- Email: [email protected]
Implementation of an antibiotic checklist increased appropriate antibiotic use in the hospital on Aruba
Accepted Manuscript Title: Implementation of an Antibiotic Checklist Increased Appropriate Antibiotic Use in the Hospital on Aruba Authors: Frederike Vera van Daalen, Anouk Lagerburg, Jaclyn de Kort, Elena S`anchez Rivas, Suzanne Eugenie Geerlings PII: DOI: Reference:
S1201-9712(17)30103-0 http://dx.doi.org/doi:10.1016/j.ijid.2017.03.019 IJID 2908
To appear in:
International Journal of Infectious Diseases
Received date: Revised date: Accepted date:
25-1-2017 3-3-2017 16-3-2017
Please cite this article as: van Daalen Frederike Vera, Lagerburg Anouk, de Kort Jaclyn, S`anchez Rivas Elena, Geerlings Suzanne Eugenie.Implementation of an Antibiotic Checklist Increased Appropriate Antibiotic Use in the Hospital on Aruba.International Journal of Infectious Diseases http://dx.doi.org/10.1016/j.ijid.2017.03.019 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Implementation of an Antibiotic Checklist Increased Appropriate Antibiotic Use in the Hospital on Aruba Frederike Vera van Daalen1 [email protected], Anouk Lagerburg2 [email protected], Jaclyn de Kort3 [email protected], Elena Sànchez Rivas4 [email protected], Suzanne Eugenie Geerlings5 [email protected]
Department of Internal Medicine, Division of Infectious Diseases, Academic Medical Centre, University of Amsterdam, Meibergdreef 9, 1105 AZ Amsterdam, the Netherlands. 1
Department of Internal Medicine, Division of Infectious Diseases, Academic Medical Centre, University of Amsterdam, Meibergdreef 9, 1105 AZ Amsterdam, the Netherlands. 2
3
Department of Internal Medicine, dr. Horacio E. Oduber Hospitaal, Dr. Horacio E. Oduber Hospital Boulevard #1, Oranjestad, Aruba. 4
Department of Internal Medicine, dr. Horacio E. Oduber Hospitaal, Dr. Horacio E. Oduber Hospital Boulevard #1, Oranjestad, Aruba. 5
Department of Internal Medicine, Division of Infectious Diseases, Academic Medical Centre, University of Amsterdam, Meibergdreef 9, 1105 AZ Amsterdam, the Netherlands. *
Corresponding author at. Department of Internal Medicine, Division of Infectious Diseases, Academic Medical Centre, University of Amsterdam, Room F4-132, Meibergdreef 9, 1105 AZ Amsterdam, the Netherlands. Tel.: +31-20-5667144; Fax: +31-20-6972286
Highlights
Antimicrobial stewardship strategies were urgently needed in the hospital on Aruba We implemented an antibiotic checklist based on seven validated quality indicators Overall, use of the antibiotic checklist improved antibiotic use Performance did not improve on each individual quality indicator The checklist, originally developed in the Netherlands, is shown to be usable on Aruba
ABSTRACT Objectives No interventions have been done yet to improve antibiotic use on Aruba. In the Netherlands, introduction of an antibiotic checklist resulted in more appropriate antibiotic use in nine
hospitals. We aimed to introduce the antibiotic checklist on Aruba, to test its effectiveness and to evaluate the possibility of implementing this checklist outside the Netherlands. Methods The antibiotic checklist included seven quality indicators (QIs) that define appropriate antibiotic use. It applied to adult patients with a suspected bacterial infection and treated with intravenous antibiotics. Primary endpoint was the QI sum score, calculated by the patient’s sum of performed checklist-items divided by the total number of QIs that applied to that specific patient. Outcomes before and after checklist introduction were compared. Results The percentage of patients with a QI sum score ≥50% was significantly increased in the intervention (n=173) compared to the baseline group (n=150) [OR 3.67, p<0·001]. However, performance did not improve on each individual QI. The checklist was used in 63.3% of the eligible patients. Conclusions Introduction of the antibiotic checklist increased appropriate antibiotic use on Aruba. Further initiatives are necessary for further improvement per QI. Our results suggest that the antibiotic checklist can be used internationally.
Keywords: antibiotic stewardship; checklist; quality indicators; intervention
INTRODUCTION
The increasing antimicrobial resistance (AMR) is a rising threat to global public health, which requires cross-sectional action by governments and society.1,2 The emergence of AMR, together with a steady decline in the development of novel antibiotics, is narrowing down our therapeutic options.3 The consumption of antibiotics is related to the development of AMR
4,5
To curb AMR it is of great importance to preserve the current agents by providing appropriate antibiotic use.
Antibiotic Stewardship Programs (ASP) have been introduced in hospitals worldwide to “measure and improve the appropriate use of antibiotic agents by promoting the selection of the optimal antibiotic drug regimen”.6 There are encouraging examples of showing beneficial effects with regard to decrease in resistance rates and costs, without an increase in clinical failures.7 However, the contents of these programs and the extent to which ASPs are implemented vary between hospitals and countries.8,9 Some European hospitals are worldwide leading in these developments.8 In the Netherlands, for example, the Ministry of Health mandated the commission of a team of infection specialists that drives ASP in each hospital since January 2014. This gives the Netherlands, together with the relative low antimicrobial resistance rates, a unique position concerning AMR compared to many countries in the rest of the world. However, no ASP has been introduced on the island Aruba, which is a constituent country of the Netherlands, located in the southern Caribbean Sea. Since AMR is a worldwide problem, and the healthcare workers in the hospital on Aruba struggle with high antimicrobial resistance rates compared to many other countries – e.g. the local resistance rate of Escherichia coli to trimethoprim/sulfamethoxazole was 56% compared to 23% in the Netherlands in 2014 [vd Berg, R. oral presentation, “DutchCaribbean mini-symposium on antimicrobial stewardship” Oranjestad, November 2015] – ASP strategies are urgently needed and should be started as soon as possible on Aruba. We recently illustrated that the implementation of an antibiotic checklist resulted in more appropriate antibiotic use in the Netherlands.10 This checklist included generic quality indicators (QIs) which define appropriate antibiotic treatment of bacterial infections in the hospital.11,12 Since these QIs have been developed by an international expert panel,11 the checklist should be useful internationally, but this must be measured in a different setting. Aruba is an interesting place to test whether this checklist can lead to similar results abroad, since on one hand similarities between Aruba and the Netherlands exist because of the political relationship, but on the other hand Aruba differs in terms of culture, climate and healthcare system and resembles more to countries in the Caribbean or Latin America.
The aims of this project were to start ASP on Aruba by introduction of the antibiotic checklist, to test its effectiveness and to evaluate the possibility of implementing this checklist outside the Netherlands.
METHODS Study design and setting We conducted a prospective cohort trial between August 1st 2015 and January 15th 2016. The antibiotic checklist was introduced at the 15th November 2015. We distinguished a baseline period (from August 1st until October 1st 2015), transition period (from October 1st until November 15th 2015) and an intervention period (from November 15th 2015 until January 15th 2016). We compared outcomes before (baseline) and after (intervention) checklist introduction. During the transition period implementation activities were started, and no data were collected. The antibiotic checklist was introduced at the Dr. Horacio E. Oduber Hospitaal (HOH), which is the only hospital on Aruba. It has a capacity of 288 beds. Each ward participated in the project, except the Intensive Care Unit (ICU) and the paediatric department, as the QIs included in the checklist do not apply to these patient populations.11 The Medical Research Involving Human Subjects Acts did not apply to this study.10 Since the study involved a quality improvement intervention with negligible risk of harming patients, individual informed consent was waived. The board of directors approved the study protocol.
Participants Eligible patients were hospitalized adults (≥ 18 years old), or adults at the ED who were admitted at a participating ward, with a suspected community-acquired and/or hospitalacquired bacterial infection, treated with intravenous (IV) antibiotics. Eligible patients in the baseline period were identified using a list that was generated by the local pharmacist from the computerized medication ordering system of all patients treated with IV antibiotics between August 1st 2015 and October 1st 2015. We performed two
exclusion rounds. In the first exclusion round we used the information on the pharmacist’s list to exclude all patients younger than 18 years old, with IV antibiotic treatment started at the ICU, and receiving cephazolin as a sole treatment (standard prophylaxis in the hospital). After exclusion round 1 we selected every second patient on the list and collected the case notes. We had to use the case notes for exclusion round 2, in which we excluded patients in case of hospital stay of less than 24 hours, other antibiotics used as prophylaxis, treatment less than 24 hours, or oral started antibiotic treatment. The remaining patients were included in the baseline group. Eligible patients admitted between November 15th 2015 and January 15th 2016 were considered as the intervention group. We included all eligible patients with a partly or total completed checklist, but also a selection of patients without a checklist to investigate whether the implementation strategies had a positive effect on the awareness of appropriate antibiotic use, even though a checklist was not completed. To select this last group of patients, we again used a pharmacist’s list of all patients treated with IV antibiotics. The patients were selected in the same manner as in the baseline measurement, but we selected every third instead of every second patient on the list.
