Using Secondary Data Sources in Prevention Research 

By Dr. Iris Smith 

The Strategic Prevention Framework (SPF) is a data-driven process for planning, executing, and evaluating prevention strategies, of which data collection, interpretation, and dissemination are an integral part.  Coordinated data collection, analysis, and synthesis at the individual, community, state, and national level are key elements of an effective prevention system. When adding in secondary sources to the process, care is needed to properly assess the validity of the data.  

Epidemiology, the foundation science of public health, helps us identify local and national patterns in behavioral outcomes and associated risk and protective factors by answering questions related to scope, frequency, and severity of prevention targets and their impact on local communities. They also hypothesize and test causal relationships between exposures and outcomes.1  In many cases, this requires us to think beyond the “person” level to consider the social environment and its impact on health outcomes. Techniques such as spatial analysis can be useful in broadening the lens of prevention to factors in the social environment that influence outcomes, helping us determine where, when, and among whom to intervene.2  

Advances in technology have led to increased use of large, publicly available, data sets (“big data”) and new methodologies for abstracting, analyzing, and applying it. Data collected by other agencies and organizations (secondary data sources) can be a cost-effective way to supplement primary data collection.  However, large data warehouses are also complex and most often consist primarily of observational and not experimental data.  This can lead to assumptions about causality that may not be valid. 3  

Wesson et al.(2022) cautions that use of secondary data, particularly “big data” has the potential to perpetuate inequities in public health, especially when vulnerable populations are not adequately represented.4 For example, data from social media platforms such as Facebook, Twitter, and Google may not capture information about populations lacking the resources to participate fully in the digital world or who have limited access to technology (for example rural populations).  Such data, while valuable, may not be representative of the general population or reflect the needs and behaviors of vulnerable populations.   

Improving the health equity of data requires that originators and users consider six ‘Vs”:  Volume: the amount of data available; Value: the usefulness of data for decision-making, Variety: the types of data included; Veracity: the trustworthiness of the data; Virtuosity: equity and ethics in design and analysis, and Velocity: the speed with which data are collected and processed, i.e., its timeliness.5 

Resources 

Eberth JM, Kramer MR, Delmelle EM, Kirby RS (2021)  What is the Place for Space in Epidemiology?  Annals of Epidemiology, 64; pg. 41-46. https://www.sciencedirect.com/journal/annals-of-epidemiology   

Li L, Novillo-Ortiz D, Azzopardi-Muscat N, Kostkova P (2021).  Digital Data Sources and Their Impact on People’s Health:  A Sytematic Review of Systematic Reviews.  Frontiers in Public Health, 05 May 2021 | https://doi.org/10.3389/fpubh.2021.645260 

Wesson P, Haswen Y, Valdez G, Stojanovski K, and Handley MA (2022).  Risks and Opportunities to Ensure Equity in the Application of Big Data Research in Public Health.  Annual Review of Public Health, 43.8; 8.1-8.20.  https://www.annualreviews.org/journal/publhealth 

_________________________________ 

[1] Eberth JM, Kramer MR, Delmelle EM, Kirby RS (2021)  What is the Place for Space in Epidemiology?  Annals of Epidemiology, 64; pg. 41-46. 

[2] IBID 

[3]Wesson P, Haswen Y, Valdez G, Stojanovski K, and Handley MA (2022).  Risks and Opportunities to Ensure Equity in the Application of Big Data Research in Public Health.  Annual Review of Public Health, 43.8; 8.1-8.20.  https://www.annualreviews.org/journal/publhealth 

[4] IBID 

[5] IBID