Symptom-specific ICD-10 codes lack sensitivity and fail to accurately capture many patients with relevant COVID-19 symptoms, a study published Aug. 17 in JAMA Network Open found.
For the cohort study, researchers reviewed EMR records of 2,201 patients who were tested for COVID-19 between March 10 and April 6 at University of Utah Health in Salt Lake City. The study authors sought to find whether the ICD-10 codes accurately captured symptoms of fever, cough and dyspnea among patients being tested for COVID-19.
Researchers found that the sensitivity, the ability to identify those with a condition, of ICD-10 codes for fever was 26 percent. Specificity, the ability to identify those without a condition, for fever was 98 percent. And the positive predictive value, or likelihood that those who test positive for a disease actually have it, was 96 percent. The negative predictive value, or the likelihood that those who test negative don't have the disease, was 41 percent.
Typically, the higher the percentage of the test, the more reliable it is.
For cough, sensitivity was 44 percent, specificity was 88 percent, the positive predictive value was 96 percent and negative predictive value was 18 percent. For shortness of breath, or dyspnea, sensitivity was 24 percent, specificity was 97 percent, positive predictive value was 93 percent and negative predictive value was 42 percent.
Because many data models and other aggregation tools rely heavily on ICD-10 codes to find clinical concepts and patterns, "inaccuracy has implications for any downstream scientific discovery or surveillance," the study authors wrote.
"Our findings highlight the importance of quality control in COVID-19 data aggregation, which has become increasingly important with recent high-profile journal retractions," study authors wrote. "As with many aspects of this pandemic, we must pay careful attention to socioeconomically vulnerable populations, including racial minorities, rural patients and low-income patients, for whom the gap between ICD-10 coding and clinical reality could be greater."
Read the full study here.