Opportunities to Improve Antibiotic Prescribing for Adults With Acute Sinusitis, United States, 2016–2020

Abstract

Background

Better understanding differences associated with antibiotic prescribing for acute sinusitis can help inform antibiotic stewardship strategies. We characterized antibiotic prescribing patterns for acute sinusitis among commercially insured adults and explored differences by patient- and prescriber-level factors.

Methods

Outpatient encounters among adults aged 18 to 64 years diagnosed with sinusitis between 2016 and 2020 were identified by national administrative claims data. We classified antibiotic agents—first-line (amoxicillin-clavulanate or amoxicillin) and second-line (doxycycline, levofloxacin, or moxifloxacin)—and ≤7-day durations as guideline concordant based on clinical practice guidelines. Modified Poisson regression was used to examine the association between patient- and prescriber-level factors and guideline-concordant antibiotic prescribing.

Results

Among 4 689 850 sinusitis encounters, 53% resulted in a guideline-concordant agent, 30% in a guideline-discordant agent, and 17% in no antibiotic prescription. About 75% of first-line agents and 63% of second-line agents were prescribed for >7 days, exceeding the length of therapy recommended by clinical guidelines. Adults with sinusitis living in a rural area were less likely to receive a prescription with guideline-concordant antibiotic selection (adjusted risk ratio [aRR], 0.92; 95% CI, .92–.92) and duration (aRR, 0.77; 95% CI, .76–.77). When compared with encounters in an office setting, urgent care encounters were less likely to result in a prescription with a guideline-concordant duration (aRR, 0.76; 95% CI, .75–.76).

Conclusions

Opportunities still exist to optimize antibiotic agent selection and treatment duration for adults with acute sinusitis, especially in rural areas and urgent care settings. Recognizing specific patient- and prescriber-level factors associated with antibiotic prescribing can help inform antibiotic stewardship interventions.

Antibiotic use is a driver of antimicrobial resistance and can lead to adverse event–related emergency department visits and Clostridioides difficile infections [1–5]. Reducing inappropriate antibiotic prescribing for common outpatient conditions can help reduce antibiotic-associated adverse events and improve patient outcomes. Sinusitis is the most common indication for antibiotics in adults in the outpatient setting, with almost 3.7 million antibiotic courses dispensed per year in the United States [6–8]. The clinical practice guideline of the Infectious Diseases Society of America (IDSA) recommends 5 to 7 days of therapy for adult patients with an uncomplicated infection and a favorable initial response (ie, improvement or no worsening of symptoms after 3–5 days) [9]. However, in 2016, about 70% of antibiotics prescribed for adults with sinusitis were for ≥10 days, and about 36% of prescriptions were for guideline-discordant agents [8]. Additionally, due to the low prevalence of acute bacterial sinusitis, antibiotic therapy is often unnecessary, with one study estimating that per 1000 population, 27 antibiotic prescriptions for sinusitis were appropriate [6, 9–11]. In contrast to patients with viral sinusitis, patients with bacterial sinusitis have persistent symptoms without improvement for ≥10 days or severe signs or symptoms (ie, high fever and purulent nasal discharge or facial pain) for at least 3 to 4 consecutive days at the beginning of illness, or they experience worsening symptoms after initial improvement [9].

Previous studies have found that outpatient antibiotic prescribing varies by patient- and prescriber-level factors, such as older adults and patients residing in the South census region having a higher likelihood of receiving an antibiotic prescription or advanced practice clinicians being more likely to prescribe broad-spectrum antibiotics than physicians [12–14]. Studies examining acute respiratory tract infections also identified differences in inappropriate antibiotic prescribing by health care setting and prescriber specialty, such as older patients being more likely to receive antibiotics and physicians practicing in rural areas and urgent care centers more likely to prescribe them [15–17]. However, associations between patient- and prescriber-level factors with inappropriate antibiotic prescribing among adults with sinusitis have not been well described by more recent national-level data [18]. Our objective was to characterize antibiotic prescribing patterns and differences by patient- and prescriber-level factors in a national cohort of commercially insured outpatient adults with acute sinusitis to inform antibiotic stewardship strategies.

METHODS

Data Source

We used administrative claims data from the Merative MarketScan Commercial Database (2016–2020), which contains person-level, deidentified, longitudinal data on demographic characteristics, inpatient and outpatient procedures and services, and outpatient pharmacy-dispensed medications [19, 20]. This database includes data from >200 million unique individuals from all 50 US states and the District of Columbia, with >120 contributing employers, 500 hospitals, and 40 health plans [20]. This activity was reviewed by the Centers for Disease Control and Prevention (CDC) and was conducted consistent with applicable federal law and CDC policy (see, eg, 45 CFR part 46, 21 CFR part 56; 42 USC §241[d]; 5 USC §552a; 44 USC §3501 et seq).

