Antibiotic usage at a clinical paediatric hospital before and after the implementation of actions related to the hospital antibiotic policy

Abstract

Objectives

The aim of this study was to determine the impact of undertaken interventions related to the hospital antibiotic policy (antibiotic stewardship programme (ASP)): participation in a point prevalence survey of healthcare-associated infections and antimicrobial use in European acute care hospitals (PPS ECDC), and Polish national programme to protect antibiotics (NPOA) on the modification of the amount and profile of antibacterial drug use at the Public Paediatric Teaching Clinical Hospital in Warsaw.

Methods

A retrospective analysis of antimicrobials (with daily defined doses (DDD) assignment) usage expressed in DDD and DDD/100 bed days (BD) in the period 2013–2017 (5 years) on 14 hospital wards at the Public Paediatric Teaching Clinical Hospital in Warsaw before and after the implementation of ASP-related actions was conducted.

Results

A total of 188 405.78 DDD were used and 553 485 paediatric BD were recorded in the wards selected for the present study in the period mentioned above. Wards with pre-authorisation duty for third-line antibiotics (group 1 of wards) used less DDD/100 BD (from 28.81 to 31.12 DDD/100 BD) than wards without such a duty (from 54.72 to 76.06 DDD/100 BD). We observed a temporary decrease of 6.37% in DDD/100 BD in group 1 of wards and a stable 9% to 21% decrease in DDD/100 BD tendency in group 2 of wards (wards without pre-authorisation duty: oncology, haematology and intensive care unit) compared with average values of DDD/100 BD in the period before ASP-related actions (2013–2014). Changes in drug utilisation (DU90%) profile were also observed, both positive and negative.

Conclusions

More frequent actions related to ASP, such as annual PPS ECDC participation and regular personnel education on the principles of antibiotic therapy, should improve and make antimicrobial treatment more rational.

Introduction

Hospital antibiotic policy is implemented mainly to limit the occurrence of antimicrobial resistance and related untreatable, severe infections that prolong hospitalisation, and increase the number of complications, patient deaths and treatment costs.^{1 2}

The actions of most significant influence on the implementation of the antibiotic stewardship programme (ASP) involve the creation of a hospital antibiotic list with a selection of antibiotics available only after pre-authorisation, the development of guidelines (recommendations) of hospital antibiotic usage, personnel education, current audits of antibiotic prescribing, and others.³ Another action supporting ASP may include participation in a point prevalence survey of healthcare-associated infections and antimicrobial use in European acute care hospitals (PPS ECDC) with analysis of obtained data and feedback to personnel.⁴

This study aimed to determine the impact of undertaken interventions related to ASP: PPS ECDC and participation in the Polish national programme to protect antibiotics (NPOA) on the modification of the amount and profile of antibacterial drug use at the Public Paediatric Teaching Clinical Hospital (PPTCH) in Warsaw.

To the best of our knowledge, this is the first study investigating the influence of PPS ECDC and educational intervention connected to participation in the Polish NPOA on the annual consumption of antibiotics at a paediatric teaching clinical hospital in Poland.

Methods

Antimicrobial (with daily defined doses (DDD) assignment) drug use was retrospectively analysed in the period 2013–2017 at the PPTCH in Warsaw before and after the implementation of ASP-related actions. The PPTCH in Warsaw has a hospital antibiotic list with a selection of third-line antibiotics available for clinicians to be prescribed only after pre-authorisation from a member of the antibiotic team and the head of the ward. Third-line antibiotics are: carbapenems, glycopeptides, cefepime, colistin, rifaximin, linezolid, ceftazidime, piperacillin and tazobactam. We divided the 14 hospital paediatric wards included in the study into two groups:

• Group 1—wards with pre-authorisation duty for third-line (restricted) antibiotics: Cardiology, Otolaryngology, Pulmonology, General paediatrics, Endocrinology, Diabetology, Gastroenterology, Cardiosurgery, Surgery and Urology, Nephrological and Neurological wards;

• Group 2—wards without pre-authorisation duty for third-line antibiotics: Oncology, Haematology and Intensive Care Unit (ICU) with postoperative care wards.

