Background: Administrative data   is a useful tool for research and quality improvement; however, validity of   research findings based on these data depends on their reliability. Diagnoses   assigned by physicians are subsequently converted by nosologists to ICD-10   codes (International Statistical Classification of Diseases and Related   Health Problems, 10th Revision). Several groups have reported ICD-9 coding   errors in inpatient data that have implications for research, quality   improvement, and policymaking, but few have assessed ICD-10 code validity in   ambulatory care databases. Our objective was to evaluate pulmonary embolism   (PE) ICD-10 code accuracy in our large, integrated hospital system, and the   validity of using these codes for operational and health services research   using ED ambulatory care databases. 

Methods: Ambulatory care data for patients (age ≥ 18 years) with a PE   ICD-10 code (I26.0 and I26.9) were obtained from the records of four urban   EDs between July 2013 to January 2015. PE diagnoses were confirmed by   reviewing medical records and imaging reports. In cases where chart diagnosis   and ICD-10 code were discrepant, chart review was considered correct.   Physicians' written discharge diagnoses were also searched using 'pulmonary   embolism' and 'PE', and patients who were diagnosed with PE but not coded as   PE were identified. Coding discrepancies were quantified and described. 

Results: One thousand, four hundred   and fifty-three ED patients had a PE ICD-10 code. Of these, 257 (17.7%) were   false positive, with an incorrectly assigned PE code. Among the 257 false   positives, 193 cases had ambiguous ED diagnoses such as 'rule out PE' or 'query   PE', while 64 cases should have had non-PE codes. An additional 117 patients   (8.90%) with a PE discharge diagnosis were incorrectly assigned a non-PE   ICD-10 code (false negative group). The sensitivity of PE ICD-10 codes in   this dataset was 91.1% (95%CI, 89.4-92.6) with a specificity of 99.9% (95%CI,   99.9-99.9). The positive and negative predictive values were 82.3% (95%CI,   80.3-84.2) and 99.9% (95%CI, 99.9-99.9), respectively. 

Conclusions: Ambulatory care data,   like inpatient data, are subject to coding errors. This confirms the   importance of ICD-10 code validation prior to use. The largest proportion of   coding errors arises from ambiguous physician documentation; therefore,   physicians and data custodians must ensure that quality improvement processes   are in place to promote ICD-10 coding accuracy.

Authors: Kristin Burles, Grant Innes, Kevin Senior, Eddy Lang, Andrew McRae