Quality of Pharmacotherapy Consultations
We evaluated the performance of 116 U.S. drug information centers in responding to specific questions about drugs. The primary measures were correctness of responses and extent of probing for patient data. Questions addressed the effect of ranitidine on blood alcohol concentrations, the potential interaction between didanosine and dapsone, prevention of nonsteroidal antiinflammatory drug (NSAID)-induced peptic ulcers, and use of erythromycin for diabetic gastroparesis. The percentages of centers providing correct overall responses were 70% for the ranitidine question, 90% for the didanosine-dapsone question, 8% for the NSAID question, and 20% for the erythromycin question. For the three patient-specific questions, the percentages of centers eliciting vital patient data were 27% for the didanosine-dapsone question, 86% for the NSAID question, and 5% for the erythromycin question. In providing pharmacotherapy consultations, drug information centers generally failed to obtain pertinent patient data, thereby risking incorrect responses and inappropriate recommendations.

There are over 100 drug information centers in the United States whose primary function is to provide accurate, unbiased, and timely consultations about drugs and drug therapy. Consultations often pertain to drug selection, dosing, and administration, as well as prevention and management of adverse reactions and interactions. The centers mainly serve physicians, other health care providers, and the general public. They are commonly located in hospitals or major medical centers, and are staffed by pharmacists and their assistants. Although the centers potentially have an important role in patient care, the quality of consultations may vary from center to center.

To our knowledge, two nationwide studies examined the quality of pharmacotherapy consultations provided by U.S. drug information centers. The first one evaluated 90 centers that were identified from two directories published in 1976. The second assessed 56 centers selected randomly from a registry of 154 centers that responded to a demographic survey in 1990. In both studies, centers were contacted by telephone, and a single, multipart drug information question was posed using a conversation protocol. To obtain responses under "usual" working conditions, the centers were not informed that they were participating in a study. The findings from both studies suggested that responses provided by drug information centers often are deficient.

We conducted the third nationwide study to evaluate the performance of drug information centers. Although it was similar in design to the first two, there were some methodologic differences; principally, we used four multipart drug information questions rather than one. These four questions covered a wide range of clinical topics, thereby allowing fuller assessment of center performance.