Projected Impact of Predefined Reflex Testing Algorithms on Thyroid Function and Prostate-Specific Antigen Testing: A Real-World Laboratory Stewardship Study

Background/Objectives: Reflex testing may improve laboratory stewardship by limiting downstream analyses to situations in which additional testing is supported by the result of an initial test. This study evaluated the projected impact of predefined reflex testing algorithms on thyroid function and prostate-specific antigen (PSA) test utilization and laboratory expenditure. Methods: This observational modelling study included clinician-initiated thyroid function and PSA testing sessions recorded at a tertiary-care hospital in Türkiye during 2022. Clinician-requested tests and corresponding laboratory results were compared with predefined reflex algorithms applied retrospectively at the testing-session level. For thyroid testing, thyroid-stimulating hormone (TSH) was used as the initial test; free thyroxine (FT4) was projected for elevated TSH, and both FT4 and free triiodothyronine (FT3) for decreased TSH. For PSA testing, free PSA was projected only when total PSA was 4.0–10.0 ng/mL. Observed clinician-initiated testing was compared descriptively with projected algorithm-based testing, including downstream test utilization and laboratory expenditure. Results: The thyroid cohort comprised 60,656 testing sessions from 43,839 patients, and the PSA cohort comprised 4583 sessions from 3631 patients. In the thyroid cohort, clinicians requested 49,858 FT4 and 19,899 FT3 measurements. Under the predefined algorithm, these numbers were projected to decrease to 6205 and 2478, respectively, reducing total thyroid test volume from 130,413 to 69,339 tests, a projected reduction of 61,074 tests (46.8%). In the PSA cohort, free PSA testing was projected to decrease from 2133 to 642 measurements, reducing total PSA-related test volume from 6716 to 5225 tests (22.2%). Projected laboratory expenditure decreased from US dollars (USD) 73,579.01 to USD 39,121.06 for thyroid testing and from USD 6468.78 to USD 4560.73 for PSA testing. Across both pathways, the projected saving was USD 36,366.00, corresponding to a 45.4% reduction in expenditure. Conclusions: Retrospective application of predefined reflex testing algorithms to clinician-initiated thyroid function and PSA requests projected substantial reductions in downstream test utilization and laboratory expenditure. Prospective implementation studies incorporating patient-specific clinical information are needed to determine whether these projected reductions can be achieved safely in routine clinical practice without limiting clinically appropriate testing.