Objective: Methods of reporting of race/ethnicity in biomedical journals are largely unknown. We aimed to systematically examine the reporting practice of race/ethnicity and socioeconomic status (SES) information in biomedical journals.
Study design and setting: All primary research articles that reported more than one racial/ethnic group, published between 1999 and 2003 in Annals of Internal Medicine, JAMA, The Lancet, and The New England Journal of Medicine (n=1,152) were reviewed for their use of race/ethnicity and SES variables. Interobserver reliability was assessed by independent abstraction of 10% of study sample.
Results: There were a total of 116 different terms used to describe various racial/ethnic groups. Assignment of race/ethnicity by self-report was stated in only 13% of papers; 52% of papers identifying race/ethnicity of study participants did not report any SES information. Overall, 16% of articles explicitly stated reasons for collecting information on race/ethnicity.
Conclusion: Our results suggest that race/ethnicity information was suboptimally reported in general medical journals. Terminology used was highly variable. Method of establishing racial/ethnic categories, rationale for collecting race/ethnicity data, and SES information were underreported.