The following interventions were undertaken between 2015 and 2017 to improve ASP at the PPTCH in Warsaw: participation in the Polish national programme to protect antibiotics (NPOA), an educational intervention connected with it in 2015, and the ECDC Point Prevalence Survey (PPS) in 2015 and 2017. NPOA is a national, governmental programme, among the others, aiming to prevent antimicrobial resistance, analyse antibiotic usage trends, prepare recommendations for antibiotic treatment, and develop antibiotic stewardship in Polish hospitals.

Activities connected with NPOA participation and PPS ECDC carried out in the PPTCH for antibiotic stewardship were:

• multiple antibiotic team meetings involving the development and implementation of recommendations of antibiotic use: empirical therapy, surgical antibiotic prophylaxis, indications for antibiotics at the PPTCH in Warsaw

• at the end of 2015, an NPOA representative educated doctors from different wards on the principles of antibiotic therapy and hospital infection prevention

• the results of ECDC PPS were discussed during meetings of the Infection Control Committee and meetings of the antibiotic team, members of which in both cases are doctors and hospital pharmacists (one of them is head of Hospital Pharmacy).

Antimicrobials encompassed in the study included a total of 53 chemical substances—that is, antibiotics and chemotherapeutics (table 1).

Table 1.

Antibacterial drugs (antibiotics and chemotherapeutics) included in the study divided into chemical subgroups

Antimicrobial subgroups	Description of subgroups
Aminoglycosides	Amikacin, gentamicin, streptomycin, tobramycin
Third-generation cephalosporins	Ceftriaxone, cefotaxime, ceftazidime, ceftibuten, cefoperazone-sulbactam
Other cephalosporins	Cephalosporins (third generation excluded)
Fluoroquinolones	Ciprofloxacin, levofloxacin, ofloxacin
Glycopeptide antibacterials	Vancomycin, teicoplanin
Carbapenems	Meropenem, imipenem and cilastatin, ertapenem
Lincosamides	Clindamycin, lincomycin
Macrolides	Azithromycin, clarithromycin, erythromycin, roxithromycin, spiramycin
Metronidazole	Specified due to higher usage
Penicillins	Penicillins with or without β-lactamase inhibitors
Sulfamethoxazole and trimethroprim	Specified due to higher usage
Other	Antimicrobials not included in other antimicrobial subgroups and very rarely used

We obtained data from the hospital pharmacy computer program (DDD) and from the hospital medical statistics department (hospitalisation bed days (BD)). Calculations were based on DDD and DDD/100 BD. Statistical analysis was performed with the Wilcoxon test and linear regression analysis with EXCEL and STATISTICA software.

The study was divided into three parts:

• Part 1: an estimation of changes observed in the value of DDD/100 BD depending on ASP intervention in different ward groups in the period 2013–2017. The Wilcoxon test was used for statistical analysis and also descriptive statistics with percentage calculations.

• Part 2: the evaluation of changes in antibiotic subgroups (table 1) use in ward groups in the period 2013–2017 with linear regression analysis.

• Part 3: analysis of drug utilisation rates (DU90%) in the period 2013–2017 in group 3. DU90% is defined as the number of the most commonly used antibiotics that accounted for 90% of the total amount of DDD prescribed.⁵

Results

A total of 188 405.78 DDD were used and a total of 553 485 paediatric BD were recorded during the above-mentioned period in the 14 wards selected for the present study.

Part 1

DDD/100 BD was calculated for two groups of wards and for all wards in the study (AWIS) (table 2).

Table 2.

DDD/100 BD values in individual years

			Interventions: ECDC PPS, participation in NPOA, personnel education		Interventions: ECDC PPS
Wards/year	2013	2014	2015	2016	2017
Group 1	28.81	26.11	30.15	25.71	31.12
Group 2	66.17	72.43	76.06	62.78	54.72
AWIS	34.34	32.58	37.13	31.49	35.12

AWIS, all wards in the study; BD, bed days; DDD, daily defined doses; ECDC, European Centre for Disease Prevention and Control; NPOA, Polish national programme to protect antibiotics; PPS, point prevalence survey.

