An attempt to simplify clinical diagnostics of poisoning in the field using the binary discrimination method

Introduction. The lack of analytical determination of toxic chemicals in biological environments and environmental objects complicates the assessment of the prognosis of the poisoned person’s condition and the choice of medical tactics at the advanced stages of medical evacuation. This problem is extremely relevant in cases of poisoning with fast-acting substances in conditions where therapeutic measures are severely limited in time. Taking this into account, an attempt was made to develop a diagnostic algorithm based on the mathematical method of binary discrimination of objective signs of poisoning.

Material and methods. Tables of 56 bivariantly manifesting (yes/no) signs of 89 etiological types of poisoning were compiled taking into account national guidelines and the authors’ experience in the presence of analytical confirmation of the diagnosis. Using the discriminant binarization method, sets of 11 signs were selected that unambiguously determine all types of poisoning and a decision rule was compiled for their express diagnostics at the advanced stages of medical evacuation.

Results. Assessments of the presence or absence of objective manifestations of intoxication in the format of a reduced matrix made it possible to formulate a version of a decision rule for diagnosing poisoning with the most common types of military, industrial and medicinal toxicants, modern narcotic and psychotropic drugs, and substances of natural origin. The compiled algorithm is currently being tested to determine the objective characteristics of prognostic significance in the context of providing medical care of unknown structure toxic agent intoxication.

Limitations. Binary discrimination acceptability for data objectification on a probable toxic chemical class does not apply to home foodborne intoxications and iatrogenies, and their methodic interpretabilities are limited by preventive prescription of symptomatic therapeutics.

Conclusion. The method of multiple binary discrimination of alternatively expressed features is applicable in the formation of express assessments of the state and prognosis of the course of the disease allows reducing redundant information and compiling primary diagnostic algorithms applicable until analytical confirmation of the etiology of poisoning is available.

Compliance with ethical standards. The study does not require the provision of an opinion from the biomedical the ethics committee.

Author contribution:
Chepur S.V. – article idea, writing and editing the text;
Vinnik P.M. – mathematical solution of the applied problem;
Yudin M.A., Mosin A.V. – compilation of the matrix based on the clinical material;
Kobelev M.V, Kraenkov M.S. – numerical solutions.
All co-authors – approval of the final version of the article, responsibility for the integrity of all parts of the article.

Conflict of interest. The authors declare no conflict of interest.

Funding. The study had no sponsorship.

Received: October 17, 2024 / Accepted: July 14, 2025 / Published: August 29, 2025

1. Nazarov V.B., Gladkikh V.D., Boyarintsev V.V., Samoylov A.S., Belovolov A.Yu. Current problems of forning reserves of antidote therapy funds for eliminating medical and sanitary consequences of emergency situations. Toksikologicheskiy vestnik. 2011; 6(111): 33–7. (in Russian)

2. The Encyclopedia of Poisons and Antidotes. Ed. Turkington C., Mitchell D. New York, 2010; 324.

3. Savelyeva Ye.I. Application areas of bioanalytical chromatograph Y-mass spectromery. Zhurnal analiticheskoy khimii. 2021; 76(10): 937–51. (in Russian)

4. Medical toxicology. National guide [Meditsinskaya toksikologiya: nachionalnoe. rukovodstvo]. Pod red. Luzhnikova Ye.A. Moscow: GEOTAR-Media, 2014; 928. (in Russian)

5. Thomsen K., Christensen A.L., Iversen L., Lomholt H.B., Winther O. Deep Learning for Diagnostic Binary Classification of Multiple-Lesion Skin Diseases. Front. Med. (Lausanne). 2020; 7: 574329.

6. Marcin T., Hautz S.C., Singh H., Zwaan L., Schwappach D., Krummrey G., et al. Effects of a computerised diagnostic decision support tool on diagnostic quality in emergency departments: study protocol of the DDx-BRO multicentre cluster randomised cross-over trial. BMJ Open. 2023; 13(3): e072649.

7. Emmert D., Szczypien N., Bender T.T.A., Grigull L., Gass A., Link C. et al. A diagnostic support system based on pain drawings: binary and k-disease classification of EDS, GBS, FSHD, PROMM, and a control group with Pain2D. Orphanet. J. Rare Dis. 2023; 18(1): 70.

8. Rebrova O.Yu., Gusev A.V. Calculation of the sample size for clinical trials of medical decision support systems with binary response. Sovremennye tekhnologii v meditsine. 2022; 14(3): 6–14. (in Russian)

9. Nelson L.S., Howland M.A., Lewin N.A., Smith S.W., Goldfrank L.R., Hoffman R.S., et al. Goldfrank’s Toxicologic Emergencies. Ed. by N.E. Flomenbaum. 11 ed. New York – Chicago – San Francisco: MacGraw Hill Education, 2019, 2070.

10. Li Y., Lu F., Yin Y. Applying logistic LASSO regression for the diagnosis of atypical Crohn’s disease. Sci. Rep. 2022; 12(1): 11340.