Chart reminders in the diagnosis of anemia.

We hypothesized that chart reminders would decrease the number of cases in which a low admission hemoglobin level was missed or ignored. Additionally, we compared the cost of generating the chart reminders by hand with the estimated cost of modifying the university hospital information system for this function. For three months, two medical students reviewed all abnormal hemoglobin results and assigned those more than 1 g below normal limits to a control or study group. A questionnaire, placed in all study group charts, requested an explanation for the low hemoglobin level. Eleven of 141 study cases and 16 of 150 control cases failed the audit criteria (differences not significant). The cost was $308 compared with $3,000 for altering the hospital information system. We conclude that chart reminders did not improve the detection of anemia and that it cost less to test this intervention by manually simulating the computer function.