Intervention: the antibiotic checklist The antibiotic checklist was previously developed in a Dutch study.13 It is meant as a supporting tool for physicians to improve the quality of their antibiotic regimens. It includes seven generic QIs that define appropriate antibiotic use in the treatment of bacterial infections in the hospital.11,12 The checklist is divided into two bundles. The first bundle (five items) has to be completed at the moment of prescribing IV antibiotics. The second bundle (two items) has to be used during the course of treatment, at the latest after 72 hours of treatment. Due to the preference of the local infectious diseases (ID) specialists the Dutch checklist was slightly adjusted for the item ‘local guideline adherence’. The option ‘no, according to other guidelines’ was replaced by the option ‘no, in consultation with the ID specialist’. The Aruban antibiotic checklist is presented in figure 1.
Procedure In the baseline period we only collected data and no interventions were done yet. In the transition period we organized a kick-off symposium to introduce the antibiotic checklist among physicians working in the hospital. During this symposium education was given about antimicrobial resistance and the importance of all steps of appropriate antibiotic use as summarized in the checklist-items. In addition, room for improvement of the current antibiotic use in the hospital was shown by presenting data of the baseline period. The intervention period started after the symposium. The checklists were displayed in printed form in doctors’ rooms, and we spread reminding materials such as laminated pocket versions and posters. Physicians working at participating wards were asked to complete checklists for all eligible patients during the intervention period. When the first bundle was completed at the ED, the checklist was taken to the ward with the patient. During the intervention period local study coordinators (AL, JK) gave additional clinical lessons on appropriate antibiotic use and promoted checklist use by face-to-face prompting.
Data collection For both baseline and intervention patients data were collected from patients’ case notes and electronic medical records. We did not use the completed checklists as a data source because the checklist was meant as a supporting tool for physicians, not as a measure instrument for actually care provided. All data were collected by one researcher (AL) who was previously trained in data collection during the Dutch antibiotic checklist study.10 Data were entered in a certificated online data entry system (OpenClinica). Another researcher (FD) checked five percent of all electronic case report forms (of both baseline and intervention patients) against the original data sources.
For all patients we collected data on adherence to the QIs.10 We also collected data on possible confounders, including age, sex, Charlson comorbidity index,14 clinical condition as
assessed by the Modified Early Warning Score,15 type of diagnosis, community- or hospitalacquired infection, and location of start of antibiotic treatment (ED or ward). For outcomes at patient level we collected data on ICU admission and in-hospital mortality.
Outcomes Our primary endpoint was appropriate antibiotic use defined by the QI sum score, calculated by the patient’s sum of performance scores on the QIs divided by the total number of QIs that applied to that specific patient. Secondary endpoints were the performance scores per QI (yes/no per QI), ICU admission and in-hospital mortality. To evaluate the possibility of implementing this checklist on Aruba we estimated the checklist use, which was defined by the percentage of eligible patients for whom a checklist was at least partly completed.
Statistical analysis We distinguished the following four patient groups for the analysis:
Baseline group (1)
Intervention group, consisting of: o
Eligible patients with a completed checklist; (2)
o
Eligible patients with a partly completed checklist; (3)
o
Eligible patients without a checklist. (4)
Differences in patient characteristics between baseline (1) and intervention (2+3+4) patients were assessed with the chi-square test for categorical, and with the unpaired t-test for numerical variables. QI performance was calculated for each patient by algorithms that were developed in a previous study.12 The algorithm of the QI “prescribe systemic antibiotic treatment according to the local antibiotic guideline” was adapted to the local antibiotic guidelines of the Dr. HOH. These algorithms took into account whether a checklist item applied to the patient. E.g., the item “adapt therapy when culture results become available” only applied to patients with a positive culture result. The QI sum score was also calculated by an algorithm, dividing the
number of performed QIs by the total number of QIs that applied to that specific patient. Per patient we determined whether the QI sum score was 50% or higher.16 To compare the percentages of patients with a QI sum score ≥ 50% in the baseline and in the intervention group we performed logistic regression analyses. We used the same covariates for adjustment that were included in the Dutch effectiveness study.10 To identify differences between the influence of the intervention period in general (independently whether a checklist was completed or not), and the influence of the implementation strategies plus the fulfilment of a checklist, we performed two analyses. First we compared the QI sum score of the baseline group (1) to the total intervention group (2+3+4), and then we compared the baseline group (1) to the patients with a completed checklist in the intervention group (2). For secondary outcomes we compared the baseline group (1) to the total intervention group (2+3+4). The scores per QI (yes/no per QI) were also assessed using logistic regression with adjustment for covariates. ICU admission and in-hospital mortality were expected to have a low incidence and therefore were assessed descriptively. Checklist use was estimated by extrapolating the percentage of eligible patients in the intervention group without a checklist after the second exclusion round, to all patients without a checklist in the intervention group. The percentage of checklist use was calculated by dividing the number of included checklists by the number of included checklists plus the number of eligible patients without a checklist according to the extrapolation.
Sample size No data were available on the QI performance in the Dr. HOH, and therefore we used literature to estimate the expected increase in QI sum score. A study aiming to improve antibiotic use for patients with a complicated urinary tract infection resulted in an increase in QI sum score of approximately 4%.17 The implementation of the antibiotic checklist in the Netherlands resulted in an increase of the QI sum score of 18.7%.10 In a simulation study based on the two group chi-square test of equal proportions we calculated that a sample size
of 300 patients detects a difference of 15% in QI sum score assuming a type I error of 0·05, and a type II error of 0·2. For the power analysis we used SAS version 9.3. All other analyses were done using IBM SPSS Statistics, version 23.0.
RESULTS Between August 1st and October 1st 2015, 150 patients were included in the baseline group. From November 15th 2015 until January 15th 2016 173 patients were included in the intervention group, of whom 110 had a completed checklist, 35 had a partly completed checklist, and 28 had no checklist (figure 2). Of the 35 incomplete checklists, all checklists had a completed first bundle, but an incomplete second bundle. Most checklists were collected from the internal (39.3%), surgical (28.3%) and lung department (9.0%). The characteristics of the included patient groups are summarized in table 1. The only significant difference between the baseline and intervention group related to the type of diagnosis (p = 0.04). In the baseline group the percentage of patients with ‘more than two possible diagnoses’ or ‘diagnosis not covered by the local guidelines’ was higher compared to the intervention group. These patients were excluded from the analysis of the QI ‘prescribe antibiotic treatment according to the local antibiotic guideline’ as it was not clear which local guideline was supposed to be followed.
In the baseline group (1) the mean QI sum score was 45% compared to 61% in the total intervention group (2+3+4) and 64% in the completed checklist group (2). The percentage of patients with a QI sum score ≥ 50% was significantly increased in the whole intervention group (n=173) compared to the baseline group (n=150) [odds ratio (OR) 3.67 (95% CI 2·2– 6·0), p<0·001]. In the comparison between the baseline- and the completed checklist group (n=110) the OR was even greater [4.74 (CI 95% 2.6–8.7), p<0.001]. Per QI, the association between the intervention and performance according to the QI differed (table 2). The largest increase was seen in the collection of blood cultures, followed
by the adjustment to renal function and documentation of the indication for antibiotic treatment. Performances on de-escalation and IV-oral switch remained similar in the intervention period, and a slight decline was indicated in the performance on guideline adherence and collection of cultures from the suspected site of infection in the intervention period. ICU admission and in-hospital mortality were sparse in both study groups. In the baseline group eleven patients were admitted to the ICU (7.3%) and two patients died (1.3%), compared to fifteen ICU admissions (8.7%) and four deaths (2.3%) in the intervention period. In total 145 eligible patients had a (partly) completed checklist. After the first exclusion round, the intervention group without a checklist contained 180 patients (figure 2). The percentage of eligible patients without a checklist after the second exclusion round in the intervention group was 46.7% (28 out of 60, see figure 2). Hence, approximately 84 patients in the intervention group without a checklist were actually eligible for the checklist (0.467 * 180 = 84). The estimated total study compliance was 63.3% (145/(145+84)).