Study Design and Population

We identified adults aged 18 to 64 years with a sinusitis diagnosis in an outpatient setting from 1 January 2016 to 31 December 2020. Sinusitis was identified by ICD-10-CM diagnosis codes (J01.00–J01.91; International Classification of Diseases, Tenth Revision, Clinical Modification). We required continuous enrollment and prescription drug coverage during the 180-day baseline period before the sinusitis diagnosis date. We considered an oral antibiotic prescription dispensed between the diagnosis date (day 0) and up to 3 days after to be associated with the encounter; encounters without an antibiotic prescription were also included. Individuals were eligible to have multiple encounters for sinusitis during the study period.

Using diagnosis codes from the ICD-10-CM, procedure codes from the ICD-10 Procedure Coding System, and codes from the Current Procedural Terminology/Healthcare Common Procedure Coding System, we aimed to include only encounters associated with incident acute sinusitis in the cohort. To capture encounters that reflected the clinical management indicated for most individuals in the general population with uncomplicated acute sinusitis, we excluded those resulting in prescriptions with unusual antibiotic treatment durations (1–2 or >14 days). We also excluded encounters of adults with a concurrent diagnosis of a condition for which antibiotics were warranted (eg, pneumonia, urinary tract infection, skin and soft tissue infection; Supplementary Table 1). To restrict the study population to adults with a lower likelihood of infection due to antimicrobial-resistant bacteria, we excluded encounters in which adults received hospice care or mechanical ventilation or had a serious underlying medical condition (eg, malignancy, autoimmune conditions, or receipt of immunosuppressants) within the previous 180 days of the sinusitis diagnosis date, as well as encounters of adults who had an inpatient or skilled nursing facility admission or prior antibiotic use (intravenous, intramuscular, oral) within the previous 30 days (Supplementary Tables 2 and 3).

Outcomes and Sinusitis-Related Antibiotic Prescription Definitions

The primary outcomes were receipt of an antibiotic prescription for acute sinusitis within 3 days of the sinusitis diagnosis date and whether that antibiotic selection was guideline concordant, as defined by the IDSA clinical practice guideline for acute bacterial rhinosinusitis [9]. First- and second-line antibiotics were categorized as guideline-concordant antibiotic selection. All other antibiotics were categorized as guideline discordant. First-line antibiotic agents included amoxicillin-clavulanate or amoxicillin. Although amoxicillin is not considered a first-line agent by the IDSA guideline, it is recommended for the initial management of acute sinusitis by other clinical practice guidelines, and recent evidence has demonstrated that amoxicillin has similar efficacy to amoxicillin-clavulanate for adults with acute sinusitis but with fewer gastrointestinal adverse events [21, 22]. Second-line antibiotics included doxycycline, levofloxacin, or moxifloxacin. The secondary outcome was guideline-concordant duration of therapy among events with a guideline-concordant antibiotic; prescriptions with a supply ≤7 days were classified as guideline-concordant duration, and those >7 days were classified as guideline-discordant duration.

Covariates

We assessed several patient- and prescriber-level factors, including patient age at the time of the encounter for sinusitis, patient sex (reported as biologic sex), rural-urban status, geographic region (Northeast, Midwest, South, and West), type of outpatient setting (office, urgent care, emergency department, outpatient hospital, and walk-in retail clinic), and prescriber specialty/type (family medicine, internal medicine, emergency medicine, pediatrics, otolaryngology, allergy/immunology, nurse practitioner, and physician associate). In-person and virtual encounters led by any prescriber specialty/type practicing at an office-based outpatient setting were categorized as office encounters. Individuals were classified as living in an urban area if the primary beneficiary resided in a metropolitan statistical area according to the US Census Bureau geographic mapping classification [20, 23]. Primary beneficiaries not residing in a metropolitan statistical area were classified as living in a rural area.

