Economic Evaluations on Antimicrobial Stewardship Programme: A Systematic Review

- Purpose : To systematically review studies on cost-effectiveness of implementing Antimicrobial stewardship programmes (ASP) in the hospital setting. Methods : A systematic literature search was performed using electronic databases, such as EMBASE, PubMed/Medline, CINAHL, NHS and CEA Registry from 2000 until 2017. The quality of each included study was assessed using Joanna Briggs Institute Critical Appraisal Checklist for Economic Evaluations and Consolidated Health Economic Evaluation Reporting Standards Statement checklist. Results : Of the 313 papers retrieved, five papers were included in this review after assessment for eligibility. The majority of the studies were cost-effectiveness studies, comparing ASP to standard care. Four included economic studies were conducted from the provider (hospital) perspective while the other study was from payer (National Health System) perspective. The cost included for economic analysis were as following: personnel costs, warded cost, medical costs, procedure costs and other costs. Conclusions : All studies were generally well-conducted with relatively good quality of reporting. Implementing ASP in the hospital setting may be cost-effective. However, comprehensive cost-effectiveness data for ASP remain relatively scant, underlining the need for more prospective clinical and epidemiological studies to incorporate robust economic analyses into clinical decisions.


INTRODUCTION
Antimicrobial resistance is a growing public health threat. It has been associated with high mortality, with more than one million people die from drugresistant infections over the last two years [1]. Managing antimicrobial resistance is a costly exercise; in the United States alone, an excess cost of 20 billion USD was spent to treat these drug-resistant infections annually [2]. Resistance to new superbugs has reached an alarming level and more importantly, no major new types of antibiotics have been developed in the past three decades [3]. This looming public health threat has attracted the attention of various governments and global organisations, and thus, numerous strategies, including antimicrobial stewardship programmes (ASP), have been deliberated to combat antimicrobial resistance.
ASP aims to improve antimicrobial use (i.e. optimal selection, dosage, and duration of antimicrobial treatment) to optimise clinical outcomes and patient safety, reduce resistant infections, and minimise costs [4]. Two core strategies, which involves (1) prospective audit with intervention and feedback, and (2) formulary restriction and pre-authorisation, are identified as the foundation of ASP [4]. Davey et al. classified the antimicrobial stewardship intervention types into three: persuasive (e.g. audit, educational programmes, reminders, feedback), restrictive (e.g. formulary restrictions, authorisation, antibiotic cycling) and structural (e.g. computerisation of records, decision support system) [5]. Numerous studies [6][7][8] have shown the effectiveness of ASP and its interventions; and recently, there is an _________________________________________ increasing emphasis in evaluating cost-effectiveness upon implementation of these interventions. The recent reviews involved a large variation in the method and depth of ASP being economically evaluated (i.e. included predominantly costing studies [9], focused mainly on the clinical outcomes [10] or a narrative review with only one database was search [11]). The studies, therefore, were limited by substantial inter-study methodological heterogeneity, hindering meaningful conclusion to be drawn. Partial economic evaluations that took only drug acquisition cost of antimicrobial agents into consideration would have underestimated cost advantage and health benefit of ASP. Accordingly, the aim of the current study was to systematically review the economic evaluations of ASP in the hospital setting, facilitating informed decision making by policy makers and healthcare providers, in particular the countries with budget constraint.

Search Strategy
A systematic search of electronic databases, including EMBASE, PubMed/Medline, CINAHL, NHS and CEA Registry website from year 2000 to 2017, was performed by two independent authors (KM and NHI). The combinations of search terms, together with MESH terms, employed in this review were as follow: "(Antimicrobial OR antibiotic) AND stewardship AND (economics OR cost). These included ((""anti-bacterial agents""[Pharmacological Action] OR ""antibacterial agents""[MeSH Terms] OR (""antibacterial""[All Fields] AND ""agents""[All Fields]) OR ""anti-bacterial agents""[All Fields] OR ""antibiotics""[All Fields]) AND ""cost""[All Fields] OR ""costs and cost analysis""[MeSH Terms] OR (""costs""[All Fields] AND ""cost""[All Fields] AND ""analysis""[All Fields]) OR ""costs and cost analysis""[All Fields])) AND stewardship [All Fields]. The final search was done in October 2017.