DISCUSSION Our study demonstrated that the implementation of an antibiotic checklist on Aruba was associated with a significant increase in QI sum score, but performance did not improve on each individual QI. The checklist was used in 63.3% of the eligible patients, which illustrates that it was possible to implement the antibiotic checklist outside the Netherlands. Performance on taken blood cultures showed the largest improvement in the intervention period, which also applied to the Dutch checklist implementation study.10 Documentation of the indication was also improved, even though the baseline score was already high. This improvement explains the differences in the type of diagnoses between the baseline and the intervention group, as in the baseline group more patients with unclear documentation of indication ended up in the category of ‘more than two possible diagnosis’, or ‘diagnosis not covered by guidelines’.
The room for improvement in guideline adherence in the baseline period is in line with previous results from different parts of the world.12,18-22 For example, an observational study in a teaching hospital in Brazil concluded that 65.4% of the parenteral antimicrobial prescriptions was appropriate according to the local guidelines.18 It is disappointing, however, that guideline adherence did not improve in our intervention period, which is in contrast with other intervention studies, including our Dutch checklist study, aiming to improve antibiotic prescription.19-22 The main difference with these studies is that in our project improvement of guideline adherence was actually one of seven aims, while they only focussed on the antibiotic prescription.19-22 Obviously the Dutch checklist study also aimed to improve performance on the seven checklist-items, but an important difference is the lack of an easy accessible digital version of the guidelines on Aruba, which possibly hinders its use. Insufficient use of the guidelines can also be an explanation for the lack of improvement in de-escalation and IV-oral switch, as instructions for these items are given in the same antibiotic guideline booklet. In general, improvement on each QI might be hindered by these kind of barriers, as, in contrast to the Dutch study,13 we did not perform a barrier identification prior to checklist implementation on Aruba.
The checklist use of 63.3% was surprisingly high compared to 23.2% in the Dutch study.10 On Aruba some extra activities were undertaken, such as clinical lectures in the intervention period,
and
face-to-face-prompting,
which
has
been
described
as
an
effective
implementation intervention.6,23 Due to this approach the effect of checklist use and the other educational efforts intervened. In contrary to the Dutch checklist study,10 we have collected data on QI performance of twenty-eight intervention patients without a completed checklist. We have shown that the increase in QI sum score was stronger in the intervention group with a completed checklist, suggesting that checklist use has an additional effect. Few studies have been performed in the Caribbean and Latin America on the introduction of an ASP.22,24 This study was the first project performing an intervention that resulted in
improvement in the process of intravenous antibiotic use on Aruba. We met several criteria that are required for successful checklist implementation, including the implementation of a checklist that is based on previous research,25 and the organization of stimulating activities for checklist use.25-27 We illustrated the additional effect of completing a checklist to the effects caused by the other implementation activities by also collecting data of patients who were eligible but for whom a checklist was not used. More general, we provided data on antibiotic use in the Dr. HOH hospital, which was absent prior to this project, and we gained awareness on appropriate antibiotic use and the problem of antimicrobial resistance among healthcare workers on Aruba.
Our study has several limitations. Due to time limits we did not collect data of all eligible patients during the baseline- and intervention period. However, we have minimalized the chance of measurement bias by a systematic approach of inclusion. A disadvantage of these small numbers is that the sample size was too small to reach significance in some sub analyses. For example, the fact that the improved performance in adjustment to renal function was not significant, can be caused by the relatively small number of patients this QI applied to. Finally, our primary endpoint is a combination of seven QIs, but the increase in QI sum score was mainly caused by the high increase in collection of blood cultures. Therefore it can be questioned if the QI sum score is the best measure to reflect the improvement of the whole process of antibiotic use.
Pros and cons of further implementation of the checklist should be considered. Implementing a checklist is a complex intervention that requires a lot of motivation and effort of physicians. On the other hand, it provides an opportunity to involve physicians with direct responsibility for patients in the aims of antimicrobial stewardship.28 In contrast with most ASP interventions,7 the presence of infectious diseases experts is not required with checklist use. Another advantage is that the tool is based on validated quality indicators.11,12 An important
con is that the impact appeared only in some QIs in this study. Therefore, before further implementation is started, barriers should be identified that hinder improvement of all QIs in daily practice in the Dr. HOH. For example, the implementation of the local antibiotic booklet seems the first important step to improvement. By addressing these barriers, the process of antibiotic use could be further improved.
In conclusion, awareness created by the ASP activities including the implementation of the antibiotic checklist resulted in an increase in sum score of QIs for appropriate antibiotic use on Aruba. Further initiatives are necessary for further improvement, especially to improve antibiotic guideline adherence. Completion of the antibiotic checklist was high, suggesting that the implementation of the checklist is possible outside the Netherlands.
Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Competing interests No competing interests declared.
Acknowledgements We are grateful to dr. B.C. Opmeer from the Clinical Research Unit of the Academic Medical Centre for his assistance in the sample size calculation. We also want to give special thanks to Ebby Ruiz from the pharmacist department in the dr. E. Oduber hospital, who provided the lists of all patients treated with intravenous antibiotics. References
[1] World Health Organization (WHO). Antimicrobial resistance: global report on surveillance 2014. .
[2]
http://www.who.int/drugresistance/documents/surveillancereport/en/.
Accessed
October 2016.
[3] The Review on Antimicrobial Resistance. Tackling drug-resistant infections globally:
Final
report
and
recommendations.
2016.
http://amr-
review.org/sites/default/files/160525_Final%20paper_with%20cover.pdf
Accessed
September 2016.
[4] Luepke KH, Suda KJ, Boucher H, et al.,Past, present, and future of antibacterial economics: Increasing bacterial resistance, limited antibiotic pipeline, and social implications. Pharmacotherapy 2017;37:71-84.
[5] Tacconelli E.Antimicrobial use: risk driver of multidrug resistant microorganisms in healthcare settings. Curr Opin Infect Dis 2009;22:352–8.
[6] Goossens H, Ferech M, Vander Stichele R, Elseviers M,ESAC Project Group. Outpatient antibiotic use in Europe and association with resistance: a cross-national database study. Lancet 2005;365:579–87.
[7] Barlam TF, Cosgrove SE, Abbo LM, et al.,Implementing an Antibiotic Stewardship Program: guidelines by the Infectious Diseases Society of America and the Society for Healthcare Epidemiology of America. Clin Infect Dis 2016;62:e51–77.
[8] Davey P, Marwick CA, Scott CL, et al.,Interventions to improve antibiotic prescribing
practices
2017;2:CD003543.
for
hospital
inpatients.
Cochrane
Database
Syst
Rev
[9] Howard P, Pulcini C, Hara GL, West RM, Gould IM, Harbarth S, et al.,An international cross–sectional survey of antimicrobial stewardship programmes in hospitals. J Antimicrob Chemother 2015;70:1245-55.
[10] World Health Organization. The evolving threat of antimicrobial resistance. Options
for
action.
http://whqlibdoc.who.int/publications/2012/9789241503181_eng.pdf.
2012. Accessed
July
2016.
[11] Van Daalen FV, Prins JM, Opmeer BC, Boermeester MA, Visser CE, van Hest RM et al The effect of an antibiotic checklist on length of hospital stay and appropriate antibiotic use in adult patients treated with intravenous antibiotics: a stepped wedge cluster randomised trial. Clinical Microbiol Infect 2017; DOI: 10.1016/j.cmi.2017.01.019 [Epub ahead of print].
[12] Van den Bosch CM, Geerlings SE, Natsch S, Wille J, Prins JM, Hulscher ME.Quality indicators to measure appropriate antibiotic use in hospitalized adults. Clin Infect Dis 2015;60:281–91.
[13] van den Bosch CM, Hulscher ME, Natsch S, Wille J, Prins JM, Geerlings SE.Applicability of generic quality Indicators for appropriate antibiotic use in daily hospital practice: a cross-sectional point-prevalence multicenter study. Clinical Microbiol Infect 2016;22:888.e1-e9.