Statistical Analyses

We summarized the proportion of eligible encounters with antibiotic selection and duration that was guideline concordant (first- and second-line) and guideline discordant, as well as visits with no antibiotic prescriptions across patient- and prescriber-level factors. To examine the relationship between patient- and prescriber-level factors and guideline-concordant antibiotic prescribing by agent and duration, we used modified Poisson regression models to estimate risk ratios and 95% CIs for each outcome, adjusting for age, patient sex, rural-urban status, geographic region, and outpatient setting, as appropriate. All analyses were conducted with SAS version 9.4 (SAS Institute Inc). This study followed the STROBE guidelines (Strengthening Reporting of Observational Studies in Epidemiology) [24].

RESULTS

Cohort Characteristics

A total of 4 689 850 encounters for acute sinusitis met our inclusion criteria (Figure 1). There were 3 141 440 unique patients included, with an average of 1.5 encounters per patient during 2016 to 2020. Acute sinusitis encounters consisted of adults who were 45 to 64 years of age (49.3%) and female (65.5%) and lived in urban areas (74.4%) and the South (51.2%). Encounters often occurred in an office setting (81.1%) and were commonly linked to a family medicine physician (39.3%; Supplementary Table 4).

In total, 815 543 (17.4%) sinusitis encounters did not result in an antibiotic prescription (Table 1). About 21% of encounters for adults aged 18 to 24 years, 18% for adults aged 25 to 44 years, and 16% for adults aged 45 to 64 years resulted in no antibiotic prescription. Most encounters that occurred at an emergency department did not result in an antibiotic prescription (57.3%). There were 3 874 307 (82.6%) sinusitis encounters with an antibiotic prescription; most were for amoxicillin-clavulanate (38.5%), followed by macrolides (22.0%), amoxicillin (13.4%), and doxycycline (9.0%; Table 2). Of all sinusitis encounters, 2 490 562 (53.1%) resulted in an antibiotic prescription for a guideline-concordant agent (42.9% first-line and 10.2% second-line) and 1 383 745 (29.5%) in a guideline-discordant agent. Encounters for adults residing in rural areas yielded a slightly higher proportion of prescriptions for a guideline-discordant agent as compared with encounters for those residing in urban areas (33.9% vs 28.6%). Encounters at an office setting and urgent care frequently resulted in a guideline-discordant agent (30.5% and 27.5%, respectively). Overall, the median treatment duration for adults with acute sinusitis was 10 days (IQR, 7–10), and the median durations for encounters resulting in guideline-concordant and guideline-discordant durations were 5 days (IQR, 5–7) and 10 days (IQR, 10–10), respectively. About 75% of first-line agents and 63% of second-line agents were prescribed for >7 days (Table 3). Additional patient information and prescriber-level factors by overall antibiotic prescribing, guideline-concordant agent selection, and guideline-concordant duration are shown in Supplementary Table 5.

Association Between Encounter Characteristics and Acute Sinusitis Management

Encounter characteristics associated with any antibiotic prescribing, guideline-concordant prescribing, and guideline-concordant durations included patient age group, patient sex, region, rural-urban status, and outpatient clinical setting (Table 4). Specifically, encounters among adults aged 25 to 44 years vs 18 to 24 years were associated with any antibiotic prescribing (adjusted risk ratio [aRR], 1.04; 95% CI, 1.04–1.04), increased prescribing for a guideline-concordant agent (aRR, 1.05; 95% CI, 1.04–1.05), and increased guideline-concordant duration (aRR, 1.21; 95% CI, 1.20–1.22). Female patients were less likely to receive a guideline-concordant agent (aRR, 0.97; 95% CI, .97–.98), but among those prescribed a guideline-concordant agent, they were more likely to receive guideline-concordant durations (aRR, 1.05; 95% CI, 1.05–1.06). Encounters among adults residing in the Southern or Northeastern regions vs the Western region were associated with increased guideline-concordant durations (South vs West: aRR, 1.03 [95% CI, 1.02–1.04]; Northeast vs West: aRR, 1.02 [95% CI, 1.01–1.02]).

Encounters among adults living in rural areas vs urban areas were 8% less likely to receive a guideline-concordant agent (aRR, 0.92; 95% CI, .92–.92) and 23% less likely to receive a prescription with a guideline-concordant duration (aRR, 0.77; 95% CI, .76–.77). Finally, encounters at an urgent care vs office setting were associated with any antibiotic prescribing (aRR, 1.03; 95% CI, 1.03–1.03) and guideline-concordant prescribing (aRR, 1.02; 95% CI, 1.02–1.03) and were 24% less likely to receive a guideline-concordant duration (aRR, 0.76; 95% CI, .75–.76).

