Measuring appropriate use of antibiotics in pyelonephritis in Belgian hospitals

https://doi.org/10.1016/j.cmpb.2008.12.002Get rights and content

Abstract

Inappropriate use of antibiotics can induce antibiotic resistance, treatment failure, increased costs and even mortality. We developed a methodology for measuring guideline compliance of hospital antibiotic prescriptions in community-acquired acute pyelonephritis in Belgium. The claims and clinical data of all Belgian hospitalizations for community-acquired acute pyelonephritis were extracted from a nationwide administrative database. In a clinically homogeneous subset of patients, the percentage of patients who received a guideline-compliant prescription was calculated according to prescription guidelines disseminated in Belgium. In the group of non-pregnant adult female patients, 31% of the prescriptions were not in strict compliance with the guideline. Interhospital variability ranged from 0% to 100% compliance. We conclude that administrative databases can be used to analyze antibiotic prescription behavior in hospitals for homogeneous and clinically relevant patient groups. The interhospital variability observed in Belgian hospitals indicate that there is a clear room for improvement.

Introduction

Antibiotic resistance has become a worldwide concern, with major consequences such as treatment failure, prolonged or additional hospitalizations, increased costs and mortality [1]. In the 1980s susceptibility of community-acquired pyelonephritis approached 75% for ampicillin and 100% for co-trimoxazole. By the mid-1990s, these susceptibilities had fallen to 50–60% and 70–82%, respectively [2], [3]. Inappropriate use of antibiotics can induce antibiotic resistance in the general population [4], [5], [6], [7], [8]. Furthermore, the inappropriate use can increase resistance to antibiotics of the same class but also of other classes. Hospitals’ antibiotic policy may affect not only current patients but also future patients [9].

Studies from different countries have already reported broad variations in antibiotic prescriptions in hospitals, often linked to different practices between disciplines [10], [11], [12], [13]. If guidelines are not implemented, additional policy measures should be considered nationally and at the level of the individual prescriber. The use of already validated retrospective administrative databases is a convenient and feasible way to analyze hospital practices without investing excessive time and resources for new data collections [14], [15], [16].

Acute pyelonephritis is a common indication for antibiotic prescription in hospitals. It can cause Gram-negative bacteremia and community-acquired sepsis. Evidence-based international and local guidelines exist for this pathology. The Belgian clinical guideline on antibiotic treatment of community-acquired acute pyelonephritis in hospitalized immunocompetent adult was published in 2002 by the Belgian Antibiotic Policy Coordination committee of the Ministry of Public Health (BAPCOC) [17]. For this feasibility study, acute pyelonephritis was chosen as it was thought to be a relatively easy and homogeneous pathology to delineate in administrative data. The data available at the time of the study dated from 2000, thus before the publication of the Belgian clinical guideline. The Belgian edition of the Sanford guide to Antimicrobial Therapy however, was a very widespread reference in Belgium at that time [18].

Section snippets

Aims of the study

To assess appropriateness of antibiotic use for acute pyelonephritis in Belgian hospitals based on nationwide administrative hospital databases.

Source data

The national administrative clinical database (Résumé Clinique Minimum/Minimale Klinische Gegevens) collected by the Ministry of Public Health twice a year from all acute care hospitals, includes administrative information (month of admission, age, sex, length of stay, death during hospitalisation, etc.), ICD-9-CM diagnoses and procedures codes for each stay. The All-Patient Refined Diagnosis Related Group (APR-DRG) and level of severity were available. On the other hand, the financial database

Results

The initial population of patients filtered on diagnoses, alive at discharge, consisted of 5275 stays spread between 124 acute care hospitals. 82.1% were female patients. Mean age was 27.7 year (median = 23, interquartile range (IQR) = 4–42). The male population was generally younger: mean = 24.8 year (median = 8, IQR = 0–50) versus mean = 28.4 year (median = 24, IQR = 7–41) for female patients. The age distribution showed a superposition of three subpopulation peaks: a pediatric population, an adult

Discussion

This feasibility study showed that administrative databases can be used to delineate a homogeneous group of patients and that they can be used to audit antibiotic prescriptions in hospitals. However, these data have their limitations too. First, tertiary validation on the field revealed that invoiced doses can be overestimated, especially when oral products are delivered to a patient for the days following discharge. In that case, the claim does not reveal the exact dose taken during the stay.

Conclusion

This feasibility study showed that, taking into account a number of methodological limitations, administrative databases can be used to analyze antibiotic prescription behavior for homogeneous patient groups in hospitals. Globally, clinicians in Belgian hospitals apply practice guidelines on antibiotic use. In a minority of the hospitals however, there is clear room for improvement in the choice of prescription. In hospitals where deviations are a matter of concern, measures should be taken to

Conflict of interest

Some of the members of the expert group declared conflicts of interest related to affiliations such as the Infectious Disease Advisory Board and the Sanford group or to funding of scientific studies by AstraZaneca, Bristol-Myers Squibb or GlaxoSmithKline. The authors of the paper have no conflict of interest to declare.

Acknowledgements

Grateful acknowledgment is hereby made to all the experts for their scientific input during the study as well as their hospital team for the data validation: Dr. Filip Ameye, urologist, General Hospital Maria Middelares (Ghent), Pr. Willy Peetermans, infectiologist, President Hospitals working group of the Belgian Antibiotic Policy Coordination committee of the Ministry of Public Health (BAPCOC) and University Hospital Leuven (Leuven); Dr. Jacky Peeters, nephrologist, Hospital

References (25)

  • R.A. Weinstein

    Controlling antimicrobial resistance in hospitals: infection control and use of antibiotics

    Emerg. Infect. Dis.

    (2001)
  • D. Guillemot et al.

    Evolution of antimicrobial resistance: impact on antibiotic use

    Semin. Respir. Crit. Care Med.

    (2002)
  • Cited by (0)

    1

    Tel: +32 3 234 45 14; fax: +32 3 234 49 16.

    View full text