[14] Van Daalen FV, Geerlings SE, Prins JM, Hulscher ME.A survey to identify barriers of implementing an antibiotic checklist. Eur J Clin Microbiol Infect Dis 2016;35:545.
[15] Quan H, Sundararajan V, Halfon P, et al.,Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care 2005;43:1130–9.
[16] Subbe CP, Kruger M, Rutherford P, Gemmel L.Validation of a modified early warning score in medical admissions. QJM 2001;94:521–6.
[17] van den Bosch CM, Hulscher MEJL, Akkermans RP, Wille J, Geerlings SE, Prins JM.Appropriate antibiotic use reduces length of hospital stay. J Antimicrob Chemother 2016; DOI: 10.1093/jac/dkw469 [Epub ahead of print].
[18] Spoorenberg V, Hulscher ME, Geskus RB, de Reijke TM, Opmeer BC, Prins JM, Geerlings SE.A Cluster-Randomized Trial of Two Strategies to Improve Antibiotic Use for Patients with a Complicated Urinary Tract Infection. PLoS One 2015;10:e0142672. .
[19] Kawanami GH, Fortaleza CM.Factors predictive of inappropriateness in requests for parenteral antimicrobials for therapeutic purposes: a study in a small teaching hospital in Brazil. Scand J Infect Dis 2011;43:528-35.
[20] Hecker MT, Fox CJ, Cydulka RK, Siff JE, Emerman CL, Sethi AK, Muganda CP, Donskey CJ.Effect of a stewardship intervention on adherence to uncomplicated cystitis and pyelonephritis guidelines in an emergency department setting. PLoS One 2014;9:e87899. .
[21] Percival KM, Valenti KM, Schmittling SE, Strader BD, Lopez RR, Bergman SJ.Impact of an antimicrobial stewardship intervention on urinary tract infection treatment in the ED. Am J Emerg Med 2015;33:1129-33. .
[22] Popovski Z, Mercuri M, Main C, Sne N, Walsh K, Sung M, Rice T, Mertz D.Multifaceted intervention to optimize antibiotic use for intra-abdominal infections. J Antimicrob Chemother 2015;70:1226-9. .
[23] Guanche Garcell H, Pisonero Socias JJ, Enseñat Sánchez R, Fiterre Lancis I, Mir Narbona I, García Arzola B, Pardo Gómez G, Gutiérrez García F.Impact of a quality control program on antibiotic prescription in a hospital in Havana, Cuba. Rev Panam Salud Publica 2011;30:598-602.
[24] Weiss CH, Dibardino D, Rho J, Sung N, Collander B, Wunderink RG.A clinical trial comparing physician prompting with an unprompted automated electronic checklist to reduce empirical antibiotic utilization. Crit Care Med 2013;41:2563-9.
[25] Okumura LM, Silva MM, Veroneze I. Effects of a bundled Antimicrobial Stewardship Program on mortality: a cohort study. Braz J Infect Dis 2015;19:246-52.
[26] Bosk CL, Dixon-Woods M, Goeschel CA, Pronovost PJ.Reality check for checklists. Lancet 2009;374:444–5.
[27] Leape LL. The checklist conundrum. N Engl J Med 2014;370:1063–4.
[28] Pronovost P, Needham D, Berenholtz S, et al.,An intervention to decrease catheter-related bloodstream infections in the ICU. N Engl J Med 2006;355:2725–32.
[29] Hamilton KW, Gerber SJ, Moehring R, et al.,Point-of-prescription interventions to Improve Antimicrobial stewardship. Clin Infect Dis 2015;60:1252–8.
Figure 1. The Aruban Antibiotic Checklist
Figure 2. Flow diagram of participants
Table 1: patient characteristics
Characteristics
Sex. male Age, mean (SD) Infection, community acquired versus hospital acquired
Baseline N = 150
Intervention N = 173
Baseline (%)
Completed Checklists (%)
N=150 74 (49.3) 61.7 (20.1) 120 (80.0) / 30 (20.0)
N = 110 56 (50.9) 63.5 (18.7) 87 (79.1) / 23 (20.9)
Partly completed checklist (%) N = 35 16 (45.7) 58.1 (18.4) 32 (91.4) / 3 (8.6)
No checklist (%)
Intervention total (%)
N = 28 9 (32.1) 63.9 (17.3) 23 (82.1) / 5 (17.9)
N = 173 81 (46.8) 62.5 (18.5) 142 (82.1) / 31 (17.9)
P
0.65 0.22 0.63
Type of diagnosis Respiratory tract infection Urinary tract infection Skin and soft tissue infection Intra-abdominal infection Other infections Two diagnoses More than 2 possible diagnoses/ diagnosis not covered by guideline Participating wards Internal medicine Surgery Pulmonology Gastroenterology Urology Other Charlson comorbidity index, median (IQR) MEWS score 0 1 2 3 4 5 6 ≥7 AB started at the ED, yes
0.04 23 (15.3) 28 (18.7) 14 (9.3)
26 (23.6) 22 (20.0) 20 (18.2)
6 (17.1) 3 (8.6) 9 (25.7)
7 (25.0) 3 (10.7) 4 (14.3)
39 (22.5) 28 (16.2) 33 (19.1)
15 (10.0) 29 (19.4) 14 (9.3) 27 (18.0)
9 (8.2) 11 (10.0) 10 (9.1) 12 (10.9)
3 (8.6) 7 (20.0) 4 (11.4) 3 (8.6)
2 (7.1) 7 (25.0) 3 (10.7) 2 (7.1)
14 (8.1) 25 (14.5) 17 (9.8) 17 (9.8)
40 (26.7) 41 (27.3) 13 (8.6) 13 (8.6) 10 (6.7) 33 (22.0) 0 (0-2)
47 (42.7) 25 (22.7) 11 (10.0) 5 (4.5) 2 (1.8) 20 (18.2) 1 (0-2)
10 (28.6) 16 (45.8) 2 (5.7) 0 (0) 3 (8.6) 4 (11.4) 0 (0-1)
5 (17.9) 7 (25.0) 5 (17.9) 4 (14.3) 3 (10.7) 4 (14.3) 0.5 (0-2)
62 (35.8) 48 (27.7) 18 (10.4) 9 (5.2) 8 (4.6) 28 (16.2) 0 (0-2)
31 (20.7) 54 (36.0) 29 (19.3) 23 (15.3) 8 (5.4) 2 (1.3) 1 (0.7) 2 (1.3) 34 (22.7)
22 (20.0) 34 (30.9) 19 (17.3) 22 (20.0) 9 (8.2) 0 (0) 3 (2.7) 1 (0.9) 34 (30.9)
11 (31.4) 14 (40.0) 3 (8.6) 4 (11.4) 2 (5.7) 0 (0) 1 (2.9) 0 (0) 11 (31.4)
6 (21.4) 11 (39.3) 4 (14.3) 5 (17.9) 2 (7.1) 0 (0) 0 (0) 0 (0) 8 (28.6)
39 (22.5) 59 (34.1) 26 (15.0) 31 (17.9) 13 (7.5) 0 (0) 4 (2.3) 1 (0.6) 53 (30.6)
0.39
0.60 0.35
0.11
Table 2. QI sum score and score per QI QI
Baseline (n = 150)
Intervention (n =173)
Comparison
numerator/ denominator (%)
numerator/ denominator (%)
Difference %
OR*
CI (95%)
P
Blood cultures
22/150 (15)
119/173 (69)
+ 54
13.7a
7.7-24.2
<0.001
Cultures of suspected site of infection Guideline adherence
33/68 (49)
41/99 (41)
-8
0.69b
0.4-1.3
0.27
67/109 (61)
79/138 (57)
-4
0.84c
0.5-1.4
0.51
Adjustment to renal function
6/16 (38)
12/23 (52)
+ 14
1.84d
0.5-6.8
0.36
Documented antibiotic plan
116/150 (77)
154/173 (89)
+ 12
1.90e
1.0-3.7
0.06
De-escalation
15/53 (28)
21/77 (27)
-1
1.07f
0.5-2.2
0.85
IV-oral switch
29/85 (34)
41/98 (42)
+8
1.17g
0.6-2.2
0.64
SUMSCORE ≥ 50%
75/150 (50)
134/173 (77)
+ 27
3.67h
2.2-6.0
<0.001
*based on logistic regression analyses, taking covariates into account. Covariates: a type of diagnosis, sex, location start antibiotics, MEWS b type of diagnosis c community vs hospital acquired infection, type of diagnosis, AB last 30 days d sex e community vs hospital acquired infection, type of diagnosis f community vs hospital acquired infection g age, community vs hospital acquired infection, type of diagnosis, location start antibiotics h community vs hospital acquired and location start antibiotics
S1201-9712(17)30103-0 http://dx.doi.org/doi:10.1016/j.ijid.2017.03.019 IJID 2908
To appear in:
International Journal of Infectious Diseases
Received date: Revised date: Accepted date:
25-1-2017 3-3-2017 16-3-2017
Please cite this article as: van Daalen Frederike Vera, Lagerburg Anouk, de Kort Jaclyn, S`anchez Rivas Elena, Geerlings Suzanne Eugenie.Implementation of an Antibiotic Checklist Increased Appropriate Antibiotic Use in the Hospital on Aruba.International Journal of Infectious Diseases http://dx.doi.org/10.1016/j.ijid.2017.03.019 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Implementation of an Antibiotic Checklist Increased Appropriate Antibiotic Use in the Hospital on Aruba Frederike Vera van Daalen1 [email protected], Anouk Lagerburg2 [email protected], Jaclyn de Kort3 [email protected], Elena Sànchez Rivas4 [email protected], Suzanne Eugenie Geerlings5 [email protected]
Department of Internal Medicine, Division of Infectious Diseases, Academic Medical Centre, University of Amsterdam, Meibergdreef 9, 1105 AZ Amsterdam, the Netherlands. 1
Department of Internal Medicine, Division of Infectious Diseases, Academic Medical Centre, University of Amsterdam, Meibergdreef 9, 1105 AZ Amsterdam, the Netherlands. 2
3
Department of Internal Medicine, dr. Horacio E. Oduber Hospitaal, Dr. Horacio E. Oduber Hospital Boulevard #1, Oranjestad, Aruba. 4
Department of Internal Medicine, dr. Horacio E. Oduber Hospitaal, Dr. Horacio E. Oduber Hospital Boulevard #1, Oranjestad, Aruba. 5
Department of Internal Medicine, Division of Infectious Diseases, Academic Medical Centre, University of Amsterdam, Meibergdreef 9, 1105 AZ Amsterdam, the Netherlands. *
Corresponding author at. Department of Internal Medicine, Division of Infectious Diseases, Academic Medical Centre, University of Amsterdam, Room F4-132, Meibergdreef 9, 1105 AZ Amsterdam, the Netherlands. Tel.: +31-20-5667144; Fax: +31-20-6972286
