Group health risk parameters in a heterogeneous cohort. Indirect assessment as per events taken in dynamics
Автор: Obesnyuk V.F.
Журнал: Анализ риска здоровью @journal-fcrisk
Рубрика: Профилактическая медицина: актуальные аспекты анализа риска здоровью
Статья в выпуске: 2 (34), 2021 года.
Бесплатный доступ
The present work focuses on describing a procedure for assessing intensive and cumulative parameters of specific risk when observing cohorts under combined exposure to several external or internal factors. The research goal was to reveal how to use well-known heuristic-descriptive parameters accepted in remote consequences epidemiology for analyzing dynamics of countable events in a cohort; analysis should be performed on quite strict statistic-probabilistic grounds based on Bayesian approach to explaining conditional probabilities that such countable events might occur. The work doesn’t contain any new or previously unknown epidemiologic concept or parameters; despite that, it is not a simple literature review. It is the suggested procedure itself that is comparatively new as it combines techniques used to process conventional epidemiologic information and a correct metrological approach based on process description. The basic result is providing a reader with understanding that all basic descriptive epidemiologic parameters within cohort description framework turn out to be quantitatively interlinked in case they are considered as conditional group processes. It allows simultaneous inter-consistent assessment of annual risk parameters and Kaplan-Meier (Fleming-Harrington) and Nelson-Aalen cumulative parameters as well as other conditional risk parameters or their analogues. It is shown that when a basic descriptive characteristic of cumulative parameters is chosen as a measure for measurable long-term external exposure, it is only natural to apply such a concept as a dose of this risk factor which is surrogate in its essence. Operability of the procedure was confirmed with an example. The suggested procedure was proven to differ from its prototype that previously allowed achieving only substantially shifted estimates, up to ~100 % even in case an operation mode was normal. Application requires creating specific but quite available PC software.
Risk, parameter, epidemiology, risk factor, competition, indirect estimate, mortality, process, cohort, strata, model
Короткий адрес: https://sciup.org/142229592
IDR: 142229592 | DOI: 10.21668/health.risk/2021.2.02