No research effort can result in a comprehensive disclosure of all relevant publications, especially on a publically available dataset which encompasses a wide range of topics. The articles presented here were obtained through an extensive search of publications indices (PubMed, ProQuest, and ScienceDirect). Within each search inquiry, keywords included “BRFSS,” “validity,” and/or “reliability.” Any article which included testing of BRFSS reliability and/or validity was included. Articles which expressed only opinions, without any comparisons or statistical testing were not considered. Given that the purpose of this research was to validate self-reported estimates in an era of declining landline telephone coverage, only those articles which have been published from 2004–2011 were included. Articles were then categorized and are presented in the following topic areas:
1. Access to health care/ general health
2. Immunization, preventive screening, and testing
3. Physical activity measures
4. Chronic disease
5. Mental health measures
6. Overweight and obesity measures
7. Tobacco and alcohol use measures
8. Responsible sexual behavior measures
9. Injury risk and violence
Quality of individual studies may vary significantly. Therefore a scoring rubric was devised to estimate the rigor of the tests of reliability and/or validity found in the literature. Higher rankings on the reliability rubric were achieved by authors who conducted reliability tests using repeated test/retest measures, used multiple samples/populations or multiple time periods. The rubric was also scored higher if authors conducted statistical tests, rather than simply comparing prevalence estimates. Authors who simply tested reliability by noting that results within the BRFSS were internally consistent were ranked lower on the reliability rubric. A similar rubric was used to rank validity assessments. Validity tests comparing the BRFSS to physical measures were ranked highest. Comparing BRFSS validity over time or comparing BRFSS against other self-reported data were ranked lower. Higher ranked assessments of validity and reliability were also characterized by more rigorous statistical comparisons, including the use of sensitivity and specificity measures [15 (link)], kappa and other statistics [16 (link)] or other statistical comparisons [17 (link)]. The rubric provided overall categorical rankings and is not intended to be interpreted as an interval measure of quality estimates. For each of the topics the following information is presented:
1. The number of articles relating to reliability of the BRFSS
2. The number of articles relating to validity of the BRFSS
3. The quality of reliability tests used by authors
4. The quality of validity tests
5. An overall assessment of the literature on reliability and validity of the BRFSS
Thus the method used to assess the literature followed the path illustrated in Figure1 .
1. Access to health care/ general health
2. Immunization, preventive screening, and testing
3. Physical activity measures
4. Chronic disease
5. Mental health measures
6. Overweight and obesity measures
7. Tobacco and alcohol use measures
8. Responsible sexual behavior measures
9. Injury risk and violence
Quality of individual studies may vary significantly. Therefore a scoring rubric was devised to estimate the rigor of the tests of reliability and/or validity found in the literature. Higher rankings on the reliability rubric were achieved by authors who conducted reliability tests using repeated test/retest measures, used multiple samples/populations or multiple time periods. The rubric was also scored higher if authors conducted statistical tests, rather than simply comparing prevalence estimates. Authors who simply tested reliability by noting that results within the BRFSS were internally consistent were ranked lower on the reliability rubric. A similar rubric was used to rank validity assessments. Validity tests comparing the BRFSS to physical measures were ranked highest. Comparing BRFSS validity over time or comparing BRFSS against other self-reported data were ranked lower. Higher ranked assessments of validity and reliability were also characterized by more rigorous statistical comparisons, including the use of sensitivity and specificity measures [15 (link)], kappa and other statistics [16 (link)] or other statistical comparisons [17 (link)]. The rubric provided overall categorical rankings and is not intended to be interpreted as an interval measure of quality estimates. For each of the topics the following information is presented:
1. The number of articles relating to reliability of the BRFSS
2. The number of articles relating to validity of the BRFSS
3. The quality of reliability tests used by authors
4. The quality of validity tests
5. An overall assessment of the literature on reliability and validity of the BRFSS
Thus the method used to assess the literature followed the path illustrated in Figure
Free full text: Click here
