Balancing acts: Deciding for or against antibiotics in acute respiratory infections

Patient, clinician, and system factors influence unnecessary antibiotic prescribing for viral acute respiratory infections (ARIs). When patients seek care for ARIs, they use a variety of tactics to obtain antibiotic prescriptions.¹ Both clinicians and patients tend to overemphasize the importance of purulent secretions, whether from the nose or lung, in deciding whether an antibiotic is needed.^2-4

Antibiotic prescribing increases with clinician age⁵ and number of years in practice,⁶ and antibiotics are more likely to be prescribed for ARIs in urban⁵ and nonteaching^7,8 practice settings.

In addition, clinicians’ perceptions that patients or parents expect antibiotics are often inaccurate. They frequently assume that patients are dissatisfied when they do not receive antibiotic prescriptions for ARIs; however, accurate explanations and reassurance, not antibiotic prescriptions, have been shown to increase patient satisfaction.^9,10

These factors, though known to be associated with excessive antibiotic prescribing, do not explain how clinicians actually make decisions to give antibiotics for ARIs. Thus, we applied grounded theory—a qualitative research method used to examine social processes—to provide insight into the antibiotic prescribing process.¹¹

Methods

Setting

Our study was conducted in a rural Western community with a population of 32,014, served by 24 primary care clinicians.

Sample

We aimed to interview all of the community’s 24 primary care clinicians. Each clinician was mailed information about the study and then received a follow-up phone call to arrange an interview. Three clinicians were not interviewed due to scheduling conflicts.

The median age of the participants was 43 years with a range of 32 to 58. Nine subjects (43%) were women. Four subjects were nurse practitioners (NPs) and 17 were MDs. Areas of practice included internal medicine (n=5), college health (n=5), family medicine (n=4), pediatrics (n=4), and emergency medicine (n=3). No clinicians were representative of minority ethnic groups.

Interviews

We gathered data in audiotaped, semi-structured interviews. Hart, a practicing NP in the study community who received doctoral training in qualitative research, interviewed the 21 clinicians. The interviews ranged from 30 to 120 minutes with an average of 1 hour.

The clinician was asked to describe how he or she decided when patients presenting with ARIs should receive antibiotics. They were also asked to describe situations in which they would “most definitely” or “definitely not” prescribe antibiotics and situations that caused uncertainty. They were also asked to describe challenges associated with ARI management.

Data analysis

Hart analyzed the interviews using the constant comparative method, a form of qualitative analysis whereby each unit of data is compared with previously identified units.¹¹ Each line of transcribed interview data was analyzed chronologically and coded for patterns, themes, and processes. The resulting codes were then compared for similarities and differences. Categories and “families” of categories were developed. This process continued until the basic social process, or one phenomenon central to each interview, was discovered. The findings were then discussed and refined by members of the research team, consisting of Pepper and Gonzales, as well as 3 PhD-prepared researchers with expertise in qualitative research.

Credibility

A recognized method for establishing credibility in qualitative research is through the use of “member checks,” wherein findings are presented to the original study participants for confirmation and clarification.¹² A group member check was obtained during a “Grand Rounds” presentation at the community hospital. Thirty-one people attended, including 9 of the original participants. During a forum following the presentation, study participants confirmed that the findings accurately described their antibiotic prescribing processes.

Results

Two main concepts emerged from the interviews: individual best practice and perceived patient/parent satisfaction. These concepts are discussed using illustrative quotes from actual interviews.

Individual best practice describes how each clinician was attempting to do what he or she believed to be clinically “best” for the patient presenting with ARI symptoms. Clinicians differed in their approaches to patients presenting with ARIs. Some prescribed antibiotics for non-specific upper respiratory infection symptoms. Others were reluctant to prescribe antibiotics unless strict diagnostic criteria were met. Ultimately, each clinician made a decision based on what he or she believed was best for the patient. What defined “best” was clinician-specific. Though “best practice” always derived from evidence that informed the individual’s practice, it should not be confused with the term “best research evidence,” as defined by Straus et al.¹³ Five categories influenced the concept of individual best practice (TABLE).