Fluoroquinolone prescribing patterns in the private healthcare sector of South Africa, 2005-2012

Abstract

Increased prescribing of fluoroquinolones leading to resistance is an issue of concern at present globally. The present study was designed to investigate prescribing patterns for fluoroquinolones in a section of the South African private healthcare sector since 2005-2012 and is a retrospective analysis of the medicine claims database of a nationally (South African) representative Pharmacy Benefit Management (PBM) company data between 2005 and 2012. The target population consisted of a total of 3788438 patients older than 18 years (male/female ratio 1.2:1) claiming at least one antibiotic prescription. Fluoroquinolones were defined as all active substances available in South Africa at the time of the study, belonging to the J01MA classification of the ATC index. Data obtained included the NAPPI (National Pharmaceutical Product Index) code (a unique product identifier), the number of dosages dispensed and the number of days’ supply. The data obtained were expressed in DDD/1000 inhabitants/day. The average DDD per prescription per patient per year was also determined to describe the trends over the study period. Descriptive and inferential statistics were used to analyse the data using the SAS Version 9.3. All statistically significant results were considered with a probability of p < 0.05. Fluoroquinolones represented 28% of all antibiotic prescriptions (N=7069563) claimed during the study period. A mean of 1.45 ± 0.92 (95% CI 1.44 - 1.45) prescriptions were claimed per patient in 2005, 1.46 ± 0.92 (95 % CI 1.45-1.46) in 2006, 1.47 ± 0.93 (95 % CI 1.47-1.48) in 2007, 1.45 ± 0.90 (95 % CI 1.44-1.45) in 2008, 1.41 ± 0.85 (95 % CI 1.41-1.42) in 2009, 1.40 ± 0.833 (95 % CI 1.39-1.40) in 2010, 1.36 ± 0.74 (95 % CI 1.33-1.34) in 2011 compared to 1.31 ± 0.71 (95% CI 1.31-1.32) during 2012 (Cohen’s d=0.2). The association between number of prescriptions and gender (p < 0.0001, Cramer’s V = 0.02) and age groups (p < 0.0001, Cramer’s V = 0.04) was statistically but not practically significant. Ciprofloxacin [1.30 DID (2005), 1.42 DID (2006), 1.70 DID (2007), 1.63 DID (2008), 1.51 DID (2009), 1.48 DID (2010), 1.29 DID (2011) vs. 1.19 DID (2012)], levofloxacin [0.46 DID (2005), 0.58 DID (2006), 0.82 DID (2007), 0.95 DID (2008), 1. 04 DID (2009), 0.97 DID (2010), 0.79 DID (2011) vs. 0.67 DID (2012)] and moxifloxacin (0.51 DID (2005), 0.60 DID (2006), 0.80 DID (2007), 0.70 DID (2008), 0.62 DID (2009), 0. 57 DID (2010), 0.51 DID (2011) vs. 0.44 DID (2012)] were most claimed. The average DDD/prescription/patient/year for ciprofloxacin increased from 4.12 ± 3.21 (95% CI 4.10-4.13) in 2005 to 4.84 ± 2.27 (95% CI 4.83-4.86) in 2012 (p < 0.0001, Cohen’s d = 0.2), vs. 5.47 ± 4.57 (5.43- .52) in 2005 to 7.52 ± 3.72 (7.48-7.56) during 2012 for levofloxacin (p < 0.0001, Cohen’s d = 0.5), and 5.89 ± 5.45 (5.85-5.94) in 2005 to 5.90 ± 4.84 (5.84-5.96) in 2012 for moxifloxacin (p < 0.0001, Cohen’s d = 0.002). The study concluded that the number of patients receiving fluoroquinolones generally decreased in the private healthcare sector during the study period. Gender and age were found as weak predictors of fluoroquinolones use. The ATC classification and the DDD unit of measurement have provided a platform to compare drug use at various healthcare levels in different settings and geographic locations.