Highlights
Antimicrobial stewardship strategies were urgently needed in the hospital on Aruba We implemented an antibiotic checklist based on seven validated quality indicators Overall, use of the antibiotic checklist improved antibiotic use Performance did not improve on each individual quality indicator The checklist, originally developed in the Netherlands, is shown to be usable on Aruba
ABSTRACT Objectives No interventions have been done yet to improve antibiotic use on Aruba. In the Netherlands, introduction of an antibiotic checklist resulted in more appropriate antibiotic use in nine
hospitals. We aimed to introduce the antibiotic checklist on Aruba, to test its effectiveness and to evaluate the possibility of implementing this checklist outside the Netherlands. Methods The antibiotic checklist included seven quality indicators (QIs) that define appropriate antibiotic use. It applied to adult patients with a suspected bacterial infection and treated with intravenous antibiotics. Primary endpoint was the QI sum score, calculated by the patient’s sum of performed checklist-items divided by the total number of QIs that applied to that specific patient. Outcomes before and after checklist introduction were compared. Results The percentage of patients with a QI sum score ≥50% was significantly increased in the intervention (n=173) compared to the baseline group (n=150) [OR 3.67, p<0·001]. However, performance did not improve on each individual QI. The checklist was used in 63.3% of the eligible patients. Conclusions Introduction of the antibiotic checklist increased appropriate antibiotic use on Aruba. Further initiatives are necessary for further improvement per QI. Our results suggest that the antibiotic checklist can be used internationally.
Keywords: antibiotic stewardship; checklist; quality indicators; intervention
INTRODUCTION
The increasing antimicrobial resistance (AMR) is a rising threat to global public health, which requires cross-sectional action by governments and society.1,2 The emergence of AMR, together with a steady decline in the development of novel antibiotics, is narrowing down our therapeutic options.3 The consumption of antibiotics is related to the development of AMR
4,5
To curb AMR it is of great importance to preserve the current agents by providing appropriate antibiotic use.
Antibiotic Stewardship Programs (ASP) have been introduced in hospitals worldwide to “measure and improve the appropriate use of antibiotic agents by promoting the selection of the optimal antibiotic drug regimen”.6 There are encouraging examples of showing beneficial effects with regard to decrease in resistance rates and costs, without an increase in clinical failures.7 However, the contents of these programs and the extent to which ASPs are implemented vary between hospitals and countries.8,9 Some European hospitals are worldwide leading in these developments.8 In the Netherlands, for example, the Ministry of Health mandated the commission of a team of infection specialists that drives ASP in each hospital since January 2014. This gives the Netherlands, together with the relative low antimicrobial resistance rates, a unique position concerning AMR compared to many countries in the rest of the world. However, no ASP has been introduced on the island Aruba, which is a constituent country of the Netherlands, located in the southern Caribbean Sea. Since AMR is a worldwide problem, and the healthcare workers in the hospital on Aruba struggle with high antimicrobial resistance rates compared to many other countries – e.g. the local resistance rate of Escherichia coli to trimethoprim/sulfamethoxazole was 56% compared to 23% in the Netherlands in 2014 [vd Berg, R. oral presentation, “DutchCaribbean mini-symposium on antimicrobial stewardship” Oranjestad, November 2015] – ASP strategies are urgently needed and should be started as soon as possible on Aruba. We recently illustrated that the implementation of an antibiotic checklist resulted in more appropriate antibiotic use in the Netherlands.10 This checklist included generic quality indicators (QIs) which define appropriate antibiotic treatment of bacterial infections in the hospital.11,12 Since these QIs have been developed by an international expert panel,11 the checklist should be useful internationally, but this must be measured in a different setting. Aruba is an interesting place to test whether this checklist can lead to similar results abroad, since on one hand similarities between Aruba and the Netherlands exist because of the political relationship, but on the other hand Aruba differs in terms of culture, climate and healthcare system and resembles more to countries in the Caribbean or Latin America.
The aims of this project were to start ASP on Aruba by introduction of the antibiotic checklist, to test its effectiveness and to evaluate the possibility of implementing this checklist outside the Netherlands.
METHODS Study design and setting We conducted a prospective cohort trial between August 1st 2015 and January 15th 2016. The antibiotic checklist was introduced at the 15th November 2015. We distinguished a baseline period (from August 1st until October 1st 2015), transition period (from October 1st until November 15th 2015) and an intervention period (from November 15th 2015 until January 15th 2016). We compared outcomes before (baseline) and after (intervention) checklist introduction. During the transition period implementation activities were started, and no data were collected. The antibiotic checklist was introduced at the Dr. Horacio E. Oduber Hospitaal (HOH), which is the only hospital on Aruba. It has a capacity of 288 beds. Each ward participated in the project, except the Intensive Care Unit (ICU) and the paediatric department, as the QIs included in the checklist do not apply to these patient populations.11 The Medical Research Involving Human Subjects Acts did not apply to this study.10 Since the study involved a quality improvement intervention with negligible risk of harming patients, individual informed consent was waived. The board of directors approved the study protocol.
Participants Eligible patients were hospitalized adults (≥ 18 years old), or adults at the ED who were admitted at a participating ward, with a suspected community-acquired and/or hospitalacquired bacterial infection, treated with intravenous (IV) antibiotics. Eligible patients in the baseline period were identified using a list that was generated by the local pharmacist from the computerized medication ordering system of all patients treated with IV antibiotics between August 1st 2015 and October 1st 2015. We performed two
exclusion rounds. In the first exclusion round we used the information on the pharmacist’s list to exclude all patients younger than 18 years old, with IV antibiotic treatment started at the ICU, and receiving cephazolin as a sole treatment (standard prophylaxis in the hospital). After exclusion round 1 we selected every second patient on the list and collected the case notes. We had to use the case notes for exclusion round 2, in which we excluded patients in case of hospital stay of less than 24 hours, other antibiotics used as prophylaxis, treatment less than 24 hours, or oral started antibiotic treatment. The remaining patients were included in the baseline group. Eligible patients admitted between November 15th 2015 and January 15th 2016 were considered as the intervention group. We included all eligible patients with a partly or total completed checklist, but also a selection of patients without a checklist to investigate whether the implementation strategies had a positive effect on the awareness of appropriate antibiotic use, even though a checklist was not completed. To select this last group of patients, we again used a pharmacist’s list of all patients treated with IV antibiotics. The patients were selected in the same manner as in the baseline measurement, but we selected every third instead of every second patient on the list.