DISCUSSION

In this nationwide cohort study of commercially insured adults diagnosed with acute sinusitis in the outpatient setting, most encounters resulted in an antibiotic prescription, with nearly one-third of encounters resulting in a guideline-discordant agent. Additionally, about 75% of encounters with a guideline-concordant agent had a prescription with a longer-than-recommended duration. We found that patient age, sex, rural-urban status, and type of outpatient clinical setting were associated with antibiotic prescribing for acute sinusitis. Although we observed a statistically significant association between guideline-concordant antibiotic agent selection and certain factors (eg, patient age, sex), such associations had a small magnitude of effect and might be less clinically relevant than the discrepancies identified with prescribing guideline-concordant durations.

Antibiotic therapy for acute sinusitis is often unnecessary [9, 10]. In our study, 17% of encounters did not result in an antibiotic prescription. Although we were unable to assess the clinical criteria used to diagnose acute bacterial sinusitis, this percentage likely reflects the overtreatment of adults with antibiotic therapy. Most cases of acute sinusitis have a viral etiology, and a recent cross-sectional study showed that approximately 50% of encounters for sinusitis do not warrant an antibiotic prescription [9, 11]. In this study, only encounters at an emergency department had such high proportion, with 57% of them not resulting in an antibiotic prescription. A cross-sectional study based on data from national health care surveys in the United States found that presenting to an emergency department for acute sinusitis was not associated with receipt of an antibiotic prescription and, in certain geographic regions, primary care office visits may be more likely to result in an antibiotic prescription [25]. Therefore, it is possible that certain strategies are being implemented in emergency departments to decrease antibiotic prescribing for certain common outpatient conditions, such as sinusitis. Watchful waiting and delayed prescribing are reasonable strategies for the initial management of patients with acute sinusitis with mild symptoms who do not meet the stringent clinical criteria for establishing a bacterial infection [9, 21]. Antibiotic stewardship programs can focus on promoting these strategies and addressing barriers that prescribers might face, such as perceived patient expectations and concern for patient satisfaction. Evidence shows that outpatient antibiotic prescribing can be improved and patients may still be satisfied without an antibiotic if their communication expectations are met [26]. Our study also shows that even when an antibiotic is prescribed, opportunities still exist to optimize guideline-concordant agent selection and duration for acute sinusitis.

Clinical guidelines recommend against the use of macrolides for acute sinusitis [9, 21]. Despite this, we found that azithromycin was the second-most prescribed antibiotic. This finding is consistent with other studies showing frequent prescribing of macrolides for acute sinusitis [8, 11]. Furthermore, we observed that even when guideline-concordant agents are selected, these are predominantly prescribed for longer than recommended, which can lead to excess patient harm. C difficile infection is a well-known antibiotic-associated adverse event, and it has been shown that the risk of this infection increases with antibiotic duration [4, 5]. This risk can be reduced by prescribing guideline-concordant durations, leading to similar outcomes and fewer adverse events as compared with longer durations [27]. Therefore, it is important that antibiotic stewardship programs develop interventions aimed at improving overall prescribing for acute sinusitis. Successful strategies that have increased guideline-concordant prescribing for acute upper respiratory tract infections, such as acute sinusitis, have followed the CDC’s “Core Elements of Outpatient Antibiotic Stewardship” [28]. For example, a before-and-after interrupted time series analysis showed that a multimodal intervention consisting of a targeted education campaign and a prescribing dashboard significantly reduced inappropriate prescribing rates in all outpatient specialties [29]. Another health care system implemented clinical decision support tools to facilitate the gathering of relevant information that would help ensure guideline-concordant diagnoses and provide nudges for guideline-concordant prescribing [30]. When these clinical decision support tools were combined with audit and feedback, the health care system observed increased guideline-concordant antibiotic use [30]. Antibiotic stewardship programs can adopt and tailor such strategies based on additional patient- and prescriber-level factors.

Age appeared to be a patient-level factor associated with antibiotic prescribing for acute sinusitis, especially regarding duration of treatment. We observed that encounters for adults aged 25 to 64 years were more likely to have a prescription with guideline-concordant duration as compared with encounters for younger adults aged 18 to 24 years. When all outpatient encounters for younger adults were evaluated, this patient population had a high proportion of encounters resulting in no antibiotic prescription (21%) or in a guideline-discordant agent (29%). These encounters were excluded from the regression model so that we could evaluate the selection of treatment duration only from encounters that resulted in a guideline-concordant agent. It is possible that our finding might reflect a difference in the clinical management of younger adults with acute sinusitis. Since incidence rates of sinusitis are higher among adults aged 45 to 74 years, there may be differences in the type or specialty of prescribers seeing young adults with sinusitis, which could result in different prescribing practices when compared with other adult patient populations [31].

