Continuous Rural-Urban Coding for Cancer Disparity Studies: Is It Appropriate for Statistical Analysis?

Area of Science:

• Epidemiology
• Geographic Information Systems (GIS)
• Public Health

Background:

• Categorizing rural-urban codes as nominal variables is common in cancer disparity research.
• This categorization loses information on continuous rural-urban transitions, potentially leading to inconsistent findings.
• Few studies have validated using rural-urban codes as continuous variables.

Purpose of the Study:

• To assess the validity of using rural-urban codes as continuous variables in cancer disparity analyses.
• To compare results from continuous rural-urban codes with those from truly continuous rurality data.
• To determine the optimal level for using continuous rural-urban codes.

Main Methods:

• Geocoded cancer cases (breast, colorectal, hematological, lung, prostate) in north central Florida (2005-2010).
• Employed a linear hierarchical model to regress late-stage cancer occurrence on continuous rural-urban codes.
• Validated findings by comparing with analyses using continuous rurality data.

Main Results:

• Regression analysis confirmed that using rural-urban codes as continuous variables yields consistent results with truly continuous rurality data for all cancer types.
• Rural-urban codes at the census tract level provided the closest estimation and are recommended.
• This approach maintains the ordering and continuum of the rural-urban landscape.

Conclusions:

• It is methodologically sound to use rural-urban codes as continuous variables in cancer studies.
• The continuous-variable approach enhances analytical flexibility and preserves crucial ordering information.
• This method offers a more nuanced understanding of cancer risk variation across the rural-urban spectrum.