Intervention: the antibiotic checklist The antibiotic checklist was previously developed in a Dutch study.13 It is meant as a supporting tool for physicians to improve the quality of their antibiotic regimens. It includes seven generic QIs that define appropriate antibiotic use in the treatment of bacterial infections in the hospital.11,12 The checklist is divided into two bundles. The first bundle (five items) has to be completed at the moment of prescribing IV antibiotics. The second bundle (two items) has to be used during the course of treatment, at the latest after 72 hours of treatment. Due to the preference of the local infectious diseases (ID) specialists the Dutch checklist was slightly adjusted for the item ‘local guideline adherence’. The option ‘no, according to other guidelines’ was replaced by the option ‘no, in consultation with the ID specialist’. The Aruban antibiotic checklist is presented in figure 1.
Procedure In the baseline period we only collected data and no interventions were done yet. In the transition period we organized a kick-off symposium to introduce the antibiotic checklist among physicians working in the hospital. During this symposium education was given about antimicrobial resistance and the importance of all steps of appropriate antibiotic use as summarized in the checklist-items. In addition, room for improvement of the current antibiotic use in the hospital was shown by presenting data of the baseline period. The intervention period started after the symposium. The checklists were displayed in printed form in doctors’ rooms, and we spread reminding materials such as laminated pocket versions and posters. Physicians working at participating wards were asked to complete checklists for all eligible patients during the intervention period. When the first bundle was completed at the ED, the checklist was taken to the ward with the patient. During the intervention period local study coordinators (AL, JK) gave additional clinical lessons on appropriate antibiotic use and promoted checklist use by face-to-face prompting.
Data collection For both baseline and intervention patients data were collected from patients’ case notes and electronic medical records. We did not use the completed checklists as a data source because the checklist was meant as a supporting tool for physicians, not as a measure instrument for actually care provided. All data were collected by one researcher (AL) who was previously trained in data collection during the Dutch antibiotic checklist study.10 Data were entered in a certificated online data entry system (OpenClinica). Another researcher (FD) checked five percent of all electronic case report forms (of both baseline and intervention patients) against the original data sources.
For all patients we collected data on adherence to the QIs.10 We also collected data on possible confounders, including age, sex, Charlson comorbidity index,14 clinical condition as
assessed by the Modified Early Warning Score,15 type of diagnosis, community- or hospitalacquired infection, and location of start of antibiotic treatment (ED or ward). For outcomes at patient level we collected data on ICU admission and in-hospital mortality.
Outcomes Our primary endpoint was appropriate antibiotic use defined by the QI sum score, calculated by the patient’s sum of performance scores on the QIs divided by the total number of QIs that applied to that specific patient. Secondary endpoints were the performance scores per QI (yes/no per QI), ICU admission and in-hospital mortality. To evaluate the possibility of implementing this checklist on Aruba we estimated the checklist use, which was defined by the percentage of eligible patients for whom a checklist was at least partly completed.
Statistical analysis We distinguished the following four patient groups for the analysis:
Baseline group (1)
Intervention group, consisting of: o
Eligible patients with a completed checklist; (2)
o
Eligible patients with a partly completed checklist; (3)
o
Eligible patients without a checklist. (4)
Differences in patient characteristics between baseline (1) and intervention (2+3+4) patients were assessed with the chi-square test for categorical, and with the unpaired t-test for numerical variables. QI performance was calculated for each patient by algorithms that were developed in a previous study.12 The algorithm of the QI “prescribe systemic antibiotic treatment according to the local antibiotic guideline” was adapted to the local antibiotic guidelines of the Dr. HOH. These algorithms took into account whether a checklist item applied to the patient. E.g., the item “adapt therapy when culture results become available” only applied to patients with a positive culture result. The QI sum score was also calculated by an algorithm, dividing the
number of performed QIs by the total number of QIs that applied to that specific patient. Per patient we determined whether the QI sum score was 50% or higher.16 To compare the percentages of patients with a QI sum score ≥ 50% in the baseline and in the intervention group we performed logistic regression analyses. We used the same covariates for adjustment that were included in the Dutch effectiveness study.10 To identify differences between the influence of the intervention period in general (independently whether a checklist was completed or not), and the influence of the implementation strategies plus the fulfilment of a checklist, we performed two analyses. First we compared the QI sum score of the baseline group (1) to the total intervention group (2+3+4), and then we compared the baseline group (1) to the patients with a completed checklist in the intervention group (2). For secondary outcomes we compared the baseline group (1) to the total intervention group (2+3+4). The scores per QI (yes/no per QI) were also assessed using logistic regression with adjustment for covariates. ICU admission and in-hospital mortality were expected to have a low incidence and therefore were assessed descriptively. Checklist use was estimated by extrapolating the percentage of eligible patients in the intervention group without a checklist after the second exclusion round, to all patients without a checklist in the intervention group. The percentage of checklist use was calculated by dividing the number of included checklists by the number of included checklists plus the number of eligible patients without a checklist according to the extrapolation.
Sample size No data were available on the QI performance in the Dr. HOH, and therefore we used literature to estimate the expected increase in QI sum score. A study aiming to improve antibiotic use for patients with a complicated urinary tract infection resulted in an increase in QI sum score of approximately 4%.17 The implementation of the antibiotic checklist in the Netherlands resulted in an increase of the QI sum score of 18.7%.10 In a simulation study based on the two group chi-square test of equal proportions we calculated that a sample size
of 300 patients detects a difference of 15% in QI sum score assuming a type I error of 0·05, and a type II error of 0·2. For the power analysis we used SAS version 9.3. All other analyses were done using IBM SPSS Statistics, version 23.0.
RESULTS Between August 1st and October 1st 2015, 150 patients were included in the baseline group. From November 15th 2015 until January 15th 2016 173 patients were included in the intervention group, of whom 110 had a completed checklist, 35 had a partly completed checklist, and 28 had no checklist (figure 2). Of the 35 incomplete checklists, all checklists had a completed first bundle, but an incomplete second bundle. Most checklists were collected from the internal (39.3%), surgical (28.3%) and lung department (9.0%). The characteristics of the included patient groups are summarized in table 1. The only significant difference between the baseline and intervention group related to the type of diagnosis (p = 0.04). In the baseline group the percentage of patients with ‘more than two possible diagnoses’ or ‘diagnosis not covered by the local guidelines’ was higher compared to the intervention group. These patients were excluded from the analysis of the QI ‘prescribe antibiotic treatment according to the local antibiotic guideline’ as it was not clear which local guideline was supposed to be followed.
In the baseline group (1) the mean QI sum score was 45% compared to 61% in the total intervention group (2+3+4) and 64% in the completed checklist group (2). The percentage of patients with a QI sum score ≥ 50% was significantly increased in the whole intervention group (n=173) compared to the baseline group (n=150) [odds ratio (OR) 3.67 (95% CI 2·2– 6·0), p<0·001]. In the comparison between the baseline- and the completed checklist group (n=110) the OR was even greater [4.74 (CI 95% 2.6–8.7), p<0.001]. Per QI, the association between the intervention and performance according to the QI differed (table 2). The largest increase was seen in the collection of blood cultures, followed
by the adjustment to renal function and documentation of the indication for antibiotic treatment. Performances on de-escalation and IV-oral switch remained similar in the intervention period, and a slight decline was indicated in the performance on guideline adherence and collection of cultures from the suspected site of infection in the intervention period. ICU admission and in-hospital mortality were sparse in both study groups. In the baseline group eleven patients were admitted to the ICU (7.3%) and two patients died (1.3%), compared to fifteen ICU admissions (8.7%) and four deaths (2.3%) in the intervention period. In total 145 eligible patients had a (partly) completed checklist. After the first exclusion round, the intervention group without a checklist contained 180 patients (figure 2). The percentage of eligible patients without a checklist after the second exclusion round in the intervention group was 46.7% (28 out of 60, see figure 2). Hence, approximately 84 patients in the intervention group without a checklist were actually eligible for the checklist (0.467 * 180 = 84). The estimated total study compliance was 63.3% (145/(145+84)).
