Evaluation of early implementations of antibiotic stewardship program initiatives in nine Dutch hospitals

Antibiotic resistance is an increasing worldwide threat to patient safety and quality of care [1, 2]. A correlation between antibiotic (over) use and increasing antibiotic resistance is widely acknowledged [3]. Hospitals can influence antibiotic use with improved, prudent antibiotic prescription, as up to 50% of prescribed antibiotics may be inappropriate or even unnecessary [4]. To curb the increasing resistance, hospitals started antibiotic stewardship programs (ASPs) as quality initiatives for infection prevention and control. An ASP ensures proper use of antibiotics with the best patient outcomes, less risk of adverse effects, optimal cost-effectiveness and to reduce or stabilise levels of resistance [5]. Guidelines from professional societies, such as the Infectious Diseases Society of America/Society for Healthcare Epidemiology of America (IDSA/SHEA) recommend expert-based strategies and interventions to implement ASPs in hospitals [6‐8]. In the Netherlands, the Dutch Working Party on Antibiotic Policy (SWAB) is responsible for guideline development, education and surveillance for optimal antibiotic use. This organisation recently published a vision document, stressing the need for Dutch hospitals to form an ASP team and start planning future ASP initiatives as per January 2014 [9].

There is no consistent approach in local implementations of ASP in hospitals [10]. Hospitals combine expert-based strategies and subsequent ASP interventions in many different variations [4]. In a prior systematic review, neither a consistent implementation strategy nor a sufficiently described implementation rationale could be identified in international ASP studies (van Limburg M, Köck R, Karreman J, Sinha B, de Jong N, Wentzel J, Friedrich A, Hendrix R, van Gemert-Pijnen J, “Towards an Implementation Strategy for Antibiotic/Antimicrobial Stewardship: A Systematic Review”, Under review). To understand how these local implementations of ASP take form it is important to use a bottom-up analysis from within hospitals. In this approach, we involve the relevant local stakeholders in the hospitals we aim to assist with implementing ASP, as opposed to an expert-driven or authoritative top-down approach.

Progression in implementation can be assessed with the concept of ‘maturity’. Maturity models were first coined by Paulk et al [11] and are used in many disciplines to model or evaluate improvement of organisations and processes [12]. Maturity models describe a levelled progress of how, in this case, an ASP implementation gradually improves from an ad-hoc state to a structured and managed state and eventually to a measurable and self-optimizing state [11].

Maturity assessments are already used in the context of ASP. Earlier maturity assessments of ASPs were done by the Antibiotic Strategy International group who introduced a comprehensive list of potential structure indicators to assess maturity of ASP in hospitals in several European countries [13]. In short, this means that maturity of ASP depends on the number of expert-based structure indicators that are implemented as part of ASP in a hospital.

This study does not attempt to assess maturity by scoring a completeness of an ideal program, but analyses differences in local implementations of ASP interventions that are recommended in guidelines. Our aim is not primarily to identify which interventions are or should be implemented, but more importantly, to identify where rationale plays a role in implementing ASP interventions and what the consequences are for expert-based ASP strategies. The present study is the first step in our bottom-up assessment with local ASP experts to give an overview of which recommended interventions are implemented, or are being developed, combined with an importance analysis of these interventions. Incongruent responses from stakeholders show where possible difficulties or barriers surface when implementing ASP conform currently available expert guidelines and indicate possibilities for decision support. A follow-up study is required to understand the deeper rationale behind these differences and barriers and how local experts can be assisted with relevant implementation advice.

An online questionnaire using LimeSurvey was made available from August to November 2013. Clinical microbiologists, clinical pharmacists and infectious disease physicians were approached as these expert types are identified as being key-stakeholders in ASP (van Limburg M, Köck R, Karreman J, Sinha B, de Jong N, Wentzel J, Friedrich A, Hendrix R, van Gemert-Pijnen J, “Towards an Implementation Strategy for Antibiotic/Antimicrobial Stewardship: A Systematic Review”, Under review). Ethical approval was given after revision by an internal representative of the ethical board (reference: BCE14039) and no further ethical review procedures were deemed necessary. We sent a questionnaire invitation to known infection control experts recruited in earlier research projects. These experts represented academic and non-academic hospitals in the Dutch eastern border region and they were asked to distribute the questionnaire among clinical microbiologists, clinical pharmacists and infectious disease physicians tasked with ASP in their hospitals. As at the time of the questionnaire ASP teams were not necessarily formed at the hospitals yet, we chose to approach key stakeholders rather than teams directly. We also analysed the response of key stakeholders individually (not clustered per hospital) as stakeholders may have possible differences in views on ASP interventions (even within hospitals).

The questionnaire was structured around ten implementation topics based on a literature review (see Table 1). The topics that were addressed involved ASP program initiatives, presence and importance of commonly used interventions and their possible characteristics, and possible outcomes and effects relevant for assessing evidence for the effectiveness of ASP. The extraction of implementation topics concerning ASP maturity or implementation assessments, attitude or opinion surveys towards ASP or practical implementation toolkits for ASP are presented in detail in Table 1.

