Changing Physician Prescribing Behavior: The Community-Acquired Pneumonia Intervention Trial

Kim C. Coley, Susan J. Skledar , Michael J. Fine, Donald M. Yealy , Patrick P. Gleason , Michael L. Ryan, Wishwa Kapoor , and Robert A. Branch


Am J Health Syst Pharm. 2000;57(16) 

In This Article


There were 105 controlgroup patients and 120 intervention-group patients. Demographic characteristics were similar between the groups (Table 1). Although mean ASG scores were slightly higher for the intervention group (2.1 versus 1.9), the difference was not considered clinically significant. Intervention- group patients were more likely than control patients to receive their first antimicrobial dose in the ED (81 patients [68%] versus 59 patients [56%]).

Table 2 compares the primary antimicrobials initially prescribed in the ED and the inpatient units during the two study periods. During the control period, ampicillin-sulbactam was the most commonly prescribed antimicrobial. Fortytwo (71%) of 59 control patients who received antimicrobials in the ED and 61 (58%) of 105 control patients on the inpatient units were given ampicillin-sulbactam. Ampicillin-sulbactam use was significantly lower both in the ED and on the inpatient units during the intervention period (p < 0.001). Conversely, the use of cefuroxime was significantly higher. There were no significant differences in the use of other antimicrobials.

Although not all patients received antimicrobials in the ED, those who did frequently continued on the same regimen once they were transferred to an inpatient unit. During the control period, 59 patients received antimicrobials in the ED, and 43 (73%) of them continued the same regimen upon transfer. Similarly, 62 (77%) of 81 patients who received antimicrobials in the ED during the intervention period continued on the same regimen when transferred to an inpatient unit. During the intervention period, 14 (17%) of 81 patients received antimicrobials in the ED that deviated from the treatment guidelines. The therapy of only 5 of these 14 patients was changed to meet the guidelines upon transfer to an inpatient unit. This underscores the importance of proper antimicrobial selection in the ED.

The addition of erythromycin to provide coverage against atypical pathogens was included in our treatment recommendations for nonsevere CAP. We therefore evaluated the number and type of concurrent antimicrobials prescribed to determine whether additional agents selected were consistent with the guidelines or whether physicians were adding antimicrobials to cover a wider spectrum of pathogens. We found that coverage against atypical pathogens with erythromycin or doxycycline accounted for the vast majority of cases in which patients received more than one antimicrobial during both study periods. Such coverage accounted for 48 (86%) of 56 patients and 57 (92%) of 62 patients receiving more than one antimicrobial during the control and intervention periods, respectively. Of the 94 intervention-group patients receiving cefuroxime, only 4 who were prescribed a concomitant antimicrobial received an agent other than erythromycin. These findings confirm that physicians were not inappropriately prescribing additional antimicrobials when prescribing cefuroxime.

Changing antimicrobial use in patients with nonsevere CAP did not have a negative impact on patient outcomes. The mean length of stay decreased from 6.7 days in the control group to 5.8 days in the intervention group (p = 0.2, unpaired t test, two tailed). There were no significant differences between the groups with respect to their discharge status or mortality rates. Of 105 control patients, 92 (88%) were discharged to home, compared with 99 (83%) of 120 intervention-group patients. Four control patients and five intervention-group patients died. There were no differences in readmission rates between the two groups (one patient in each group was readmitted within one month of the previous discharge).

Antimicrobial costs decreased significantly under the intervention program. The cost of antimicrobial therapy decreased from $180 per patient during the control period to $120 per patient during the intervention period (p = 0.03, unpaired t test). The majority of the costs during the control period were driven by agents that did not meet the treatment guidelines. This pattern was reversed in the intervention period.

The intervention program was discontinued at the end of June 1997. To determine whether changes in prescribing behavior were sustainable, we conducted a follow-up evaluation for eight months after the intervention period ended. Compliance with the treatment recommendations continued to be high, with 85 (79%) of 107 patients who met the treatment criteria receiving cefuroxime as the initial antimicrobial.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.