We observed that encounters among adults residing in rural areas were more likely to result in an antibiotic prescription and were also more likely to have a longer duration as compared with encounters in urban settings. These findings are consistent with other studies showing that rural residence is associated with a higher likelihood of receiving an antibiotic prescription and for that prescription to be inappropriate [32–34]. These differences in prescribing practices for patients could be attributed to various patient- and prescriber-level factors. For example, patients residing in rural areas may have to travel longer distances to see a prescriber and may be more likely to request an antibiotic prescription during their encounter [35, 36]. In addition to potential patient expectations, prescribers may feel pressured to prescribe broader-spectrum antibiotics and longer durations if they are concerned that patients may develop complications from the infection and be unable to access follow-up care [36]. Finally, the number of late-career prescribers practicing in rural areas continues to increase, and some studies suggest that they are more likely to prescribe longer courses of antibiotics [37, 38].

The type of outpatient setting was associated with prescribing practices; as compared with office settings, urgent care encounters were less likely to have a guideline-concordant duration. A cross-sectional study assessing prescribing practices across an integrated health care system also found that urgent care encounters for acute sinusitis were significantly less likely to result in a prescription with a guideline-concordant duration [39]. Urgent care prescribers might be more inclined to prescribe an antibiotic and a longer course of therapy potentially due to less familiarity with treatment recommendations and guidelines, a lack of a relationship with patients, or concerns that patients may not seek additional follow-up if their condition worsens or will not use urgent care in the future because their expectations were not met [40, 41]. Antibiotic stewardship programs that have implemented interventions aimed at addressing patient- and prescriber-level factors influencing prescribing in the urgent care setting, such as initiatives addressing patient expectations, have successfully improved antibiotic prescribing [42].

This study has several potential limitations. First, using ICD-10 diagnosis codes to identify the study population may result in misclassification. Second, administrative claims cannot ascertain the severity of infection and other factors (eg, signs or symptoms of infection) that could influence antibiotic prescribing and the assessment of whether the antibiotic prescriptions were clinically indicated. Third, the Merative MarketScan Commercial Database does not capture dosing information, and we were unable to include antibiotic agent dosing in the definition of guideline-concordant prescribing. Fourth, we did not exclude encounters for patients with a previous sinusitis diagnosis (5% of total encounters had a sinusitis diagnosis in the preceding 30 days). Prescribers may choose non–first-line antibiotics with a broader or different activity spectrum than first-line agents for patients with recurrent acute sinusitis. However, given the categorization of second-line agents (eg, fluoroquinolones) more commonly prescribed for recurrent cases as guideline concordant, the effect of including these encounters may be minimized. Fifth, we did not include prescriber type/specialty in our regression analysis because of potential collinearity with type of outpatient setting. Sixth, additional patient characteristics related to health inequities in health care access and antibiotic prescribing, such as race and ethnicity, income, or education level, were not assessed since these variables were not available in these data [43]. Last, this study included only commercially insured adults aged 18 to 64 years and thus may not be generalizable to other age groups or populations. Despite these limitations, this study characterized patient- and prescriber-level factors associated with appropriate antibiotic prescribing for acute sinusitis in a large cohort of commercially insured adults from all 50 US states and the District of Columbia.

In conclusion, this study demonstrates that specific patient- and prescriber-level factors are associated with the guideline-concordant management of acute sinusitis among commercially insured adults. Specifically, our results indicate that rural areas and urgent care settings had lower guideline-concordant antibiotic prescribing for acute sinusitis. Antibiotic stewardship activities should be tailored to the patient population, prescriber type/specialty, and type of outpatient setting where they are being implemented. Opportunities exist to optimize antibiotic prescribing, especially to reduce unnecessary prescribing and duration of therapy for acute sinusitis.

Supplementary Data

Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Author contributions. A. A. V. D., A. M. B., L.A.H., and S. K. conceived and designed the study. D. J. B., and J. M. S. performed the statistical analyses. A. A. V. D. drafted the initial manuscript. All authors contributed to the interpretation of the results and revised the manuscript for substantial content. All authors read and approved the final manuscript.

Disclaimer. The findings and conclusions in this article are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.

Patient consent. This study does not include factors necessitating patient consent.

Financial support. No funding external to the Centers for Disease Control and Prevention was provided for this study.

References

Author notes