Our study showed that most ARTIs treated with antibiotics in primary care and EDs were bronchitis, tonsillitis, COPD exacerbation, rhinopharyngitis and sinusitis. The most used classes of antibiotics were penicillins accounting for more than 58% of the total antibiotics prescribed for ARTIs. Among these, the most commonly prescribed penicillin was amoxicillin clavulanate followed by amoxicillin. Fluoroquinolones accounted for 17.6% of all antibiotic prescriptions, and 49% of these were levofloxacin. Macrolides and cephalosporins were far less frequently prescribed. In 75.8% of cases, antibiotic therapy should not be prescribed. Inappropriate antibiotic prescription as assessed by MAI was mostly observed in acute bronchitis and in patients treated with amoxicillin-clavulanic acid or levofloxacin. Comorbidities were significantly associated with inappropriate antibiotic prescription.
There is clear evidence that antibiotics are heavily overprescribed for respiratory infections because most of these infections are of viral origin and self-limited conditions (17–19). Their prescription rate ranged between 20 and 90% in Europe (12, 20, 21) and 50 to 70% in United States (21). Our study highlighted the worldwide variation in types of RTIs treated and patterns of antibiotics used. In a study conducted in the UK (22) targeting primary care settings, 73% of antibiotic prescriptions used in the treatment of upper respiratory tract infections were penicillins which is similar to our findings. According to a tertiary medical institution study conducted in Beijing (23), the most commonly prescribed classes of antibiotics for ARTIs were cephalosporins (41%). In Japan, cephalosporins constituted 41.9% of all antibiotic prescriptions and penicillins accounted for just 8.0% (24). In our study, we noted a frequent use of broad-spectrum antibiotics, amoxicillin clavulanic acid and levofloxacin represented almost two thirds of all antibiotics prescribed. This practice is not appropriate as it is recommended that narrow-spectrum antibiotics should be maintained at ≥80% in cases prescribed an antibiotic, while the proportion of fluoroquinolones should be maintained at ≤5% (25, 26). Overall, the quality of prescribing was inappropriate in our study as attested by MAI score. Similar results were observed in the United States and other developed countries (27–30). The most common MAI item involved was expensiveness and indication while the antibiotics that were most often prescribed inappropriately were amoxicillin clavulanic acid and levofloxacin. In countries with limited health resources, this indiscriminate use of antibiotics in ARTIs may result in increased health care cost. In the era of increased bacterial resistance, the need to restrict antibiotic prescription with special emphasis to narrow spectrum ones is more than urgent. Our study is the first to investigate physician practice in Tunisian EDs where the utilization rate of antibiotics for ARTIs could exceed the rate of ambulatory setting. High-volume workload, high-acuity nature of ED clinical presentation, and specificity of patient-physician relationships in the ED could explain why ED physician are more exposed to prescribe antibiotics inappropriately. In a study conducted in United States including ED visits with a diagnosis of ARTI, it was found that approximately 40% of antibiotic prescriptions were inappropriate (9). Improving the appropriate use of antibiotics in ARTIs in primary care or EDs should take into account the factors that could be implicated in this phenomenon. Available data indicate the existence of a great variation between countries with regard to the factors associated with inappropriate antibiotic prescription (12, 31). Patient expectation and physicians related factors such as diagnostic uncertainty, lack of awareness of specific guideline recommendations, and lack of time necessary to reassure the patient were among the principal reasons of antibiotic overprescription. Our study was focused on patients’ characteristics and we showed that history of coronary artery disease, asthma, and diabetes were the most important factors associated with antibiotherapy inappropriateness. Patients with diagnosis of acute bronchitis were also more likely to receive antibiotics inappropriately.
There are a number of potential limitations to note. First, although our study included a large sample representing overall clinical practice in Tunisia, we acknowledge that we did not include children who represent some of the highest users of antibiotic prescriptions. Second, in this study we applied the MAI score to assess prescribing appropriateness in primary care and ED practice. Although this experience indicated that MAI score could be used successfully in different setting to detect indicators of inappropriate prescribing, we did not assess the correlation of this score with clinical outcome. Hence, the question whether patients taking antibiotics with more appropriate ratings had better clinical outcomes than patients taking drugs with less appropriate ratings remains unanswered. Third, the rate of underprescription of antibiotics was not assessed because our study limited the analysis to patients for whom antibiotics were prescribed. Finally, it is possible that, for some prescriptions classified as inappropriate, there could be individual patient factors unknown to reviewers that might justify a provider’s decision to deviate from the guidelines.