ICD codes poorly correlate with skin cancer incidence in individual patients

CPT codes rather than ICD codes may be a more effective proxy to determine the actual number of skin cancers experienced by an individual patient, according to a study.

“Patients can develop multiple skin cancers, and their medical data can be spread over multiple health care systems,” Nimay Anand, BA, of the school of medicine at Meharry Medical College in Nashville, Tennessee, and colleagues wrote. “This fragmented care, combined with the lack of skin cancer registries, has limited our ability both to provide accurate estimates of incidence and to study the pathogenesis of multiple skin cancers.”

In the retrospective cohort study, the researchers evaluated whether standard diagnostic and procedural codes found in electronic health records serve as a valid proxy to estimate overall skin cancer incidence. All patient data were culled from Vanderbilt University Medical Center from July 1, 2008, to June 30, 2018.

Eligibility criteria included at least one ICD or CPT code and at least one pathology report for any skin cancer found in the EHR.

Participants had a mean age of 70.4 years. The cohort was 56.8% men and 88.1% white.

Overall, 130,285 clinical codes and 29,041 biopsy-confirmed skin cancers among 35,901 patients were included in the analysis.

Results showed that 6,307 patients had at least one ICD or CPT code or pathology report for a skin malignancy; 5,688 had both a relevant CPT code and a histologically verified skin malignancy, and 619 had at least one ICD code and at least one confirmed malignancy but no CPT code.

The researchers observed a strong linear correlation between the number of CPT codes and pathology records (r = 0.87). Conversely, ICD codes and pathology records poorly correlated (r = 0.22).

For the 619 patients without CPT codes, the correlation between ICD codes and pathology reports was very poor (r = 0.06).

The researchers noted that CPT codes accounted for more than 75% of the variation in the overall number of skin cancer cases described in the analysis (R² = 0.76).

“This cohort study found that the use of ICD codes was a poor proxy measure for the number of skin cancers per patient,” the researchers wrote. Rather, they suggested that CPT codes are an “adequate proxy” for the number of malignancies per patient.