Comparison of dried blood spot to venous methods for hemoglobin A1c, glucose, total cholesterol, high-density lipoprotein cholesterol, and C-reactive protein.

BACKGROUND: Compared to venipuncture, dried blood spots (DBS) can be collected by non-phlebotomists in non-clinical settings, is relatively inexpensive, more easily transported and stored conveniently. Disadvantages of DBS include difficult assay development and validation. This study compared DBS to venous methods for hemoglobin A1c, glucose, total cholesterol, high-density lipoprotein cholesterol, and C-reactive protein (CRP). METHODS: DBS collection and venipuncture were performed on 401 participants. The DBS were collected on Whatman 903 protein saver card. For analysis, 3.2mm blood punches were placed into a 96-well microtiter plate for elution and then analyzed. RESULTS: The Pearson squared correlation coefficients were high for hemoglobin A1c (0.92), CRP (0.84) and glucose (0.81) and low for total cholesterol (0.34) and high-density lipoprotein (HDL) cholesterol (0.30). Sensitivity (>82%) and specificity (>90%) were high for CRP, glucose and hemoglobin A1c at selected clinical cut-points. Low sensitivity (<41%) and high specificity (>87%) were seen for total and HDL cholesterol. CONCLUSIONS: The hemoglobin A1c, glucose and CRP correlated well between DBS and venous methods (r(2)>0.80), but there was a poor correlation for total and HDL cholesterol (r(2)<0.34). This resulted in low sensitivity of DBS methods for total and HDL cholesterol.

Developing a metadata data model for the National Health and Nutrition Examination Survey (NHANES).

The National Health and Nutrition Examination Survey (NHANES), developed by the Centers for Disease Control and Prevention (CDC), is a large and comprehensive health survey utilizing leading edge technologies to produce national estimates of health measures and the nutritional status of the U.S. population. Early NHANES metadata models grouped data by categories with little specificity and often not capturing the complexity of the survey. Subsequently, existing models at the Census Bureau, CDC, and the EPA were evaluated in addition to industry standards, such as DDI, Dublin Core, and ISO 1179. For the NHANES metadata model, the DDI standard and CDC Public Health Conceptual Model were chosen as the backbone for constructing the data model. The new model has led to increased data accuracy and several value-added products for producing codebooks, automatically checking questionnaire skip patterns, and producing questionnaire instrumentation.