June 2012

Document Type


Degree Name



Dept. of Public Health and Preventive Medicine


Oregon Health & Science University


Background: Evaluating surveillance data – such as notifiable disease reports and vital statistics records – over time using direct or indirect rates is a common public health tool for discerning and predicting disease trends. In turn, these assessments may help explain the etiology of health outcome and inform prevention and planning efforts. While rates by time period and age are easily calculated, they may obscure factors that influence disease or mortality risk. Age, period, and cohort (APC) analyses seek to uncover these influences by partitioning trends into components associated with changes over time within a given age group (age effects), time period (period effects), and birth cohort (cohort effects). Because available surveillance data for viral hepatitis predominantly represent prevalent disease, it is difficult to track changes in incidence over time or anticipate the magnitude of the disease burden. Recent studies of viral hepatitis infection in the United States have proposed that generational differences exist in morbidity and mortality of this disease.[superscript 1-3] Consequently, an APC analysis of viral hepatitis mortality may contribute information about factors perpetuating infection, suggest whether current trends are likely to be sustained, and guide public health planning efforts. Objective: To evaluate Oregon viral hepatitis mortality for the presence of cohort effects in the baby boomer generation – individuals born between 1950 and 1965 – by applying the multiphase method, a novel method of APC analysis[superscript 4,5] Methods: Deaths related to viral hepatitis were abstracted from multiple cause mortality variables from Oregon death certificates for 1995 to 2010 using International Classification of Diseases 9th and 10th revision codes (ICD-9: 070; ICD-10: B15-19, B94.2). These data were evaluated for the presence of cohort effects using the three stage multiphase method: (1) data were assessed using pairwise graphical inspection; (2) log-additive components of age and period were removed using a median polish;[superscript 6,7] the remaining cohort effect was separated from error using a linear regression model and its relative magnitude estimated. Results: Qualitative evidence from the first two steps of the multiphase method suggested the presence of age and cohort effects in viral hepatitis deaths. After removal of the log-additive effects of age and period, estimated viral hepatitis mortality rates remained significantly higher for individuals born 1950–1965 than preceding birth cohorts for 1910–1949 (p=0.03) or subsequent birth cohorts for 1966–1985 (p=0.003). No significant period effects occurred at the population level. Conclusion: We demonstrate a computationally straightforward method for assessing temporal trends using aggregate data. By applying this method to mortality rates for viral hepatitis, a clear pattern of increase in deaths is discernible between 1950 and 1965 compared with the cohorts before and after. These findings may contribute to public health surveillance and planning efforts.




School of Medicine



To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.