S21: When the Answer is “Big(ger) Data” in Environmental Epidemiology: What are the Questions?, Room 315, Floor 3, August 26, 2019, 1:30 PM - 3:00 PM As discussed in other talks in this session, pooling multiple generally similar studies allows more flexibility in data analysis compared to meta-analysis methods but requires careful attention to details of harmonization. Though the design of pooling projects may consider similarities in available data when selecting cohorts for inclusion, it is nearly inevitable that harmonization of exposures, outcomes, and covariates pose specific challenges. This talk will discuss approaches to harmonization for cohort studies, acknowledging the tradeoffs with each choice made. Even when cohorts use the same methods and tools to collect data, there are differences in cohort design including differences in time periods data were collected, differences in locations of participants, and meaningful differences in how environmental exposures differ over time and space. Harmonization of outcome measures may pose additional challenges, particularly due to changes over time in both standard definitions of health endpoints and data collection methods. For example, blood pressure may be measured, taken from medical records, a self-reported value or self-reported as “high”. It may be reported at baseline only, or reported during follow-ups. Also, standard definitions of high blood pressure have changed over time. When considering how to harmonize all of these aspects must be considered. Most existing pooling projects have harmonized covariates by attempting to create unified variables; is this sufficient? What other considerations should we be taking into account, particularly in environmental epidemiology? Are there other methodological approaches that can be considered? An open discussion of the advantages and disadvantages to different approaches to harmonization will help guide future pooling projects.