DISCUSSION Our study demonstrated that the implementation of an antibiotic checklist on Aruba was associated with a significant increase in QI sum score, but performance did not improve on each individual QI. The checklist was used in 63.3% of the eligible patients, which illustrates that it was possible to implement the antibiotic checklist outside the Netherlands. Performance on taken blood cultures showed the largest improvement in the intervention period, which also applied to the Dutch checklist implementation study.10 Documentation of the indication was also improved, even though the baseline score was already high. This improvement explains the differences in the type of diagnoses between the baseline and the intervention group, as in the baseline group more patients with unclear documentation of indication ended up in the category of ‘more than two possible diagnosis’, or ‘diagnosis not covered by guidelines’.
The room for improvement in guideline adherence in the baseline period is in line with previous results from different parts of the world.12,18-22 For example, an observational study in a teaching hospital in Brazil concluded that 65.4% of the parenteral antimicrobial prescriptions was appropriate according to the local guidelines.18 It is disappointing, however, that guideline adherence did not improve in our intervention period, which is in contrast with other intervention studies, including our Dutch checklist study, aiming to improve antibiotic prescription.19-22 The main difference with these studies is that in our project improvement of guideline adherence was actually one of seven aims, while they only focussed on the antibiotic prescription.19-22 Obviously the Dutch checklist study also aimed to improve performance on the seven checklist-items, but an important difference is the lack of an easy accessible digital version of the guidelines on Aruba, which possibly hinders its use. Insufficient use of the guidelines can also be an explanation for the lack of improvement in de-escalation and IV-oral switch, as instructions for these items are given in the same antibiotic guideline booklet. In general, improvement on each QI might be hindered by these kind of barriers, as, in contrast to the Dutch study,13 we did not perform a barrier identification prior to checklist implementation on Aruba.
The checklist use of 63.3% was surprisingly high compared to 23.2% in the Dutch study.10 On Aruba some extra activities were undertaken, such as clinical lectures in the intervention period,
and
face-to-face-prompting,
which
has
been
described
as
an
effective
implementation intervention.6,23 Due to this approach the effect of checklist use and the other educational efforts intervened. In contrary to the Dutch checklist study,10 we have collected data on QI performance of twenty-eight intervention patients without a completed checklist. We have shown that the increase in QI sum score was stronger in the intervention group with a completed checklist, suggesting that checklist use has an additional effect. Few studies have been performed in the Caribbean and Latin America on the introduction of an ASP.22,24 This study was the first project performing an intervention that resulted in
improvement in the process of intravenous antibiotic use on Aruba. We met several criteria that are required for successful checklist implementation, including the implementation of a checklist that is based on previous research,25 and the organization of stimulating activities for checklist use.25-27 We illustrated the additional effect of completing a checklist to the effects caused by the other implementation activities by also collecting data of patients who were eligible but for whom a checklist was not used. More general, we provided data on antibiotic use in the Dr. HOH hospital, which was absent prior to this project, and we gained awareness on appropriate antibiotic use and the problem of antimicrobial resistance among healthcare workers on Aruba.
Our study has several limitations. Due to time limits we did not collect data of all eligible patients during the baseline- and intervention period. However, we have minimalized the chance of measurement bias by a systematic approach of inclusion. A disadvantage of these small numbers is that the sample size was too small to reach significance in some sub analyses. For example, the fact that the improved performance in adjustment to renal function was not significant, can be caused by the relatively small number of patients this QI applied to. Finally, our primary endpoint is a combination of seven QIs, but the increase in QI sum score was mainly caused by the high increase in collection of blood cultures. Therefore it can be questioned if the QI sum score is the best measure to reflect the improvement of the whole process of antibiotic use.
Pros and cons of further implementation of the checklist should be considered. Implementing a checklist is a complex intervention that requires a lot of motivation and effort of physicians. On the other hand, it provides an opportunity to involve physicians with direct responsibility for patients in the aims of antimicrobial stewardship.28 In contrast with most ASP interventions,7 the presence of infectious diseases experts is not required with checklist use. Another advantage is that the tool is based on validated quality indicators.11,12 An important
con is that the impact appeared only in some QIs in this study. Therefore, before further implementation is started, barriers should be identified that hinder improvement of all QIs in daily practice in the Dr. HOH. For example, the implementation of the local antibiotic booklet seems the first important step to improvement. By addressing these barriers, the process of antibiotic use could be further improved.
In conclusion, awareness created by the ASP activities including the implementation of the antibiotic checklist resulted in an increase in sum score of QIs for appropriate antibiotic use on Aruba. Further initiatives are necessary for further improvement, especially to improve antibiotic guideline adherence. Completion of the antibiotic checklist was high, suggesting that the implementation of the checklist is possible outside the Netherlands.
Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Competing interests No competing interests declared.
Acknowledgements We are grateful to dr. B.C. Opmeer from the Clinical Research Unit of the Academic Medical Centre for his assistance in the sample size calculation. We also want to give special thanks to Ebby Ruiz from the pharmacist department in the dr. E. Oduber hospital, who provided the lists of all patients treated with intravenous antibiotics. References
[1] World Health Organization (WHO). Antimicrobial resistance: global report on surveillance 2014. .
[2]
http://www.who.int/drugresistance/documents/surveillancereport/en/.
Accessed
October 2016.
[3] The Review on Antimicrobial Resistance. Tackling drug-resistant infections globally:
Final
report
and
recommendations.
2016.
http://amr-
review.org/sites/default/files/160525_Final%20paper_with%20cover.pdf
Accessed
September 2016.
[4] Luepke KH, Suda KJ, Boucher H, et al.,Past, present, and future of antibacterial economics: Increasing bacterial resistance, limited antibiotic pipeline, and social implications. Pharmacotherapy 2017;37:71-84.
[5] Tacconelli E.Antimicrobial use: risk driver of multidrug resistant microorganisms in healthcare settings. Curr Opin Infect Dis 2009;22:352–8.
[6] Goossens H, Ferech M, Vander Stichele R, Elseviers M,ESAC Project Group. Outpatient antibiotic use in Europe and association with resistance: a cross-national database study. Lancet 2005;365:579–87.
[7] Barlam TF, Cosgrove SE, Abbo LM, et al.,Implementing an Antibiotic Stewardship Program: guidelines by the Infectious Diseases Society of America and the Society for Healthcare Epidemiology of America. Clin Infect Dis 2016;62:e51–77.
[8] Davey P, Marwick CA, Scott CL, et al.,Interventions to improve antibiotic prescribing
practices
2017;2:CD003543.
for
hospital
inpatients.
Cochrane
Database
Syst
Rev
[9] Howard P, Pulcini C, Hara GL, West RM, Gould IM, Harbarth S, et al.,An international cross–sectional survey of antimicrobial stewardship programmes in hospitals. J Antimicrob Chemother 2015;70:1245-55.
[10] World Health Organization. The evolving threat of antimicrobial resistance. Options
for
action.
http://whqlibdoc.who.int/publications/2012/9789241503181_eng.pdf.
2012. Accessed
July
2016.
[11] Van Daalen FV, Prins JM, Opmeer BC, Boermeester MA, Visser CE, van Hest RM et al The effect of an antibiotic checklist on length of hospital stay and appropriate antibiotic use in adult patients treated with intravenous antibiotics: a stepped wedge cluster randomised trial. Clinical Microbiol Infect 2017; DOI: 10.1016/j.cmi.2017.01.019 [Epub ahead of print].
[12] Van den Bosch CM, Geerlings SE, Natsch S, Wille J, Prins JM, Hulscher ME.Quality indicators to measure appropriate antibiotic use in hospitalized adults. Clin Infect Dis 2015;60:281–91.
[13] van den Bosch CM, Hulscher ME, Natsch S, Wille J, Prins JM, Geerlings SE.Applicability of generic quality Indicators for appropriate antibiotic use in daily hospital practice: a cross-sectional point-prevalence multicenter study. Clinical Microbiol Infect 2016;22:888.e1-e9.
[14] Van Daalen FV, Geerlings SE, Prins JM, Hulscher ME.A survey to identify barriers of implementing an antibiotic checklist. Eur J Clin Microbiol Infect Dis 2016;35:545.
[15] Quan H, Sundararajan V, Halfon P, et al.,Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care 2005;43:1130–9.
