Generating Comorbidity Indices and Determining Their Association with Socioeconomic Status

About the Author

Hi! My name is Cassie and I’m studying computational and systems biology. As someone involved in politics and interested in social justice and health sciences, I am exploring pursuing a career in epidemiology or public health in order to create a safer, healthier, and more equitable community. In my free time I like to play softball, volleyball, and do art as a hobby.


This experience working with Dr. Lacey, Dr. Wang, and the California Teachers Study team has introduced me to the field of epidemiological research and data analysis. I’ve learned about the daily life of a researcher, some of the basic computational and statistical tools we can use to analyze data, and the overarching steps to creating a research project. I’m extremely grateful to have had this opportunity and look forward to continuing work in the epidemiological field.


My Research Project

In this project, my mentor, Dr. Lacey, and I used California Teachers Study data to investigate changes in Charlson Comorbidity Index (CCI) over time. The Charlson Comorbidity Index is a weighted index of 16 comorbid conditions that can be used to predict risk of mortality. We also evaluated the association between socioeconomic status (as measured by healthcare insurance status, highest level of education, annual household income, and retirement status) and CCI over time, and whether those associations differed before, during, and after the Great Recession. The Great Recession was a global financial crisis that took place from December 2007 to June 2009.


We used data from the subset of the CTS population that filled out Questionnaire 4 in 2005-2006, which included information about socioeconomic status, and we linked hospitalization data from California’s Office of Statewide Health Planning and Development (OSPHD) to generate an aggregate CCI value and analyze its association with socioeconomic status.


The data demonstrated that within our subset of the CTS population, there was an increase in CCI over time. Our preliminary results from the univariate analyses, which analyzed the relationships between these variables one at a time, indicated that older age, lower annual household income, being retired, lower levels of education, and being on Medicare were each statistically significantly associated with higher CCI values.


However, when we analyzed these variables together in our multivariate analysis, the findings indicated that age had the strongest association with increases in CCI. In 2005, CTS participants’ CCI ranged from 0 to 4 or more. The chart below shows CCI by age group for participants age 50 and younger compared with participants older than 80.



Education and retirement status were not strongly associated with increases in CCI. Compared with participants on Medicare, participants with managed-care insurance plans, which involve Health Maintenance Organizations to manage an individual’s healthcare, were less likely to have an increase in their CCI. Lower income was more strongly associated with an increase in CCI during the Great Recession compared with the time periods before or after the Great Recession.


Future Goals

I would like to further study these relationships between health insurance status, annual income, and CCI over time in other study populations. I would also like to analyze the impact of the Great Recession on healthcare utilization in the California Teachers Study population, and how that relates to CCI.