The final questionnaire was adapted for the Dutch healthcare context, reduced and validated by microbiologists to have a final scope of 47 questions for an acceptable answering duration of 15–20 minutes. A literal English translation of the Dutch questionnaire can be found in Additional file 1 for reference.

Responses were imported into SPSS21 for statistical analysis. Incomplete responses were excluded. Frequency cross-tables of responses stratified by hospital type (academic/non-academic) were used for reporting all results, as we anticipated differences between the two hospital types [10].

Both academic hospitals and non-academic hospitals in the Dutch border region are busy with antibiotic stewardship initiatives. When assessing local implementations of ASPs, a clear difference in academic and non-academic hospitals can be seen. Generally, academic hospitals are experimenting with comprehensive programs of recommended interventions in pilot ASPs. Non-academic hospitals have implemented interventions that are required by national guidelines and are expanding the program with other important interventions for a comprehensive ASP. Obviously the academic nature and larger size of academic hospitals support a more experimental, comprehensive approach, whereas non-academic hospitals seem to take a more practical approach by focusing on easily obtained gains first. Some existing processes can be reused as many activities that are now under the ASP umbrella are not necessarily new. For example, many hospitals have had remote consults with microbiologists in place for years. Also, this difference between academic and non-academic hospitals is strengthened by the fact that non-academic hospitals seem to have more difficulties in securing adequate resources for ASP.

When we combine maturity modelling with the findings of the current progress in ASP in hospitals included in our questionnaire, a number of conclusions can be drawn:

The assessment of the implementation of ASP can be expanded further by understanding the rationale and considerations local experts have based on expert-based recommendations in available regulatory documents. Patel et al. concluded that little guidance is offered on the practical aspects of implementing ASPs and that non-academic hospitals in the US need to overcome implementation issues by accounting for unique characteristics of their institutions [18]. Understanding these issues combined with these unique characteristics can eventually be used for parameters to customise local implementations of expert-based recommendations.

Earlier maturity assessments identified a set of structure indicators or characteristics that need to be present [13, 19]. Prudence should be shown to assess the maturity of ASP implementation when using these structure indicators, as local experts may have valid reasons to deviate from expert recommendations. For example, a non-academic hospital may not have infectious disease physicians but a clinical microbiologist performing similar tasks. If available maturity assessments are strictly observed, this example hospital is missing an important structure indicator, while in practice the ASP was implemented differently from the ideal program as proposed and assessed by experts. Therefore, tallying whether expert-recommended interventions are present is not enough to capture the progress and maturity of ASP implementations.

By applying convenience sampling, we contacted respondents indirectly via infection control experts in the Dutch eastern border region and we depended on their willingness and involvement with ASP to find and persuade colleagues to partake. Due to the novelty of ASP, some experts declined participation as they claimed they were not sufficiently progressed with ASP initiatives yet. As a result, this study assessed only nineteen stakeholders from nine hospitals in the Dutch border region. Conclusions of this study do not reflect the entirety of Dutch hospitals. However, this was not the primary goal of the study, as we wanted to assess the implementation process of early ASP initiatives, for which the sample size sufficed. Another limitation was that by reducing the number of questions, not every intervention was assessed in high detail. We also omitted questions regarding microbiological diagnostics, infrastructure or tests and policy forming to keep the questionnaire at an acceptable length.

To keep the momentum going of ASP implementations and to improve the chances of success of these implementations, future research is necessary to obtain further insights into the rationale and issues that local experts consider during local implementations. What considerations did they have and why? At academic hospitals it would be interesting to assess what was learned with experimentation in pilot implementations. How did they arrive at the interventions as used in pilots and how will these pilots be scaled up? Case studies of these pilots can lead to concrete examples of potential and different implementations that can be used for advising other hospitals various configuration possibilities of ASPs. At non-academic hospitals, future research will be more focussed on practical issues in implementing ASP. What do local experts in non-academic hospitals consider in practice when implementing ASP? What are easily obtainable gains? What are barriers?

These different case studies, potential issues and rationale behind characteristics between interventions can be used as parameters that influence the configuration of to-be implemented ASPs. By inventorying all these parameters, local experts can be supported in customising and implementing an ASP that fits their hospital. Eventually, these parameters could be used to synthesise an implementation maturity toolkit. We plan to design a decision aid for experts in academic and non-academic hospitals that will generate a customised implementation advice for ASP fit for their local conditions.

Advising an ASP implementation is not straightforward. Experts who are tasked to introduce ASPs in their hospitals use expert-driven guidelines but need to transpose these guidelines to locally implemented interventions. This transposition leaves leeway for experimentation and considerations and leads to local differences in implemented intervention and ASPs. Progress can be made by assisting local experts with implementing ASPs in their hospitals. This assistance, however, needs to take into account that local conditions need to be translated into practical implementation advice and that a ‘one-ASP-for-all’ advice does not meet the needs of local experts. A bottom-up assessment with local experts can find parameters that influence local implementations of ASPs. These parameters can be used as input for a decision aid that generates a customised advice for a local implementation of an ASP.