J Am Board Fam Pract: first published as 10.3122/15572625-13-6-462 on 1 November 2000. Downloaded from

Antibiotic Prescribing for : How Low Can We Go?

By the time I graduated from medical school in of that wise philosopher Pogo, "We have seen the 1975, I had learned that most enemy, and he is us." in otherwise healthy children and adults No doubt, we can do better. Several investiga­ were caused by , such as parainfiuenza, in­ tors have described successful methods of reducing fluenza, adenovirus, and respiratory syncytial . use for respiratory tract infections. This learning was reinforced during my family Gonzales et al2 found that combined patient and practice residency training in Charleston, Se. I was clinician education was effective in reducing anti­ taught that the challenge of primary care practice is biotic use for from 74% to 48% in to distinguish the many patients with benign, self­ a health maintenance organization setting. Using a limited respiratory tract infections from those pa­ quality improvement approach and a computer­ tients who are more seriously ill with bacterial based patient record in an academic family practice and who need an antibiotic to recover setting, Ornstein and colleagues3 reduced antibi­ more quickly and more certainly. Anned with this otic prescribing for acute bronchitis from 60% to scientific knowledge, I marched valiantly into prac­ less than 30%. tice, determined to base my prescribing on good These and other initiatives show that it is pos­ science. It was only in rural practice that I clearly sible to reduce use of for acute respira­ recall regularly confronting syndromes called tory tract infections, but how low can we go? The 4 "acute bronchitis" and "," for which pa­ report by Hueston et al in this issue of the JABFP suggests that, for patients with the diagnosis of tients seemed to expect an antibiotic and for which acute bronchitis, the answer might not be 0%. their previous had prescribed one. "I get After the successful campaign to reduce antibiotic this bronchitis every fall. Dr. Smith gives me a shot use for acute bronchitis by Ornstein et aV a stub­ of and I'm fine 2 days later." With mod­ born residual of 26% of patients received an anti­ est success I was able to talk patients out of this biotic. Why? To answer this question, Hueston approach, but I more frequently failed than suc­ and colleagues compared clinical characteristics of ceeded, though many were willing to take my pre­ patients who received an antibiotic with those who scription for pills as an acceptable alternative to the did not in a retrospective chart review conducted shot in the buttock. http://www.jabfm.org/ after the largely successful intervention. They com­ A quarter-century later I am still fighting the pared clinician characteristics as well. The investi­ good battle of appropriate antibiotic use for respi­ gators nearly came up empty handed. Patients' ratory tract infections. After a 25-year cold war, were poor predictors of anti­ however, the battle is heating up! The Centers for biotic prescribing in these recalcitrant cases. There Control and Prevention has entered the were no clinician factors that predicted antibiotic fray, urging us and our patients to use antibiotics prescribing-no "resistant" physicians were found. on 27 September 2021 by guest. Protected copyright. sparingly and wisely for acute respiratory tract in­ The study was, however, underpowered to detect fections. 1 Family physicians have been named as a some potentially important clinical predictors. For major contributor to the problem of bacterial re­ example, an antibiotic was prescribed for 36% of sistance, and the reasons we participate in the prob­ patients with , 42% of patients with shortness lem have been thoroughly described. In the words of breath, 56% of patients with chest , and 75% of patients with sweats, though a small proportion of patients had these complaints. Presence of pu­ Submitted, 14 August 2000. rulent , a common reason for antibiotic pre­ From the Department of Family Medicine, Michigan State University College of Human Medicine, East Lansing. scribing, was not assessed. Address reprint requests to John M. Hickner, MD, MS, Children were much less likely to receive an Department of Family Practice, Michigan State University College of Human Medicine, B III Clinical Center, Mich­ antibiotic than adults, 3% compared with 34%. igan State University, East Lansing, MI 48824. Patient expectation could not be measured in this