Study Selection
Economic evaluations [i.e. cost-effectiveness analysis (CEA), cost-utility analysis or cost-benefit analysis] of ASP were included in this systematic review. The exclusion criteria were as follow: studies that included only direct cost of antimicrobial agents, studies presented only as abstracts with no full reporting of findings, review papers, editorial letter, non-English literatures, studies involving ASP in the outpatient setting, studies comparing the effectiveness of different antibiotic regimens and studies before year 2000 were all excluded. Partial economic evaluations were excluded due to the fact that the lack of information for performing an indepth quality assessment while studies before year 2000 were excluded due to the rapid advancement in ASP. The eligibility of all potential economic studies identified for inclusion was independently assessed by two review authors (KH and NHI). Any discrepancies on study inclusion were resolved through discussion and consensus.

Data Extraction and Collection
A standardised, electronic form was used to extract data from each economic study. Data (e.g. country, type of economic analysis, year of costing, perspective, time horizon, comparators, cost components, outcome measure, sensitivity analysis, economic findings) obtained from the included studies were independently extracted by two authors (KH and NHI).

Assessment of Methodological Quality
Quality assessment for all included economic studies were independently assessed (KM and NHI) using criteria as outlined in the Joanna Briggs Institute Critical Appraisal Checklist for Economic Evaluations [12]. The quality of reporting was evaluated using the Consolidated Health Economic Evaluation Reporting Standards checklist [13]. Any disagreement was resolved by discussion between the authors.

Description of Included Studies
A total of 313 studies was identified via the search. After removal of duplicates (n=34), 244 studies were excluded based on the pre-specified criteria ( Figure  1). A total of five economic studies investigating the cost-effectiveness of ASP in the hospital setting were included in the qualitative synthesis. Four studies were published within the past five years and were from Western countries. All studies were carried out in the different settings of hospital, with two studies conducted in both general and critical care units [14,15], one study in critical care unit only [16], one in urology ward [17] and one in surgical and nonsurgical wards and emergency department [18].  Table 1). The majority of them were conducted from provider (hospital) perspective [14,15,17,18] with only one study was from a payer (i.e. National Health System) perspective [16]. Hence, only direct medical costs were taken into account, which comprised of hospitalisation, healthcare providers and drug costs. Apart from that, additional cost such as computerised clinical decision support system was included in the study by Scheetz et al. [14] while Okumura et al. included research cost in the analysis [15]. Development costs of antibiotic checklist, implementation costs (i.e. web site, e-learning, briefing, posters and laminated pocket version) and operational costs were included in the study by Van Daalen et al. [18]. All costs were obtained retrospectively in all studies except one study [16] obtained the average cost from the published literature.
Four studies compared ASP to standard care [14,[16][17][18] [18] developed and implemented the use of antibiotic checklist (i.e. persuasive intervention) in nine Dutch hospitals and evaluated the cost-effectiveness of checklist usage as ASP intervention. Antimicrobial stewardship interventions in the study by Ruiz-Ramos et al. [16] was comprehensive, which included antimicrobial restriction (both antibiotics and antifungals), formal consultation, implementation of protocols for de-escalation and guidelines for antibiotic prophylaxis or treatment, formal reassessment of antimicrobials and implementation of computer-assisted decision support (i.e. persuasive, restrictive and structural interventions).
Several modelling techniques were reported across the four included studies, including decision tree [14,16], a 30-day Markov model [15] and costminimisation model [17]. Van Daalen et al. study [18] did not adopt any modelling techniques. Four studies employed incremental cost-effectiveness ratio (ICER) as model outcome [14][15][16]18], except one used return on investment [17]. All included studies, except one [18], conducted both deterministic and probabilistic sensitivity analyses to compute the effect of uncertainty in input variables on model outcome. Only one study ran the subgroup analysis since the cohort were from two different groups of patients (i.e. one had infection-related indications and the other had severe underlying diseases such as cancer) [17]. All studies concluded that ASP was cost-effective in either short-or longterm setting.

Quality Assessment
The quality assessment and quality of reporting of the five studies were summarised in the Table 2 and  Table 3, respectively. In general, most studies were able to adequately report at least 16 items that essential for performing an economic evaluation, indicating that these studies were of good quality of methodology.

