The present study tests the hypothesis that a quality control chart methodology can be developed that will allow the early detection of unexpected patterns of occurrence of laboratory critical values, representing variations in hospital processes. As defined by Lundberg, a critical value indicates a pathophysiological state at such variance with normal as to be life-threatening, unless something is done promptly and for which some corrective action could be taken. For this study, aggregate data were drawn from the Information Warehouse at the Ohio State University Medical Center (OSUMC), beginning in April 2000 to January 2005 (58 months). All critical laboratory value occurrences were captured for that time period, and sorted by medical service. The critical values were plotted in time-series fashion and analyzed for aberrances in monthly critical value rates, analogous to clinical control charts. Thirty-four aberrances, or “spikes”, were detected in the 234 time-series charts (14.53%), representing special cause variations. This aggregate data was traced back to the patient level, where extensive chart review yielded demographic and risk factor information. Analysis of the study data showed both a significant increase in critical value occurrence rates compared to previous studies, and a heterogeneous patient population, with different risks depending on the type of medical care service. A hospital-based case-control study tested the hypothesis that monitoring and identification of special causes in monthly critical value occurrence rates would yield a higher percentage of patient adverse events, then through random selection. A patient adverse event was defined as cardiac injury, cerebrovascular adverse events, major bleeding episode, prolongation of hospital stay, or death. Multiple logistic regression analyses revealed several significant predictor variables in this case-control study: the occurrence of patient adverse events, advanced age, gender, current tobacco smoking, prior myocardial infarction, number of critical values per patient, and assigned medical care unit risk. The most relevant independent predictor variable of a spike month relative to a non-spike month was the occurrence of adverse events (adjusted odds ratio=2.54, 95% confidence interval = 1.64-4.70, p=0.004). Thus, the risk of a spike month increased more than 2.5 times with occurrence of an adverse event.