Discussion

Over a 6-year period, there was an overall decline in ambulatory antibiotic prescription rates, but trends were not consistent across all antibiotics, demographics, and prescribing providers. Macrolides and fluoroquinolones showed the largest decrease, and lincosamides was the only class that increased. This finding contrasts with data from the United States from 2006 to 2010, which did not show a significant decline in macrolides or fluoroquinolones.^{Reference Suda, Hicks, Roberts, Hunkler and Taylor8} This difference may indicate a recent trend toward decreased macrolide use, likely driven by reduced susceptibility of Streptococcus pneumonaie.^{Reference Zhang, Chen, Yu, Pan and Liu11} The decreased fluoroquinolone use may be due to the recent US Food and Drug Administration (FDA) safety alerts discouraging their use for minor infections.^{Reference Yarrington, Anderson and Dodds Ashley12}

Antibiotic prescription rates were significantly higher in females, particularly in young adults in whom women received more than twice as many antibiotics, which is similar to previous reports.^{6,Reference Patrick, Marra, Hutchinson, Monnet, Ng and Bowie13,Reference Kristensen, Johnsen and Thomsen14} We postulate that the higher rates in women are due to higher healthcare utilization and certain infections (eg, urinary tract infections) that occur more often in females. This trend, unfortunately, leads to more community-onset C. difficile infections in women.^{Reference Younas, Royer and Rac15}

The greatest decrease in antibiotic prescription rates was observed in children. This significant decrease (26%) is consistent with results of a recent study reporting significant decline in antibiotic prescribing for children from 2000 to 2014.^{Reference Finkelstein, Raebel, Nordin, Lakoma and Young16} One possible explanation is the impact of recent guidelines for management of otitis media, although without additional interventions this may have had only a modest effect.^{Reference Lieberthal, Carroll and Chonmaitree17,Reference Deniz, van Uum, de Hoog, Schilder, Damoiseaux and Venekamp18} Another potential factor is the increase in childhood vaccinations against S. pneumonaie, Haemophilus influenzae, and influenza leading to a reduction of common respiratory infections.^{Reference Morris, Moss and Halsey19–Reference Younas, Royer and Weissman21}

During each year, the highest prescription rates were observed during winter months. This finding is similar to previously reported seasonal variations.^{Reference Suda, Hicks, Roberts, Hunkler and Taylor8,Reference Patrick, Marra, Hutchinson, Monnet, Ng and Bowie13,Reference Elseviers, Ferech, Vander Stichele and Goossens22,Reference Curtis, Walker, Mahtani and Goldacre23} These trends are likely due to increases in acute respiratory infections during winter, although many of these are viral. Educational campaigns for the general population and healthcare providers may be useful tools for improvement, but implementation in community settings remains challenging. Increased utilization of rapid diagnostics may also reduce antibiotic use.^{Reference Cals, Butler, Hopstaken, Hood and Dinant24,Reference Andreeva and Melbye25} These efforts are critical because seasonal variation in resistance rates has been associated with seasonality in ambulatory antibiotic prescriptions.^{Reference Hoberman, Paradise, Greenberg, Wald, Kearney and Colborn26,Reference Ramsey, Royer and Bookstaver27}

The finding of higher antibiotic prescription rate in rural settings is consistent with previous results.^{Reference Curtis, Walker, Mahtani and Goldacre23,Reference Gonzales, Steiner and Sande28–Reference Barlam, Soria-Saucedo, Cabral and Kazis30} This finding may be partly due to concerns of patients’ long travel distance and difficulty returning to care if symptoms persist, making watchful waiting less applicable.^{Reference Chow, Benninger and Brook31} This phenomenon may explain overall high antibiotic prescription rates in the southeastern United States, where large proportion of residents live in rural communities.

This study demonstrates only 66.2% of all antibiotics are prescribed by MDs or DOs, which contrasts with the findings of a national study from 2005 to 2010, which found that 81% of antibiotics were prescribed by physicians.^{Reference Suda, Roberts, Hunkler and Taylor32} This shift toward less MD/DO prescribing and more NP/PA prescribing is multifactorial but is greatly influenced by increasing numbers of advanced practice providers (69% increase).

Unfortunately, prescriber rates did not decrease as much among NPs and PAs as among MDs and DOs. This difference may be due to disproportional numbers of NPs and PAs working in urgent care centers or retail clinics, which have been associated with increased antibiotic prescribing.^{Reference Palms, Hicks and Hersh33} This finding highlights the need to intervene to improve prescribing by NPs and PAs, who are not reached by typical medical societies or academic meetings. Dentists prescribed 5.9% of antibiotics, compared with 10.4% in the prior national study; however, dentist prescriber rates increased over the study period.^{Reference Suda, Roberts, Hunkler and Taylor32} Dentists prescribed more than one-third of all clindamycin. Because of this increase in antibiotic prescribing overall and for antibiotics with high risk of C. difficile infection, educational and clinical stewardship interventions tailored for dental providers are necessary and have been shown to be effective.^{Reference Deshpande, Pasupuleti and Thota34,Reference Gross, Hanna, Rowan, Bleasdale and Suda35}

This study has several limitations. Antibiotic prescription is a surrogate for antibiotic use because prescription filled does not necessarily mean the medication was administered. Also, we did not assess indications to allow assessment of appropriateness. These data represent only 30% of the South Carolina population and may not be representative of other demographics. Individuals ≥65 years old are not represented due to lack of Medicare prescription claims.

In conclusion, ambulatory antibiotic prescription rates in South Carolina have decreased from 2012 to 2017. Understanding differences in prescribing rates with regard to patient and prescriber characteristics is an important tool to develop specific ambulatory antibiotic stewardship interventions to influence prescription behaviors for these populations. Statewide efforts are needed to further improve ambulatory antibiotic stewardship focused on older adults and acute respiratory infections during the winter months. Other states can take similar approaches to evaluate prescribing practices and designate regional interventions for improvement.