462 JABFP November-December 2000 Vol. 13 No. 6 J Am Board Fam Pract: first published as 10.3122/15572625-13-6-462 on 1 November 2000. Downloaded from study, but it is a likely explanation for this striking was prescribed: Other than the diagnosis, what difference. Perhaps parents are more willing to take other factors did you consider when prescribing an a watch-and-wait approach for their children be­ antibiotic for this patient? Clinicians reported cause the current national antiantibiotic campaign other factors in 95% of cases. The most common has begun to have some effect, whereas adults who reasons were patient not improving (21 %), patient might have received an antibiotic for a similar in­ getting worse (19%), and patient has been sick for fection in the past are more likely to demand an too long (19%). Clinicians prescribed an antibiotic antibiotic. Recently, Dosh and colleaguess found in 98% of cases in which they believed the likeli­ that patient demand for an antibiotic for an upper hood of an adverse outcome was high if no antibi­ respiratory tract , acute bronchitis, and otic was prescribed. acute sinusitis was strongly predicted by having How low can we go with antibiotic prescribing received an antibiotic for a previous similar illness. for acute respiratory infections? Here are my tar­ I doubt that patient expectation is the only factor gets for the year 2005: upper respiratory tract in­ driving continued antibiotic prescribing for acute fections 10%; acute bronchitis 25%; acute sinusitis bronchitis as well as other common acute respira­ 50%. Until the prognostic uncertainty of pro­ tory tract infections, or ARIs. (I prefer this termi­ longed symptoms is resolved (or not!) by a large nology rather than the anatomic nomenclature of randomized clinical trial, many physicians will con­ bronchitis, sinusitis, and upper respiratory tract in­ tinue to reach for the prescription pad for patients fection, because the same organism might cause who have acute respiratory tract infections who acute bronchitis in one person, simple upper respi­ have been sick for too long or who are not getting ratory tract infection in another, and acute sinusitis better. in yet another.) Based on my many discussions with John M. Hickner, MD, MS family physicians and from our own research, phy­ East Lansing, Mich sicians do believe that some patients with ARls benefit from antibiotic treatment. In a prospective References descriptive study of diagnosis and treatment of 1. Schwartz B, Bell DM, Hughes JM. Preventing the ARIs in a rural primary care research network, we emergence of : a call for ac­ tion by clinicians, public health officials, and pa­ found a typical pattern of antibiotic prescribing: tients. JAMA 1997;278:944-5. 21 % for upper respiratory tract infections, 80% for 2. Gonzales R, Steiner JF, Lum A, Barrett PH Jr. De­ acute bronchitis, and 98% for acute sinusitis.s In creasing antibiotic use in ambulatory practice: im­ this "untrained" group of physicians and patients, pact of a multidimensional intervention on the treat­

antibiotic prescribing was not independently pre­ ment of uncomplicated acute bronchitis in adults .. http://www.jabfm.org/ dicted by any symptoms but was predicted by a JAMA 1999;281:1512-9. small number of physical findings: discolored nasal 3. Ornstein SM, Jenkins RG, Lee FW, et. al. The computer-based patient record as a CQI tool in a discharge, rales or rhonchi, tenderness, and family medicine center. Jt Comm J Qual Improv postnasal drainage. Surprisingly, patient expecta­ 1997;23:347-61. tion was not an independent predictor of antibiotic 4. Hueston W}, Hopper JE, Dacus EN, Mainous AG, prescribing. Roetzhein RG. Why are antibiotics prescribed for Hueston and colleagues speculate that factors patients with acute bronchitis? A postintervention on 27 September 2021 by guest. Protected copyright. other than signs, symptoms, and clinician charac­ analysis. J Am Board Fam Pract 2000;13:398-402. teristics predict antibiotic prescribing for acute 5. Dosh S, Hickner JM, Mainous AG 3rd, Ebell M. Predictors of antibiotic prescribing for nonspecific bronchitis. I agree. In our study of ARIs, we ex­ upper respiratory infections, acute bronchitis, and plored this issue by asking clinicians the following acute sinusitis: an UPRNet study. Upper Peninsula question for each patient for whom an antibiotic Research Network. J Fam Pract 2000;49:407-14.

Editorials 463