Factors associated with antibiotic prescribing for acute bronchitis at a university health center

In the United States, 30% of outpatient antibiotic prescribing is estimated to be unnecessary, resulting in almost 47 million unnecessary antibiotic prescriptions each year [1, 2]. Excessive antibiotic prescribing drives the spread of antibiotic resistance, which contributes to increased morbidity, mortality, and economic costs associated with infections [3‐5]. In response, the 2015 U.S. National Action Plan for Combatting Antibiotic-Resistant Bacteria set a goal of reducing inappropriate antibiotic prescribing in outpatient settings by 50% by 2020 [6].

A major source of unnecessary outpatient antibiotic prescriptions is acute bronchitis cases [7‐10]. Acute bronchitis is a common self-limited respiratory illness, characterized predominantly by cough, typically lasting less than 3 weeks [7, 11]. In the US in 2011, cough was the most common illness-related reason for ambulatory care visits, accounting for 2.6 million outpatient visits [12]. A study in the UK estimated that 44/1000 adults are affected by acute bronchitis each year [13]. Antibiotics are not effective for treating acute bronchitis, which is usually of viral etiology [11], and long-standing professional guidelines recommend against antibiotics for uncomplicated cases [14, 15]. Nevertheless, US adults are prescribed antibiotics for acute bronchitis approximately 60–70% of the time [7‐9, 16]. Further, relative to other upper respiratory tract infections for which antibiotic treatment is not indicated (e.g., nasopharyngitis, laryngitis), providers are especially likely to prescribe for acute bronchitis [8, 17‐20]. Due to the prevalence of overprescribing, the U.S. Centers for Infectious Disease Control (CDC) has identified acute bronchitis cases as a major opportunity for reducing unnecessary outpatient antibiotic prescribing [21].

Although acute bronchitis presents an opportunity to improve antibiotic stewardship, there is little consensus regarding effective stewardship interventions for ambulatory care [22‐24]. A diversity of interventions have been proposed, but evidence supporting their effectiveness remains sparse [23, 24]. Implementation of outpatient stewardship programs could be aided by identifying the factors driving overprescribing, which might point to interventions that target those drivers [22]. Factors driving antibiotic overprescribing may differ between hospital and outpatient settings and could include diagnostic uncertainty, real or perceived patient expectations for antibiotics, time pressures, or gaps in provider knowledge [25, 26].

Identifying drivers of prescribing for acute bronchitis could suggest potential interventions, but relatively few studies have focused on identifying these predictors. Prior studies of upper respiratory tract infection prescribing (including for acute bronchitis) in the USA have shown higher rates of antibiotic prescribing in rural (vs. urban) practices [8, 10], when patients have multiple diagnoses [27] or illness of longer duration [28], when providers are advanced practitioners rather than physicians [9], and when providers experience greater diagnostic uncertainty [27]. Since most studies have utilized data reported to insurance companies or national agencies [13, 16, 23], few previous studies have examined how physical exam findings influence prescribing for acute bronchitis. In the few studies that have included data from patient charts, purulent nasal discharge, purulent sputum, abnormal respiratory exam, tonsillar exudate, and sinus tenderness have been reported to be moderately associated with prescribing [20, 29]. In addition, US prescribing rates for uncomplicated acute bronchitis are higher for younger adults (18–39) than older adults (40+) [16], suggesting that factors influencing bronchitis prescribing for young adults are particularly good targets for evaluation and intervention.

University student health clinics provide an opportunity to study antibiotic prescribing in young adult patient populations. In the US, college students comprise a sizeable cohort of the population, with 20.1 million students enrolled in higher education, including 13.8 million students enrolled at 4-year degree-granting institutions [30]. At these 4-year institutions, there are 165.5 annual visits to student health centers for every 100 enrolled students, 37% of which are for respiratory tract infections [31]. Despite evidence that unnecessary antibiotic prescribing is high in young adult populations [16], antibiotic stewardship programs are almost nonexistent at most student health centers, and best stewardship practices are not yet defined. Understanding what drives unnecessary antibiotic prescribing in student health centers is a first step towards evidence-based stewardship policies in these settings, and findings can also inform stewardship efforts with providers treating young adults in similar contexts (e.g., urgent care clinics).

The goal of this study was to identify patient and visit factors associated with antibiotic prescribing for young adults diagnosed with acute bronchitis at a high-volume student health center at a large US university. We conducted a retrospective chart review of all visits with an acute bronchitis diagnosis for the 2015–2016 academic year at the Pennsylvania State University’s Student Health Center. This work is part of a multi-study interdisciplinary effort to improve antibiotic stewardship in emerging adult populations, with an initial focus on students at residential colleges.

