As facilities begin their ICD-10 preparations, one thing to keep in mind is coding audits. Coding is not a “once you learn the basics, you are set” type of job. To keep coding skills strong and healthy requires on-going exercise and stretching. You can never have too much education, training, or exercise. Coding requires a strict diet of education, focus, ongoing review, and feedback to keep the skills healthy and sharp.
Quarterly coding audits offer consistent and continuous feedback to coders, which will be imperative as they begin their ICD-10 training. Budgets for external coding audits will definitely need to be increased, not just after implementation, but also before as you prepare for the go-live date.
Productivity is expected to be impacted by ICD-10 for coders, but it will also have a significant impact on productivity of the coding auditors as well. Today, an experienced auditor can skim the codes on a coding summary form, and often identify obvious errors before they even dig into the chart. However, with ICD-10, auditors will no longer know the codes and every code will have to be looked up for verification.
We are anticipating up to a 40 percent, possibly even 50 percent reduction in total data quality coding auditing productivity initially, which will increase the price and cost of the audit.
Hospitals usually base their audits on either random selection or targeted selection. For future coding audits under ICD-10, we recommend quarterly audits that include both targeted and random chart selections. For example: first- and third-quarter audits could be based on a random chart selection, and second- and fourth-quarter would be based on a targeted selection.
We recommend both random and targeted approaches because how the charts are selected can drive the audit results. There are benefits and limitations to each selection method.
Targeting specific records at high risk for error and external agency scrutiny will promote accurate coding and reduce overall coding errors on the most complex cases. As external agencies audit these records and determine they are coded correctly, this will discourage them targeting the facility’s records moving forward. Another major benefit is that these cases often offer the greatest benefit in coder education. However, targeting specific records will involve only a small sample of the facility’s overall patient population.
The benefit a random selection offers is that it will cross over the entire patient population, allowing for a broader sampling of the overall patient clinical types. These results can more readily offer a picture of overall coding quality for the payer types included. The limitation is that it can result in less complex cases falling into the sample selection, reducing identification of coder educational opportunities.
A record sampling that includes both random and targeted records from your top MS-DRGs would involve randomly selecting records from the facility’s top reported MS-DRGs by MS-DRG frequency. Errors noted and education offered would have a broader impact, as it would involve and impact a larger number of records.
The selection method would limit the charts reviewed to only the most frequently reported MS-DRGs, excluding other clinical scenarios and possibly missing issues and educational needs.
Having quarterly reviews that mix the chart selection methods from quarter to quarter allows the facility to take advantage of the benefits each selection method offers and will likely include a larger number of varying ICD-10 codes for review. Just use caution when comparing the results, ensuring to only compare results with prior audits when the sample method is the same.
About the Author
Lisa Marks, RHIT, CCS, is director of client audits for Precyse, a leader in Health Information Management (HIM) technology and services.
To comment on this article please go to email@example.com