The testing of gastric pH in the ICU has become the standard of care for most critically ill patients. It has been demonstrated that maintaining a gastric pH of greater than 3.5 confers protection from upper GI bleeding, while lesser pH values subject the patient to hemorrhagic risk. We compared nasogastric pH, as measured by pH electrode, to color-scaled pH paper in 16 critically ill patients hospitalized 3 to 10 days in the surgical ICU. Statistical analysis of 370 gastric specimens revealed a sensitivity of 66.7% and specificity of 94.5% when a paper pH (pH[p]) of greater than or equal to 4 was used as the therapeutic end-point. The sensitivity and specificity of the same pH(p) for clear buffered solutions were 100%. We conclude that the use of pH(p) lacks the clinical accuracy for determining the effects of therapy for the prophylaxis of stress gastritis and will lead to a significant degree of undertreatment. This lack of accuracy is not due to observer error or the quality or age of the testing paper. Our results suggest that if the measurement of gastric pH by pH(p) analysis is to be used as a guide for the prevention of stress-related hemorrhage, a more accurate method of monitoring may be warranted.