The findings were as follows:

• In group 1 DDD/100 BD values in each year were 1.8 to 2.8 times lower than in group 2, respectively

• The analysis of changes in antibacterial drug use each year with the Wilcoxon test revealed statistical significance (p<0.05) only for group 1 of wards in the years 2013–2014 (before the intervention related to ASP, p=0.023) and in 2016–2017 (after over 1 year following the intervention related to ASP, p=0.0499)

• Changes in antimicrobial drug use between the average value of DDD/100 BD in 2013–2014 and values of DDD/100 BD in 2015–2017 expressed in percentages were also calculated. We found that in 2015 (year of interventions related to ASP) the value of DDD/100 BD of antibacterial drugs increased by nearly 10–11% in all groups of wards.

However, a reduction in antibiotic DDD was noted a year after the intervention (data for 2016): it decreased by 9.41% of DDD/100 BD in group 2, 6.37% of DDD/100 BD in group 1, and 5.89% of DDD/100 BD in AWIS. Regrettably, after another year following the intervention (2017), despite PPS intervention, an increase in DDD/100 BD was noted in group 1 and AWIS compared with the mean values reported 2 years before the intervention (2013 and 2014). The respective values of the increase were 13.3% and 4.96%. However, group 2 maintained the downward trend and the antibiotic usage decreased by slightly over 21% compared with mean values of DDD/100 BD reported before the intervention. On the ward level changes in antibiotic use were diversified. In group 2 of wards, ICU with postoperative care wards achieved stabile 2 years lasting decrease in antibiotic use (over 20% reduction to the values of DDD/100 BD: 61.89 in 2016 and 56.86 in 2017) and in group 1 – Cardiology (17.7–21.3% reduction to the values of DDD/100 BD: 6.6 in 2016 and 6.3 in 2017, respectively). In contrast, General paediatrics presented the highest percentage 2 years lasting increasing trend in DDD/100 BD, from an average value of 12.91 in 2013–2014 to 19.55 (51.5% increase) in 2016 and 22.92 (77.6% increase) in 2017.

Part 2

As regards the years 2013–2017, statistically significant downward trends of DDD/100 BD (for average values) were observed in both groups of wards in the usage of aminoglycoside antibiotics (p=0.015 in group 1, p=0.036 in group 2) and sulfamethoxazole and trimethoprim (p=0.022 in group 1, p=0.003 in group 2). Moreover, an upward, statistically significant trend was noted in the usage of fluoroquinolones in group 1 (p=0.027). Figure 1 presents these trends for group 1 and group 2 as coloured lines (the same colour are average values of DDD/100 BD). Statistically significant differences in the use of antibiotic subgroups were observed in group 1 of the wards (figure 1) for aminoglycosides (decrease in DDD/100 BD between years 2016 and 2017, p=0.028) and cephalosporins (increase in DDD/100 BD between years 2015 and 2017, p=0.020).

Figure 1.

Use of antibiotic subgroups (average values with SD): aminoglycosides, sulfamethoxazole and trimethoprim, fluoroquinolones and cephalosporins expressed in DDD/100 BD in group 1 and group 2 of wards in 2013–2017. BD, bed days; DDD, daily defined doses.

