Valid estimates for the number of SARS-CoV-2 infections is imperative for assessing the impact of the COVID-19 pandemic within specific populations. Here, we discuss ongoing efforts aimed at understanding the state of the pandemic in two different contexts. First, we focus on seven low- and middle-income countries where COVID-19 testing has been limited. We propose using aggregated health systems data to perform syndromic surveillance and detect potential outbreaks. Second, we focus locally on the City of Holyoke, Massachusetts where testing is readily available, but racial and ethnic disparities in testing may obscure the toll of COVID-19 in historically marginalized communities. Specifically, progress is limited by: (1) missing information on race and ethnicity in the testing data and (2) selection bias resulting from access to testing. We provide a discussion on statistical approaches that can account for these complexities.