Abstract Objectives To establish and validate age- and sex-specific reference intervals (RIs) for eight common tumor markers using large-scale real-world data (RWD) and to compare the performance of five indirect statistical algorithms for RI derivation. Methods This retrospective study included 404,347 individuals undergoing health examinations at The University of Hong Kong-Shenzhen Hospital between October 2022 and December 2024. Distribution characteristics and age-related trends were evaluated using kernel density estimation and generalized additive models. After age and sex stratification, RIs for AFP, CA125, CA15-3, CA19-9, CEA, fPSA, tPSA, and the fPSA/tPSA ratio were established using the Tukey, Z-score, Gaussian mixture model (GMM), isolation forest (IF), and local outlier factor (LOF) methods. Performance was assessed by bias ratios and validated using an independent cohort. Results Most tumor markers exhibited markedly right-skewed distributions, whereas the fPSA/tPSA ratio showed only moderate skewness. AFP and CEA showed negligible sex differences. Age-specific stratification was identified for male AFP, female CA125 and CA15-3, CA19-9, CEA, and prostate-related markers. Tukey and Z-score methods showed the greatest agreement and achieved validation pass rates of ≥90 % across all demographic and analyte subgroups. In contrast, GMM and IF failed validation in 11 subgroups, and LOF failed in 6 subgroups, mainly because of the over-exclusion of high-end values. Conclusions Large-scale RWD enabled robust age- and sex-specific RIs for eight tumor markers. For highly right-skewed tumor marker data, the Tukey and Z-score methods outperformed GMM, IF, and LOF, providing reliable and practical approaches for clinical RI establishment.
Medical history taking serves a critical role in improving diagnostic safety and patient outcomes. Despite advances in technology, a patient's medical history remains the cornerstone of accurate diagnosis, often yielding diagnoses before tests are performed. This article explores how patient-centered communication, contextual awareness, and evidence-based strategies reduce diagnostic errors, which often stem from miscommunication during the clinical encounter. It highlights practical tools for clinicians and patients, addresses systemic challenges, and examines how innovations like AI and digital platforms can augment, not replace, the human connection essential to care. By reinforcing the value of thorough, empathetic history taking, this resource advocates for a future where technology and human expertise work hand-in-hand to achieve diagnostic excellence.
OBJECTIVES:Many, if not most, diagnostic errors go unrecognized by clinicians unless they cause significant patient harm. Reducing the gap between clinicians' perceived and actual diagnostic performance (i.e., improving calibration) may improve patient outcomes. However, existing approaches to measuring calibration were developed for use in experimental settings rather than realistic practice environments. We reviewed the scientific literature and solicited expert field input to improve approaches for assessing calibration in clinical practice. METHODS:We performed a targeted literature review to identify studies that compared clinicians' diagnostic performance or accuracy to perceived accuracy or confidence. Key findings from the review were provided to 10 subject matter experts in clinical reasoning and diagnostic excellence. We then interviewed these subject matter experts in a series of four small-group teleconferences. Transcripts of group discussions were analyzed for thematic content relevant to improving the measurement of calibration in real-world care settings. RESULTS:Discussions with experts highlighted that, while calibration reflects skill and expertise relevant to the task at hand, it is also fluctuating and context-dependent. Experts recommended that measurement of calibration should focus on clinically meaningful decisions that may be upstream of a final, correct diagnosis. Experts also recommended assessment of contextual factors that shape diagnostic decision-making and self-assessment. CONCLUSIONS:Measures of calibration should focus not only on outcomes but also on processes, which are often more actionable and proximal to clinical decision-making. In response to expert input, we propose a measure of real-world calibration that builds upon a previous measure to include several contextual variables.
OBJECTIVES:Discrepancies have been observed in the rerun results of plasma alkaline phosphatases (ALP) on Cobas c503 Roche analysers. The aims were to 1) determine the percentage of discrepancies in ALP reruns and 2) evaluate the effect of an automatic dilution on discrepancies. METHODS:1) Raw plasma ALP results from patients were extracted from Cobas pro c503 analysers over a 2.7-year period. Reruns were triggered for any ALP above 200 U/L with no history or with a previous normal ALP, and those with a variation exceeding the total allowable error (±15 U/L if ≤150 U/L or ±10 % if >150 U/L) were classified as discrepancies. 2) During five weeks, a systematic 1:5 dilution was triggered on the analysers for all pure tests, concomitantly. RESULTS:1) Among 9,574 reruns, 677 (7.1 %) were discordant (5.8 %, 1.2 % and 0.1 % after one, two and three reruns, respectively), and 90 % of which were decreases. 2) 9,062 pairs of concomitant pure and diluted ALP were measured, with respective medians (2.5th-97.5th percentiles) of 94 (43-559) and 90 (41-528) U/L (median of differences -4.4 %). Of these, 274 were retested, of which 16 pure reruns (5.8 %) were discordant, compared to only two diluted reruns (0.7 %). There were greater negative variations in ALP reruns as the leukocyte count increased (p<0.001). CONCLUSIONS:A substantial rate of discrepancies was observed when elevated ALP were retested on Cobas c503 analysers. This issue was almost eliminated by applying a first-line 1:5 dilution. A modification of the assay method is eagerly awaited from the manufacturer.