Part 3

It was demonstrated that before 2016 the number of antibiotics included in DU90% had increased by one every year: from 16 in 2013 to 18 in 2015. After completing the intervention the number of DU90% antibiotics was stable at a level of 17 in 2016 and 2017. Both before the intervention in the years 2013–2014 and after the intervention in the years 2016–2017 the most commonly used antibiotic in AWIS was amoxicillin with clavulanic acid. Some changes were observed as regards the composition of antibiotics making up DU90%. More significant changes were qualified as positive or negative. Positive changes included the increased percentage of DDD of amoxicillin, ampicillin and cephazolin and their shift onto higher positions on the list of the most commonly used antibiotics following the ASP-related intervention in the years 2016–2017 (from position numbers 11, 14 and out of the DU90% to position numbers 6, 7 and 14, respectively). Similarly, a decrease in the use of amikacin and lowering its rank in the DU90% group to position number 17 or even a complete loss of position in this group was also considered as a positive change. Regrettably, implemented interventions failed to stop the increased use of ceftriaxone (from position number 7 to 4 in DU90%) and cefuroxime (from position number 3 to 2).

Discussion

There is satisfactory evidence in the literature that undertaking interventions related to ASP leads to a reduction in the quantity of DDD used. However, some studies also indicate that no such results have been obtained. A Polish study by Kuziemski from University Hospital in Bydgoszcz showed that ASP implementation led to a reduction of DDD/100 BD from 59.52 to 47.88 over several years.⁶ Nitsch-Osuch also described a reduction in antibiotic usage after ASP implementation at a paediatric department. It changed from 36.3 to 24.9 DDD/100 BD.⁷ However, other studies by the same author showed an increase in DDD/100 BD at oncological surgery wards in Warsaw⁸ (from 25.88 to 108.17 DDD/100 BD over 5 years) and a neonatal ward⁹ (from 28.9 to 30.8 DDD/100 BD after a year since the intervention) despite ASP-related actions. Data obtained from an NPOA pilot study report¹⁰ as of 2012 regarding antibiotic consumption showed that the average DDD/100 BD was at 43 (ranging from 28 to 56 DDD/100 BD) at Polish hospitals included in the study. The analyses comprised systemic antibacterial medications excluding drugs used for the treatment of conditions such as tuberculosis. The present paper included antituberculosis drugs, as it was assumed that, for example, rifampicin, which is a typical antituberculosis drug (J04 according to ATC¹¹), is also used in the treatment of various infections, usually severe ones, as part of a combination antibiotic treatment if previous treatment regimens proved ineffective. Notably, the use of antituberculosis drugs at the PPTCH in Warsaw is still at a very low level. Therefore, it is justified to refer our data to this report. As regards the years included in the study (2013–2017) we did not exceed the average of 37.13 DDD for all the wards in the study, which placed us below the average indicated in the NPOA report for 2012. It is worth noting that the comparison of data concerning antibiotic usage between multispecialty paediatric teaching hospitals in Poland might constitute a more reliable reference as regards all analyses.

Analysed antibiotics and chemotherapeutics with DDD assignment demonstrated that after a year following the intervention a reduction of DDD used was noted in all groups of wards included in the study. However, after another year following the intervention, in spite of conducting a PPS intervention at that time (2017), the positive trend of reducing the quantities of DDD used was present only in group 2, which initially had markedly higher values of DDD/100 BD than group 1. As regards AWIS and group 1, not only did the values return to the average values reported before the intervention, but an increased antibiotic usage was noted. Therefore, apart from group 2, the positive effects of interventions were only temporary, with the reservation that the results of analyses were not statistically significant (the Wilcoxon test and linear regression analysis). Annually conducted PPS intervention and a permanent training programme in antibiotic treatment may increase the consistency and scope of the temporary positive trend.

Undoubtedly, group 1 which included wards with a formal duty of additional pre-authorisation of third-line antibiotics was characterised by a markedly lower DDD/100 BD, which may suggest a significant influence of this type of formal restriction on antibiotic usage. Conversely, group 2 wards (ICU, oncohaematology) used higher DDD of antibiotics in numerous hospitals because of the status and complexity of treatment in patients commonly colonized with multidrug-resistant pathogens and with severe infections. Therefore, making a definite conclusion that the formal restrictions concerning antibiotic usage exert a positive effect would require a comparison of wards with a similar specificity. However, the fact that ASP-related interventions in group 2 wards resulted in a downward trend in antibiotic usage lasting 2 years (until the end of 2017) may indicate that despite the special needs of patients treated, there was a certain reserve and possibility of reducing the quantity of DDD/100 BD. It may suggest that the introduction of additional pre-authorisation of at least a part of the third-line antibiotics in group 2 might contribute to a further reduction of DDD/100 BD.

