked separately.

In my time managing coders, I always found it a challenge to accurately measure for coding quality or productivity using a numeric measure or on a 'per code' basis because encounters vary so much from one specialty to the next and it is like trying to compare apples and oranges. I took the approach of making quality and production measures a tool for improvement rather than a way to compare quality between coders. In other words, I found that whatever way that I measured a coder's work, it was most useful for measuring how an individual was improving over time, but less useful for seeing who was doing better coding. Unless your coding team is all sharing the same or very similar type of work, I think that a quality measure is something that has to be individually tailored based on the type of encounters each individual coder is working on. I know this probably doesn't help much, but perhaps it's some food for thought. I'll be interested to hear others' input.