The incidence and severity of phlebitis associated with the i.v. push method versus the i.v. piggyback method of administration of antimicrobial agents were evaluated. All patients admitted to two nursing units of an 886-bed, tertiary-care hospital from March through June 1987 who received intravenous cephalosporins or extended-spectrum penicillins were considered for inclusion in the study. The method of administration, either i.v. push or i.v. piggyback, was randomly assigned. Each patient's catheter site was evaluated every 24 hours according to standardized definitions of phlebitis and patency. Information on each patient and venipuncture site also was collected. A total of 319 catheter sites were studied in 155 patients. There was no significant difference in the incidence or severity of postinfusion phlebitis between the i.v. push and the i.v. piggyback groups (41% versus 47%). When studying only those catheter sites discontinued because of phlebitis, the time until development of phlebitis averaged 45 +/- 20.5 hours in the i.v. push group, compared with only 36.2 +/- 17.6 hours in the i.v. piggyback group; this difference was significant. The fact that the catheter sites lasted significantly longer in the i.v. push group, combined with elimination of the cost of syringe infusion pumps or i.v. tubing and minibags, suggests that use of the i.v. push method may result in substantial cost savings. Other advantages associated with the i.v. push method include time savings for nursing personnel and patients and greater ease in scheduling doses of multiple antimicrobial agents.(ABSTRACT TRUNCATED AT 250 WORDS)