Generating Evidence Based Interpretation of Hematology Screens via Anomaly Characterization
Gil David*, 1, Larry Bernstein*, 2, Ronald R. Coifman1
Identifiers and Pagination:Year: 2011
First Page: 10
Last Page: 16
Publisher Id: TOCCHEMJ-4-10
Article History:Received Date: 25/11/2010
Revision Received Date: 28/12/2010
Acceptance Date: 5/11/2011
Electronic publication date: 1/3/2011
Collection year: 2011
open-access license: This is an open access article distributed under the terms of the Creative Commons Attribution 4.0 International Public License (CC-BY 4.0), a copy of which is available at: https://creativecommons.org/licenses/by/4.0/legalcode. This license permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Introduction: We propose a simple, workable algorithm that provides assistance for interpreting any set of data from the screen of a blood analysis with high accuracy, reliability, and inter-operability with an electronic medical record. This has been made possible at least recently as a result of advances in mathematics, low computational costs, and rapid transmission of the necessary data for computation.
Materials and Methods: The database used for this study is a file of 22,000 laboratory hemograms generated by two Beckman-Coulter Gen-S analyzers over a two month period in a 630 bed acute care facility in Brooklyn. All control samples, patient identifiers, and patients under 23 years old were stripped from the dataset. An experienced medical practitioner reviewed all of the data used in generating the algorithm described. The differential diagnoses were outlined prior to beginning the study, and preliminary studies were done to determine the reference ranges for each predictor. An algorithm for anomaly detection and classification via anomaly characterization is proposed. For each patient, the algorithm characterizes its anomalous profile and builds a differential metric to identify similar patients who are mapped into a classification.
Results: The algorithm successfully classified patients into the diagnosis that were sufficient in sample size, and others are still under observation. The algorithm correctly classified the patients as follows: Microcytic Anemia - 99.63%, Normocytic Anemia - 98.03%, Mild SIRS - 73.42%, Thrombocytopenia - 99.52%, Leukocytopenia - 84.83%, Moderate / Severe SIRS - 96.69% and Normal - 93.18%.
Discussion: This limited analysis of automated hematological results can be extended to the case of more complicated conditions than presented, and can be extended to a combination of chemistry, hematology, immunology, and other data.