[16] Subbe CP, Kruger M, Rutherford P, Gemmel L.Validation of a modified early warning score in medical admissions. QJM 2001;94:521–6.
[17] van den Bosch CM, Hulscher MEJL, Akkermans RP, Wille J, Geerlings SE, Prins JM.Appropriate antibiotic use reduces length of hospital stay. J Antimicrob Chemother 2016; DOI: 10.1093/jac/dkw469 [Epub ahead of print].
[18] Spoorenberg V, Hulscher ME, Geskus RB, de Reijke TM, Opmeer BC, Prins JM, Geerlings SE.A Cluster-Randomized Trial of Two Strategies to Improve Antibiotic Use for Patients with a Complicated Urinary Tract Infection. PLoS One 2015;10:e0142672. .
[19] Kawanami GH, Fortaleza CM.Factors predictive of inappropriateness in requests for parenteral antimicrobials for therapeutic purposes: a study in a small teaching hospital in Brazil. Scand J Infect Dis 2011;43:528-35.
[20] Hecker MT, Fox CJ, Cydulka RK, Siff JE, Emerman CL, Sethi AK, Muganda CP, Donskey CJ.Effect of a stewardship intervention on adherence to uncomplicated cystitis and pyelonephritis guidelines in an emergency department setting. PLoS One 2014;9:e87899. .
[21] Percival KM, Valenti KM, Schmittling SE, Strader BD, Lopez RR, Bergman SJ.Impact of an antimicrobial stewardship intervention on urinary tract infection treatment in the ED. Am J Emerg Med 2015;33:1129-33. .
[22] Popovski Z, Mercuri M, Main C, Sne N, Walsh K, Sung M, Rice T, Mertz D.Multifaceted intervention to optimize antibiotic use for intra-abdominal infections. J Antimicrob Chemother 2015;70:1226-9. .
[23] Guanche Garcell H, Pisonero Socias JJ, Enseñat Sánchez R, Fiterre Lancis I, Mir Narbona I, García Arzola B, Pardo Gómez G, Gutiérrez García F.Impact of a quality control program on antibiotic prescription in a hospital in Havana, Cuba. Rev Panam Salud Publica 2011;30:598-602.
[24] Weiss CH, Dibardino D, Rho J, Sung N, Collander B, Wunderink RG.A clinical trial comparing physician prompting with an unprompted automated electronic checklist to reduce empirical antibiotic utilization. Crit Care Med 2013;41:2563-9.
[25] Okumura LM, Silva MM, Veroneze I. Effects of a bundled Antimicrobial Stewardship Program on mortality: a cohort study. Braz J Infect Dis 2015;19:246-52.
[26] Bosk CL, Dixon-Woods M, Goeschel CA, Pronovost PJ.Reality check for checklists. Lancet 2009;374:444–5.
[27] Leape LL. The checklist conundrum. N Engl J Med 2014;370:1063–4.
[28] Pronovost P, Needham D, Berenholtz S, et al.,An intervention to decrease catheter-related bloodstream infections in the ICU. N Engl J Med 2006;355:2725–32.
[29] Hamilton KW, Gerber SJ, Moehring R, et al.,Point-of-prescription interventions to Improve Antimicrobial stewardship. Clin Infect Dis 2015;60:1252–8.
Figure 1. The Aruban Antibiotic Checklist
Figure 2. Flow diagram of participants
Table 1: patient characteristics
Characteristics
Sex. male Age, mean (SD) Infection, community acquired versus hospital acquired
Baseline N = 150
Intervention N = 173
Baseline (%)
Completed Checklists (%)
N=150 74 (49.3) 61.7 (20.1) 120 (80.0) / 30 (20.0)
N = 110 56 (50.9) 63.5 (18.7) 87 (79.1) / 23 (20.9)
Partly completed checklist (%) N = 35 16 (45.7) 58.1 (18.4) 32 (91.4) / 3 (8.6)
No checklist (%)
Intervention total (%)
N = 28 9 (32.1) 63.9 (17.3) 23 (82.1) / 5 (17.9)
N = 173 81 (46.8) 62.5 (18.5) 142 (82.1) / 31 (17.9)
P
0.65 0.22 0.63
Type of diagnosis Respiratory tract infection Urinary tract infection Skin and soft tissue infection Intra-abdominal infection Other infections Two diagnoses More than 2 possible diagnoses/ diagnosis not covered by guideline Participating wards Internal medicine Surgery Pulmonology Gastroenterology Urology Other Charlson comorbidity index, median (IQR) MEWS score 0 1 2 3 4 5 6 ≥7 AB started at the ED, yes
0.04 23 (15.3) 28 (18.7) 14 (9.3)
26 (23.6) 22 (20.0) 20 (18.2)
6 (17.1) 3 (8.6) 9 (25.7)
7 (25.0) 3 (10.7) 4 (14.3)
39 (22.5) 28 (16.2) 33 (19.1)
15 (10.0) 29 (19.4) 14 (9.3) 27 (18.0)
9 (8.2) 11 (10.0) 10 (9.1) 12 (10.9)
3 (8.6) 7 (20.0) 4 (11.4) 3 (8.6)
2 (7.1) 7 (25.0) 3 (10.7) 2 (7.1)
14 (8.1) 25 (14.5) 17 (9.8) 17 (9.8)
40 (26.7) 41 (27.3) 13 (8.6) 13 (8.6) 10 (6.7) 33 (22.0) 0 (0-2)
47 (42.7) 25 (22.7) 11 (10.0) 5 (4.5) 2 (1.8) 20 (18.2) 1 (0-2)
10 (28.6) 16 (45.8) 2 (5.7) 0 (0) 3 (8.6) 4 (11.4) 0 (0-1)
5 (17.9) 7 (25.0) 5 (17.9) 4 (14.3) 3 (10.7) 4 (14.3) 0.5 (0-2)
62 (35.8) 48 (27.7) 18 (10.4) 9 (5.2) 8 (4.6) 28 (16.2) 0 (0-2)
31 (20.7) 54 (36.0) 29 (19.3) 23 (15.3) 8 (5.4) 2 (1.3) 1 (0.7) 2 (1.3) 34 (22.7)
22 (20.0) 34 (30.9) 19 (17.3) 22 (20.0) 9 (8.2) 0 (0) 3 (2.7) 1 (0.9) 34 (30.9)
11 (31.4) 14 (40.0) 3 (8.6) 4 (11.4) 2 (5.7) 0 (0) 1 (2.9) 0 (0) 11 (31.4)
6 (21.4) 11 (39.3) 4 (14.3) 5 (17.9) 2 (7.1) 0 (0) 0 (0) 0 (0) 8 (28.6)
39 (22.5) 59 (34.1) 26 (15.0) 31 (17.9) 13 (7.5) 0 (0) 4 (2.3) 1 (0.6) 53 (30.6)
0.39
0.60 0.35
0.11
Table 2. QI sum score and score per QI QI
Baseline (n = 150)
Intervention (n =173)
Comparison
numerator/ denominator (%)
numerator/ denominator (%)
Difference %
OR*
CI (95%)
P
Blood cultures
22/150 (15)
119/173 (69)
+ 54
13.7a
7.7-24.2
<0.001
Cultures of suspected site of infection Guideline adherence
33/68 (49)
41/99 (41)
-8
0.69b
0.4-1.3
0.27
67/109 (61)
79/138 (57)
-4
0.84c
0.5-1.4
0.51
Adjustment to renal function
6/16 (38)
12/23 (52)
+ 14
1.84d
0.5-6.8
0.36
Documented antibiotic plan
116/150 (77)
154/173 (89)
+ 12
1.90e
1.0-3.7
0.06
De-escalation
15/53 (28)
21/77 (27)
-1
1.07f
0.5-2.2
0.85
IV-oral switch
29/85 (34)
41/98 (42)
+8
1.17g
0.6-2.2
0.64
SUMSCORE ≥ 50%
75/150 (50)
134/173 (77)
+ 27
3.67h
2.2-6.0
<0.001
*based on logistic regression analyses, taking covariates into account. Covariates: a type of diagnosis, sex, location start antibiotics, MEWS b type of diagnosis c community vs hospital acquired infection, type of diagnosis, AB last 30 days d sex e community vs hospital acquired infection, type of diagnosis f community vs hospital acquired infection g age, community vs hospital acquired infection, type of diagnosis, location start antibiotics h community vs hospital acquired and location start antibiotics