The Intrinsic Estimator Method for Control of Age, Period, Cohort, and Its Comparison With the Descriptive Classic Model



Sahebi R¹ ; Motevalian A² ; Sahebi L³ ; Sharifi H⁴ Show Affiliations
Source: Iranian Journal of Epidemiology Published:2017

Abstract

Background and Objectives: The aim of this study was to compare the intrinsic estimator method and the age, period and cohort accounting model used for identifying age, period, and cohort trends in incidence, prevalence, and mortality rates. Methods: The data of 2 studies Age, Period, and Cohort Effects on Alcohol and Drug use Among Students of Tehran University of Medical Sciences from 2006 to 2009 and The Role of Age, Period, and Cohort Effects on Smoking among the Students of Tehran University of Medical Science were used. First the age, period and cohort accounting model and then The IE method are explained, their advantages and disadvantages are discussed, and their results are compared. Results: Both methods showed an increasing, decreasing, and increasing trend for age, cohort, and cohort effect for the prevalence of smoking. IE could predict a preventing effect for an early age factor for smoking. In the descriptive model, a decreasing trend was seen from old cohorts to younger cohorts in nearly all age groups. However, the IE method did not show any cohort effects for alcohol consumption in female studnets. Conclusion: Currently, the IE method is the best method for solving linear dependence between three variables of age, birth cohort, and period in this type of studies. However, researchers should use it with caution because it has many pitfalls. © 2017, Iranian Epidemiological Association. All Rights Reserved.