Clinical Reasoning as the Guiding Framework in Medical Biochemical Test Reporting

Clinical biochemical reports should reflect the clinical perspective, be well organized and biologically oriented, and present accurate reference values based on clinically relevant distributions. To illustrate possible quality issues, we selected Clinical Cases from The New England Journal of Medicine as it publishes comprehensive reports to teach internationally relevant high-level workups. We evaluated their organization with a system organ grouping and the consistency of reference values between comparable cases. We found that the ordering of tests varied across issues, and associated interpretative tests were not organized in a scholarly manner. In several cases, the reference values for hemoglobin did not align with WHO recommendations, the upper reference value for glycohemoglobin did not correspond to established guidelines, and the upper reference value for creatinine did not reflect internationally standardized methods. The ferritin reference values in young women were not adjusted for the postmenopausal age group. When searching for incident blood loss using a common reference in young women and those of postmenopausal age in a large primary health care cohort, the frequency of flagged low ferritin results was so low that the ratio of flagged results to the incidence rate of colorectal cancer was only 1.1 (0.95 confidence interval 1.0 to 1.3). In men, low ferritin flagging was far more prevalent, resulting in a ratio of 11.7 (0.95 confidence interval 11.2 to 12.1). Thus, men had an 11-fold higher theoretical relative sensitivity when detecting occult iron loss compared with women. System organ assessment is a basic clinical competence, and system organ classification for clinical biochemical tests could be used to organize clinical laboratory reports. The reference values provided should be biologically based, harmonized, and readily recognizable by the clinician to accurately guide decisions, in contrast to local procedural determinations. Examination of local laboratory data may reveal biased laboratory support.