A system for supporting expert decision-making in the medical and social examination of children with cerebral palsy

Authors:

Mikhailishin Viktor Valerievich – Junior Researcher at the Laboratory of Innovative and Expert Rehabilitation Technologies, Albrecht Federal Scientific and Educational Centre of Medical and Social Expertise and Rehabilitation, 50 Bestuzhevskaya Street, 195067 St. Petersburg, Russian Federation; e-mail: doompro@mail.ru; https://orcid.org/ 0000-0002-9518-1945.

Golovina Yulia Aleksandrovna – PhD student, А1ЬгссЫ Federal Scientific and Educational Centre of Medical and Social Expertise and Rehabilitation, 50 Bestuzhevskaya Street, 195067 St. Petersburg, Russian Federation; e-mail: doc.yuliagolovina@gmail.com; https://orcid.org/ 0009-0008-3903-2241.

Ponomarenko Gennadiy Nikolaevich — Corresponding Member of the Russian Academy of Sciences, Honored Scientist of the Russian Federation, Grand PhD in Medical sciences (Dr. Med. Sci), Professor, Director General of the Albrecht Federal Scientific and Educational Centre of Medkal and Social Expertise and Rehabilitation, 50 Bestuzhevskaya Street, 195067 St. Petersburg, Russian Federation; Head of the Department of Physical and Rehabilitation Medicine of the North-Western State Medical University named after I.I. Mechnikov, 47 Piskarevskiy Avenue, 195067 St. Petersburg, Russian Federation; e-mail: ponomarenko_g@mail.ru; https://orcid.org/0000-0001-7853-4473.

Bolshakov Vladimir Aleksandrovich – senior researcher of the design department of the Institute of Prosthetics and Orthotics, Albrecht Federal Scientific and Educational Centre of Medical and Social Expertise and Rehabilitation, 50 Bestuzhevskaya Street, 195067 St. Petersburg, Russian Federation; e-mail: pko09_903@mail.ru; https://orcid.org/ 0000-0002-5889-375

Annotation:

Introduction. Cerebral palsy is the most common cause of primary disability in children with diseases of the nervous system. Due to the digitalization of the medical and social sphere in the Russian Federation, an urgent task is to develop a system to support medical decision-making based on artificial intelligence technologies to improve the efficiency of medical and social expertise.

Aim. To develop and evaluate an expert artificial intelligence system to assist an expert in conducting a medical and social examination of children with cerebral palsy.

Materials and methods. The study used a database containing information about the clinical and functional status of patients from the acts and protocols of the medical and social examination of 1163 children with cerebral palsy. Crossvalidation was used to assess the quality of the models, and the following metrics were used: the proportion of correct outcomes (accuracy), accuracy (precision), sensitivity (recall), f-measure. The target features included the maximum degree of severity of persistent violations of body functions, the degree of restriction of the main categories of human activity, the category of “disabled child”. The quality assessment of the artificial intelligence system was carried out in accordance with GOST R 59898- 2021.
The substantiation of the target features for use in the decision support system has been carried out. A comparative analysis of artificial intelligence algorithms in the classification problem on this data set is performed. Mathematical models predicting target features are constructed and evaluated. An expert artificial intelligence system has been developed and evaluated to assist an expert doctor in examining children with cerebral palsy.

Results. Based on the results of the comparative analysis, a classifier based on the Random forest algorithm was selected. The developed models for determining the maximum severity of persistent disorders of body functions and the degree of restriction of the main categories of vital activity showed average and high results. An artificial intelligence system has been developed, the proportion of correct outcomes in establishing the category of “disabled child” by the artificial intelligence system was 0.95. The integral quality indicator of the developed system was 0.944.

Discussion. According to the results of the analysis, the best quantitative metrics for evaluating the quality of the model when determining target variables were demonstrated by classifiers based on the Random forest algorithm. Models for determining the maximum severity of persistent disorders of body functions and the degree of disability with its use showed average and high quality of work.

Conclusion. The developed artificial intelligence system, due to the high quality metrics of the mathematical models underlying it and the high integral indicator of the quality of its work, is applicable in the practice of expert doctors.

List of cited literature:

1. Bax MCO. Terminology and classification of cerebral palsy. Developmental Medicine & Child Neurology. 1964; 6(3): 295-7.

