Using Analytics to Reduce Unspecified Diagnoses


While most physician practices and health systems implemented ICD-10 with only minor snares along the way, some groups did report a decline in productivity as coders rushed to become familiar with the new code set. Fortunately, the specialty of radiology only uses about 4 percent of the 68,000 codes within the ICD-10 set, significantly narrowing the learning curve for radiologists and radiology-focused coders.

The year-long grace period offered by the Centers for Medicare & Medicaid Services (CMS) served as another aid in the transition. For many payers, the use of unspecified ICD-10 diagnosis codes would lead to many claim denials for failing to meet medical necessity or the payer’s specific policy for the procedure performed. For this reason, CMS gave physicians and coders one year to become familiar with the new ICD-10 coding set and the greater level of specificity it required.

As of Oct. 1, 2016, however, that grace period came to a close. If there is an established policy in place (Medicare National Coverage Determinations (NCD) or Local Coverage Determinations (LCD)), Medicare may now deny certain unspecified diagnoses. If other payers offered a similar grace period, they also may follow suit and deny these claims.

To mitigate the impact this change may have on practice revenue and coder productivity, it is critical to develop and implement an advanced analytics platform to track and cut down on unspecified codes. Many industry leaders have already realized the potential of analytics to generate positive changes in healthcare: In a 2012 executive report titled “The value of analytics in healthcare,” the IBM Institute for Business Value states: “Using analytics to gain better insights can help demonstrate value and achieve better outcomes, such as new treatments and technologies. Information leading to insight can help informed and educated consumers become more accountable for their own health. Analytics can improve effectiveness and efficiency.”

Analytics can help to significantly reduce the number of unspecified diagnoses and therefore protect practices’ cash flow. Radiology groups should aim for unspecified codes to account for less than 10 percent of gross charges, keeping in mind there are times when radiologists do not have additional information regarding the patient’s condition or medical history, and that there are a few unspecified ICD-10 diagnoses that do not provide a more specific diagnosis option.

Achieve a 360-degree practice view
An advanced analytics platform that can leverage both structured and unstructured data can help physician practices better prepare for life in a post-grace period world. Because between 60 and 70 percent of all radiology exams are normal, it is critical for radiologists to obtain full clinical history to optimize the quality of care and minimize the use of unspecified codes.

Analytics are able to filter data by provider, modality, payer, location, place of service and referring provider, granting physician leaders and administrators a 360-degree view of their coding performance. These types of platforms can also help highlight areas of improvement down the line.

Gain recommendations to reduce unspecified codes
Analytics platforms can also sort the top 10 or 25 unspecified ICD-10 codes used by practices and make helpful recommendations to improve dictation and coding while reducing denials. These unspecified diagnoses are often missing anatomical- or morbidity-related information, such as laterality, body location and severity.

For example, the I65.29 code for occlusion and stenosis of unspecified carotid artery merely requires providers to document the affected carotid arteries (right, left or bilateral). By adding this piece of information, physicians can reduce claim denials and help drive uninterrupted cash flow.

While ICD-10 did not cause the turmoil many industry leaders anticipated, it did increase the workload of physicians and coders alike. To learn more about the power of analytics to reduce unspecified ICD-10 codes and increase practice and coder productivity, please email us at