Search

Additional file 3. Histogram for previous opioid prescriptions.

Additional file 2. Logistic regression results for non-opioid prescriptions.

Introduction: Willingness to participate in COVID-19 testing and tracing programs has potential to exacerbate health disparities due to systematic variation by social determinants of health. This study evaluates sociodemographic, economic, and psychological factors associated with COVID-19 testing motivations.Methods: We used data from the Person to Person Health Interview Study, collected by telephone in April and May 2020, using a representative sample of Indiana residents (n=943). Testing intentions were assessed across six motivations: treatment affordability, ability to work if positive, hospital effectiveness, symptom severity, proximity to infected, and risk of transmitting to others. Bivariate analyses and ordinal logistic models were used to assess the influence of race/ethnicity, education, financial strain, and COVID-19 worries on each outcome. Results: Respondents strongly agree that symptom severity, proximity to infected, and risk of transmitting to others would be important considerations in COVID-19 test decisions. Black and Latino respondents were more likely than white respondents to agree that ability to work if positive (b = 0.48 and 0.96) and hospital effectiveness (b = 0.92 and 0.96) would influence their decision to get tested. Test affordability (b = 2.80), hospital effectiveness (b = 2.00), and risk of transmission (b = 0.99) were considerations among participants reporting financial strain. Desire to reduce transmission (b = 1.49) and concern about proximity to infected (b = 2.66) were reported among those who expressed COVID-19 worries.Conclusions: In a representative sample in Indiana, people consider their own and others’ vulnerability to COVID-19 in testing decisions. Lower-SES groups and Black and Latino people were more likely to cite resource constraints. Policies expanding financial assistance and prosocial message framing may increase testing rates, particularly among vulnerable subpopulations.

Social network analysts typically study node prominence in bipartite social networks with methods that do not fully account the two-mode topology of the underlying social network. Recent studies have generated a number of algorithms that might estimate node prominence in bipartite graphs with greater accuracy; however, it is unclear how these algorithms perform on real world social networks. To provide a comparison of bipartite centrality algorithms and recommendations regarding their appropriate use cases, this article examines their performance in a real-world network of patient-provider ties. We build on a recent body of research showing that network centrality in patient-provider networks is a promising indicator of prescription drug-seeking. We measure centrality with recently-developed bipartite methods, and we find that BiRank and CoHITS provide significantly more robust measures of prescription drug-seeking and better predictors of subsequent opioid overdose than traditional centrality estimates, including PageRank in a one mode network projection. Our results suggest that these algorithms may be very useful for identifying centrality across network contexts.

Dissertation, Research in organizational ecology demonstrates that an organization’s competitive position within its market is highly associated with its survival chances, and that patterns of competitive constraint among organizations influence how markets evolve. However, the literature’s conceptualization of market structure is relatively coarse and static; it does not explore how individual organizations’ competitive positions shift or how market offerings change over relatively short intervals of time. In this study, I use social network analysis to study the structure of organizations’ competitive relationships directly. I examine both how changes in the structure of an organization’s competitive environment influence its survival chances, and how the structure of organizations’ competitive relationships affect the stability of market offerings. With a combination of a large crowd-sourced restaurant dataset from Yelp.com and census tract information from the American Community Survey (Census Bureau, 2009; Yelp, 2019), I apply methods from social network analysis, text analysis, and geographic information systems to track how restaurants’ competitive relationships change over time and space, and to study how these changes influence restaurants’ survival chances and overall market stability. This study provides evidence for new mechanisms of competitive constraint among organizations (niche centrality and niche compression) and new mechanisms of market stability (niche redundancy), offers a new theoretic framework for studying market structure and organizational evolution, and has critical implications for theory in the field of organizational ecology.