2. Vovchenko AV, Fisenko VS, Gribov DE, et al. Sostoyanie i dinamika invalidnosti, kompleksnaya reabilitatsiya i abilitatsiya invalidov i detei-invalidov v Rossiiskoi Federatsii: ezhegodnyi doklad (pod red. M.A. Dymochki). M.: FGBU FB MSE Mintruda Rossii [The state and dynamics of disability, comprehensive rehabilitation, and habilitation of disabled persons and children with disabilities in the Russian Federation: annual report (edited by Dymochka MA). Moscow: Federal Bureau of Medical and Social Expertise], 2022. 436 p. (In Russian).

3. Gombolay GY et al. Review of machine learning and artificial intelligence (ML/AI) for the pediatric neurologist. Pediatric Neurology. 2023; 141: 42-51.

4. Duran I et al. Artificial intelligence to improve efficiency of administration of gross motor function assessment in children with cerebral palsy. Developmental Medicine & Child Neurology. 2022; 64(2): 228-34.

5. Golovina YA, Mikhailishin VV, Ponomarenko GN, Shcherbina KK, et al. Database of clinical-functional indicators in children with cerebral palsy during medical and social expertise. Available at: https:// www1.fips.ru/ofpstorage/Doc/PrEVM/RUNW DB/000/002/024/621/653/2024621653-00001/ DOCUMENT.PDF. (accessed 20.06.2024). (In Russian).

6. Golovina YA, Mikhailin VV, Ponomarenko GN. Prediktory razvitiya stoikikh narushenii funktsii organizma u detei s detskim tserebral’nym paralichom [Predictors of the development of persistent disorders of body functions in children with cerebral palsy]. Fizioterapevt [Physiotherapist]. 2024;20(4):99-107. (In Russian).

7. Browne MW. Cross-Validation Methods. J Math Psychol. 2000;44(1):108-32. DOI: 10.1006/ jmps.1999.1279.

8. Klassifikatsii i kriterii, ispol’zuemye pri osushchestvlenii mediko-sotsial’noi ekspertizy [Classifications and criteria used in the implementation of medical and social expertise]. Prikaz Mintruda Rossii ot 27.08.2019 N 585n [Order of the Ministry of Labor of Russia of August 27, 2019 N 585n. Available at: https:// mintrud.gov.ru/docs/mintrud/orders/1386. (accessed 20.06.2024). (In Russian).

9. O sotsial’noi zashchite invalidov v Rossiiskoi Federat-sii [On the social protection of disabled people in the Russian Federation]. Federal’nyi zakon ot 24 noyabrya 1995 g. № 181-FZ. [Federal Law of November 24, 1995 N 181-FZ]. Available at: https://mintrud.gov.ru/docs/ laws/75. (accessed 20.06.2024). (In Russian).

10. O priznanii litsa invalidom [On the recognition of a person as disabled]. Postanovlenie Pravitel’stva Rossiiskoi Federatsii ot 5 aprelya 2022 g. N 588 [Decree of the Government of the Russian Federation of April 5, 2022 N 588]. Available at: http://government.ru/ docs/all/140380. (accessed 20.06.2024). (In Russian).

11. Li S, et al. Understanding and addressing quality attributes of microservices architecture: A Systematic literature review. Information and software technology. 2021;131:106449.

12. GOST R 59921.0-2022. Sistemy iskusstvennogo intellekta v klinicheskoi meditsine. M.: Standartinform [Artificial intelligence systems in clinical medicine. Moscow: Standartinform], 2023. 12 p. (In Russian).

13. GOST R 59277-2020 [Classification of artificial intelligence systems. Moscow: Standartinform], 2023. 16 p. (In Russian).

14. GOST R 59898-2021. Otsenka kachestva sistem iskusstvennogo intellekta. M.: Standartinform [Quality assessment of artificial intelligence systems. Moscow: Standartinform], 2023. 24 p. (In Russian).

15. O personal’nykh dannykh [On Personal Data]. Federal’nyj zakon ot 27.07.2006 N 152-FZ (red. ot 06.02.2023) [Federal Law N 152-FZ of July 27,2006 (edition of February 6, 2023)]. Available at: https://base.garant. ru/12148567. (accessed 14.06.2024). (In Russian).