Prescribing Antibiotics for Upper Respiratory Infections
Drug-resistant bacteria have become a major public health problem in the United States. Physicians have contributed to this resistance by prescribing antibiotics for conditions for which those drugs are not effective or indicated.1,2 For example, although viruses cause nearly all upper respiratory tract infections (URIs), physicians frequently prescribe antibiotics for these conditions. When presented with the case of an infant with scant green mucopurulent nasal secretions of a day’s duration, 71% of family physicians and 53% of pediatricians would immediately prescribe an antibiotic.3 The studies by Hueston and colleagues4 and Dosh and coworkers5 in this issue of the Journal provide additional insight into the diagnosis, coding, and antibiotic prescribing practices of primary care clinicians with patients presenting with URIs.
Acute sinusitis, acute bronchitis, URIs, and colds
Hueston and colleagues propose that the common cold, acute sinusitis, and acute bronchitis should be considered as a single clinical entity: a URI. Their chart audit of a university faculty-resident practice included patients with the diagnosis of acute bronchitis and URI as coded in an electronic medical record. Although cough and wheezing were associated with the diagnosis of bronchitis, all signs and symptoms explained only approximately a third of the differences between the 2 conditions.
In the study by Dosh and coworkers of patients presenting with an acute respiratory infection in 15 rural family practices in Northern Michigan, the diagnoses were: common cold in 34% of patients, sinusitis in 37%, and bronchitis in 29%. Although patients and physicians attach considerable importance to the findings of productive cough, purulent nasal drainage, chest sounds, and sinus tenderness, these signs and symptoms are nonspecific. As mentioned by Hueston and colleagues, categorizing signs and symptoms as either sinus- or bronchial-dominant may be more useful in deciding on comfort care measures than attaching a label of sinusitis and bronchitis to an antibiotic prescription.
How do clinicians decide to prescribe antibiotics for URIs?
Dosh and coworkers studied antibiotic prescribing and patient expectation models. The 5 variables independently associated with antibiotic prescribing were positive associations with sinus tenderness, purulent nasal discharge, rales/ronchi, postnasal discharge, and a negative association with clear nasal discharge. They identified secondary factors in 95% of patients receiving an antibiotic. Those factors included assessments that the patient was not improving, the patient was getting worse, and the patient was sick for too long. Secondary factors are associated with clinical experience and reflect the uncertainty of distinguishing a “bad cold” from a bacterial illness. In the study by Hueston and colleagues, patients with a diagnosis of bronchitis received an antibiotic 6 times more frequently than patients with a diagnosis of a URI. Clinical experience may account for the differences in frequency of diagnosing bronchitis by faculty and resident physicians. In the study by Dosh and coworkers, there was no difference in the antibiotic prescribing patterns of nonphysician and physician clinicians. The clinicians in that study prescribed an antibiotic 98% of the time when they felt that there was a high likelihood of adverse outcome for a patient not receiving an antibiotic. The accuracy of clinician perception in this regard is associated with clinical experience and warrants further study. A chart review cannot be used to define the elements of clinical experience.
Patients’ expectations and physicians’ perceptions of expectations
Sixty-nine percent of patients report that they expect antibiotics when they experience a discolored nasal discharge.6 Such patient expectations are strongly associated with antibiotic prescriptions. One study showed that patients who want antibiotics were more than 3 times as likely to receive them as those who do not specifically want them.7 Parental expectations that an antibiotic prescription would be given also increase the probability of receiving one.8,9 Dosh and coworkers did not identify patient expectations or clinician belief that patients expected an antibiotic as an independent factor associated with an antibiotic prescription. They reported, however, that patient expectation of an antibiotic is driven by past physician behavior.
Physicians may resort to the “fudge factor,” delayed antibiotic prescription technique when they are uncertain about patient outcome or to avoid patient disappointment. In this scenario, the antibiotic is the magic potion. There is no evidence that the delayed prescription technique is more effective than a natural cure.
How accurate are physicians’ perceptions of patient expectations? A 1996 study of 113 Oklahoma family physicians showed that for 25% of patients seeking care for upper respiratory illnesses, physicians’ perceptions of patients’ desire for antibiotics were inaccurate. An additional 26% of physicians were unsure about patients’ expectations.10
Physicians perceive that patients will be unhappy or seek care elsewhere if they do not receive antibiotics for URIs. Chagrin has been described as a factor in clinical decision making.11,12 The “chagrin factor” wins out even when the patient has a low likelihood of bacterial infection. Physicians may fear that if patients do not get what they want they will see another physician who will tell them they have “walking pneumonia” and prescribe a “strong” antibiotic.