This study’s results indicated two key drivers of antibiotic prescribing: variation between individual providers and diagnostic uncertainty. We take each of these in turn. Individual providers had extraordinarily variable rates of antibiotic prescribing for acute bronchitis (ranging from 0 to 80%), despite treating the same patient population at the same clinic. These results suggest that a subset of providers can drive a disproportionate amount of unnecessary antibiotic prescribing for acute bronchitis. In the current study, provider traits (e.g. provider specialty, age) were not included in the logistic regression analysis due to the small number of providers in the data set (21 total). Previous studies have identified provider specialty, provider age, and perceived patient demand for antibiotics as factors influencing provider prescribing rates for upper respiratory tract infections [9, 20, 26, 36, 37].

A second important driver may be diagnostic uncertainty. In the present study, prescriptions were much more likely when findings of rales, rhonchi, or percussion abnormalities were recorded during lung examination, and somewhat more likely when external anterior cervical lymphadenopathy was reported. Rales and percussion abnormalities increased prescribing 13-fold, and rhonchi increased prescribing 5-fold. This increase in prescribing may reflect suspicion of pneumonia. Orders of chest x-rays, which also indicate suspicion of pneumonia, were a significant predictor of prescribing in a bivariate analysis, but were not significant in a multivariate model due to high correlation with other lung exam findings. Providers may prescribe antibiotics when there is suspicion of a condition that would respond to antibiotics or general diagnostic uncertainty [27], and this may not be reflected in the diagnosis code.

Other predictors of prescribing in this study included symptoms of sneezing and vomiting, reported worsening of symptoms, diagnosis of an upper respiratory tract infection, duration of illness, and patient gender. Duration of illness has previously been associated with prescribing for upper respiratory tract infections [28]. Patient gender has not typically been associated with prescribing rates for acute bronchitis [16, 20, 28, 29], although some studies have reported that males are more likely to get antibiotic prescriptions for upper respiratory tract infections [8, 17].

The identification of provider variation and diagnostic uncertainty as drivers of prescribing suggests possible interventions for this clinic and similar settings. Provider variation points to a need for provider-targeted interventions such as audit and feedback, communication training, provider education, or clinical decision support tools [22, 23]. In an ‘audit and feedback’ intervention, individual clinicians receive personalized, ongoing feedback on their prescribing rates [22‐24, 38, 39]. In one study, quarterly feedback resulted in a 50% relative reduction in broad-spectrum antibiotic use for respiratory tract infections [24]. Provider communication training has also been shown to decrease unnecessary antibiotic prescribing [23]. Communication training addresses provider concerns related to patient satisfaction and patient expectation for antibiotics [23]. In some cases, diagnostic uncertainty may be addressed through point of care diagnostic testing [22]. Point of care diagnostics are available for respiratory tract infections including Group A Streptococcus and influenza [22]. There is some evidence supporting point of care testing to reduce antibiotic prescribing for respiratory tract infections [23, 40].

While unnecessary prescribing for acute bronchitis was common in our data, the rate of prescribing was substantially lower than the nationwide average. In the 2015–2016 academic year, antibiotics were prescribed at less than a third of acute bronchitis visits, compared to national rates near 70% [7‐9, 16]. There is still room for improvement, but overall, this suggests that lower rates of prescribing for acute bronchitis are achievable.

Our study is unique in its focus on antibiotic prescribing practices at a university health center. University health services are important centers for antibiotic prescribing serving millions of patients, yet they have largely been overlooked as sites for antibiotic stewardship. To our knowledge, the Pennsylvania State University is the first university with a student antibiotic stewardship program. This study is the first to identify drivers of antibiotic prescribing in a university health center, and one of the few to focus on young adults or consider exam findings and symptoms from patient charts as possible predictors of prescribing. We hope that these findings can be used to inform antibiotic stewardship initiatives at university health centers and similar clinical contexts. Our results suggest that unnecessary antibiotic prescribing is disproportionately driven by a subset of clinicians, and interventions targeting providers may be effective at reducing unnecessary prescribing.

Reducing unnecessary antibiotic prescribing for acute bronchitis cases is a national antibiotic stewardship goal, yet rates of unnecessary antibiotic prescribing remain stubbornly high nationwide. Here we identified factors that influence antibiotic prescribing for acute bronchitis cases at a large university health center. Suspicious findings during lung examinations (rales, rhonchi, percussion abnormality) and individual healthcare providers were the most influential factors affecting antibiotic prescribing rates for acute bronchitis visits. Patient gender, worsening symptoms, duration of illness, symptoms associated with viral infections or allergies, and anterior cervical lymphadenopathy also influenced prescribing rates.