Inappropriate antibiotic use is a significant problem in healthcare. The CDC estimated in 2013 that 2,049,442 illnesses and over 23,000 deaths were caused by antibiotic resistance. Clinicians in the emergency department often have to start antibiotic treatment immediately without knowing the organism. Therapy is initiated using broad spectrum antibiotics with the aim to transition to targeted, narrow spectrum therapy as soon as possible. These best guesses lead to overuse of antibiotics, delay the best treatment, and increase the likelihood of breeding further antibiotic resistance. We are developing assistance for selecting the best empiric antimicrobial treatment when the causal microorganism is still unknown. Existing knowledge about antibiotic resistance in local communities and hospitals are the basis for this new antibiogram. The new antibiogram will integrate information about past antibiotic resistance in the patient and in her community with information about side effects of possible antibiotics in a transparent manner to facilitate effective human-computer teamwork. We are collaborating with the emergency and infectious disease departments at UCSD to integrate these technologies into daily workflow.