Short-term, within-person variability in clinical chemistry test results. Experience from the Atherosclerosis Risk in Communities Study.

We report the within-person, between-person, and methodological variances of a number of chemical analytes for free-living, middle-aged adults who participated in the Atherosclerosis Risk in Communities Study's Intraindividual Variability Study. The Atherosclerosis Risk in Communities Study is a National Institutes of Health-sponsored multicenter study of atherosclerotic risk factors. In the Atherosclerosis Risk in Communities Study and its Intraindividual Variability Study, concentrations of the following 12 chemical analytes were measured in serum from fasting individuals: sodium, calcium, potassium, creatinine, albumin, total protein, magnesium, phosphorus, urea, insulin, glucose, and urate. The analytes are listed in order of the increasing reliability coefficient (ie, the fraction that between-person variance represents of the total observed population variance), which ranged from .59 for sodium to .91 for urate. The reliability coefficient is a strong predictor of the possibility of finding statistical correlations between measured analyte concentrations and disease occurrence in an epidemiological study like the Atherosclerosis Risk in Communities Study. The within-person variance and methodological variance are also useful in computing confidence intervals for sequential laboratory test results in patients and evaluating limits for internal quality control and proficiency testing programs.