The changes concerning the profile of antibiotics used after ASP-related interventions are positive—that is, the post-interventional shift of the most commonly used antibiotics (amoxicillin, ampicillin and cephazolin) to higher ranks in DU90% and the shift of amikacin to a lower rank. Regrettably, implemented interventions failed to stop the increased use of ceftriaxone and cefuroxime from the group of cephalosporins. It may be correlated, at least partially, with internal recommendations concerning empirical and targeted antibiotic therapy, where those two cephalosporins have a wide range of indications. The analysis of antibiotic groups revealed a statistically significant downward trend as regards the quantity of DDD/100 BD of aminoglycosides and trimethoprim/sulfamethoxazole, which is a positive effect of ASP. However, the upward trend of fluoroquinolones in group 1 is not a positive phenomenon. The situation requires further monitoring.

When analysing the effect of introducing ASP-related interventions it should be remembered that the PPTCH in Warsaw had already established an antibiotic team, and in 2015 the actions were intensified by adding a PPS intervention and external training sessions. Furthermore, obtaining a radical reduction in DDD use at such a facility may be challenging, considering that initially we did not exceed the average of 37.13 DDD in the years 2013–2017, while an NPOA pilot study in Poland showed the average at 43 DDD/100 BD.

Moreover, antibiotic resistance is a rapidly progressive phenomenon, and obtaining the opposite effect, that is, reduced antibiotic resistance of bacteria, may require longer periods. For example, up to 5 years after a single course of clarithromycin, patients continued to be colonised with resistant Staphylococcus epidermidis. Therefore, changes in the hospital-wide antibiogram may not be seen after decreasing antibiotic usage.¹² Obtaining permanent positive effects of ASP at one hospital, especially a teaching hospital, to some extent depends also on therapeutic decisions made at outpatient facilities and primary or secondary referral hospitals, where patients are often treated with multiple antibiotics and get colonised with resistant bacteria before they are admitted to a tertiary referral teaching hospital.

Conclusions

• Lower values of DDD/100 BD characterise the wards with pre-authorisation duty for third-line antibiotics. It may be assumed that additional pre-authorisation might be beneficial (reduced antibiotic usage) in wards without such a duty, in spite of the specificity of patient groups treated, for example, in ICU or Oncology.

• Additional ASP-related activities such as conducting a PPS intervention and training on antibiotic treatment result in diminishing the values of DDD/100 BD. However, the trend is permanent (2 years) only in wards without duty for restricted antibiotics. PPS ECDC alone has no diminishing impact on antimicrobial use in group 1 of the wards. Beneficial changes in the profile of drugs used are reported, but they are accompanied with some negative changes as well. Regularly scheduled training sessions on antibiotic treatment and annual PPS interventions may contribute to obtaining even better ASP-related results.

• More reliable reference to comparisons between centres characterised by a similar profile and capacity may be facilitated by a larger number of ASP-related studies conducted at various centres with detailed descriptions of antibiotics included.

What this paper adds.

What is already known on this subject

• Antibiotic resistance threatens the health and safety of patients in hospitals.

• The consumption of antibiotics in hospitals can be modified by different interventions.

• The Polish national programme to protect antibiotics (NPOA) has been developed to reduce irrational antibiotic usage and occurrence of antimicrobial resistance.

What this study adds

• Hospital antibiotic use can be reduced due to NPOA participation and PPS ECDC.

• Trend in the wards without pre-authorisation duty for restricted antibiotics was more stable compared to wards with such duty.

• PPS ECDC alone has no diminishing impact on antimicrobial use in wards with pre-authorisation duty for third-line (restricted) antibiotics.

Data availability statement

Data are available upon reasonable request.

Ethics statements

Patient consent for publication

Not required.