Impact of pre-analytical variables on Point-of-Ccare CBC accuracy in hospital wards

Thesis Abstract

The reliability of point-of-care complete blood count (POC-CBC) testing in hospital wards is increasingly pivotal for timely clinical decision-making, yet pre-analytical variables threaten its accuracy and may lead to erroneous patient management. This study investigates how pre-analytical factors—specimen collection timing, appropriateness of anticoagulant, tourniquet duration, sample handling, transport conditions, and processing time—affect POC-CBC results compared with central laboratory measurements. The aim is to quantify the impact of these variables on key CBC parameters (white blood cell count, hemoglobin, hematocrit, platelets, and red cell indices) and to identify modifiable contributors to discrepancy between POC and standard laboratory results. Specific objectives are to (i) determine the prevalence of pre-analytical errors in ward-based blood sampling, (ii) assess the agreement between POC-CBC and central lab CBC under varying pre-analytical conditions, (iii) quantify the magnitude of bias introduced by each variable using Bland-Altman analyses, (iv) explore staff knowledge and practices related to pre-analytical steps through targeted surveys, and (v) develop evidence-based recommendations to minimize pre-analytical error. The study adopts a prospective, cross-sectional design conducted across three tertiary hospital wards over a six-month period. A total of 750 paired samples will be collected from adult inpatients requiring CBC testing, with strict inclusion criteria ensuring representative case-mix and exclusion of samples with hemolysis or visible clots. Data collection will involve (i) POC-CBC results obtained with a validated handheld analyzer operated by trained ward nurses, (ii) parallel central laboratory CBC results using standard hematology analyzers, and (iii) a structured observational checklist and time-stamped logs to capture pre-analytical variables such as collection site, venipuncture technique, tourniquet use, sample volume, anticoagulant type and ratio, transport time, ambient temperature during transport, and time to analysis. Additional data will include patient demographics, clinical indication for testing, and ward-specific workflow characteristics. Instruments will be pilot-tested for validity and reliability, with inter- and intra-rater reliability assessments for observational data. Statistical analysis will incorporate descriptive statistics to quantify pre-analytical error prevalence, and inferential methods to evaluate agreement and bias. Paired comparisons will use Intraclass Correlation Coefficients (ICC) and Bland-Altman plots to assess agreement between POC-CBC and central lab results across stratified pre-analytical conditions. Multivariate linear regression and mixed-effects models will quantify the contribution of individual pre-analytical variables to discrepancies in CBC parameters, adjusting for patient-related covariates. Regression diagnostics will guide model refinement. Subgroup analyses will examine parameter-specific susceptibility to pre-analytical influences. The study will be interpreted through the lens of the Theory of Planned Behavior to contextualize staff practices and potential barriers to optimal pre-analytical care, supplemented by a systems perspective grounded in Donabedian’s structure-process-outcome framework. Expected findings include (1) measurable discordance between POC-CBC and central lab CBC associated with prolonged transport time (>30 minutes), inadequate tourniquet duration (<60 seconds or >120 seconds), improper blood-to-anticoagulant ratio, and delays from collection to analysis; (2) statistically significant biases in white blood cell counts and platelet counts under suboptimal handling, with narrower limits of agreement in controlled pre-analytical conditions; (3) identification of ward-specific workflows contributing to error clusters; and (4) evidence that staff training and procedural checklists mitigate discrepancies. The study anticipates a moderate to high agreement (ICC >0.75) under ideal pre-analytical conditions, but poorer agreement (ICC <0.60) with common pre-analytical lapses. The anticipated contribution to knowledge includes quantification of the relative impact of discrete pre-analytical factors on POC-CBC accuracy, development of a validated pre-analytical risk model, and practical recommendations for ward-based protocols, including standardized sample handling, transport timetables, and decision rules for when confirmatory central lab testing is warranted. The main conclusion will emphasize the critical role of controlled pre-analytical processes in maintaining POC-CBC reliability and patient safety. Recommendations will target hospital policy updates, staff training curricula, implementation of pre-analytical checklists, and periodic auditing of ward-based CBC practices to sustain quality improvement.