DISCUSSION
Whilst the five economic evaluations included in the current review suggested that implementation of ASP was a cost-effective option in the hospital setting, these findings need to be interpreted with caution. The lack of standardisation in outcome measure of the economic evaluations in the ASP setting, in addition to the inconsistencies in the study design and depth of the ASP interventions employed, has hindered the usefulness of data on costs and benefits in the current evidence-based practice. It is important to note that there is a wide variation in the outcome measure reported among the economic studies in the current review [e.g. ICER per life-years gained, ICER per quality-adjusted life years (QALYs) gained or ICER per averted death in 30 days)]. In addition, return on investment was reported as an outcome measure in one of the studies [17]. Return on investment is a form of cost-benefit analysis that measures the cost of program versus the financial return from that programme, calculated as total benefit minus total cost (net benefit) over total cost [19]; the benefits of intervention are converted into monetary [20]. However, a return on investment analysis typically relies on short-term returns and often ignores the health of beneficiaries or patients [20]. Future studies should consider QALY, which takes into account the quality of life of those who experience the health outcomes, as a standardised outcome measure (effectiveness) [16] since it permits comparability across the economic analyses. The outcome probabilities (i.e. input variables for effectiveness data) for all included studies were mainly obtained from the published literatures [14,16] or historical cohort studies [15,17,18]. These retrospective data could be subject to bias due to incomplete record and loss of follow up. Furthermore, estimation of the levels and long-term effects of antimicrobial resistance that will have on patient evolution as well as disease transmission are not being taken into consideration in these economic studies, and thus, underestimating the impact of ASP. All studies attained the cost inputs retrospectively and through gross costing except the study by Ruiz-Ramos et al. used average cost from published literatures in other countries to estimate antimicrobial cost per patient in critical care unit [16]; the appropriateness of adopting cost data from other countries into the analysis is a major consideration.
The included economic studies had high level of heterogeneity due to differences in the study setting, ASP strategies, clinical benefits and economic measurements. Therefore, it is difficult to determine which type of ASP strategies is the most costeffective. This is further complicated by the dearth of economic evaluations in determining the costeffectiveness of ASPs. The ASP interventions implemented in the economic studies included in the current review fulfilled most of the core elements of hospital ASP requirements set by Centers for Disease Control and Prevention, which includes leadership commitment, accountability, drug expertise, action, tracking, reporting and education [21]. Whilst multiple approaches in ASP interventions (i.e. bundled ASP strategy) are expected to provide better health outcomes, however, the driver of the cost-effectiveness cannot be distinguished and determined. Of note, none of these economic evaluations was conducted from societal perspective. Most economic studies were conducted from institutional (i.e. payer and provider) perspective given that the funding of ASP comes from the hospital administration; the societal benefits of ASP (e.g. loss of productivity due to multidrug resistant infections), however, should not be underestimated [14].
The difficulty in directly extrapolating the published economic findings to the appropriate patient populations that reflect the clinical caseload encountered in daily practice remains to be resolved since the economic studies included in the current review were conducted in critical care, urology wards, surgical and non-surgical wards and emergency department. Apart from that, other factors such as variability in healthcare systems and the resistance pattern which may differ according to the geographical areas pose great challenges to transfer cost-effectiveness data between countries.
The current review emphasises the need for research on a more systematic approach to evaluate the cost-effectiveness of individual ASP programmes. Robust health economic evaluations will provide a reasonable foundation for decisionmaking and thus, facilitating the ideal allocation for limited resources countries to fight against antimicrobial resistance. In general, the quality and execution of the included economic studies on economic evaluation were considered satisfactory. In the current review, the Joanna Briggs Institute Checklist for Economic Evaluation [12] was used to appraise the economic studies in addition to the Consolidated Health Economic Evaluation Reporting Standards checklist [13]. A 'YES' on the Consolidated Health Economic Evaluation Reporting Standards checklist may not adequately assess the quality of the criteria, but may only to indicate the completeness of the reporting rather than whether the choices were appropriate or justified. The present work, however, has shortcomings. Only studies published in English were included, and thus, leading to the small numbers of retrieved articles. The high level of the methodological heterogeneity, in terms of ASP interventions, that was noted in the current review is another limitation.

CONCLUSION
Although implementing ASP in the hospital setting is considered to be cost-effective, existing economic evaluations are limited by their great variation in the study design, outcome measure, types of ASP intervention and clinical settings. Therefore, future research evaluating the economic impact of ASP should consider using a standardised outcome measure with a longer time horizon of analysis. Robust economic studies to assess the cost-effective component of ASP across an extended clinical setting are anticipated.

TRANSPARENCY DECLARATIONS
All authors have no conflicts of interest to declare. This review was partially supported by UiTM internal grant (600-IRMI/DANA5/3/BESTARI (044/2017).   Is there a well-defined question/objective?

2.
Is there comprehensive description of alternatives?

3.
Are all important and relevant costs and outcomes for each alternative identified?

4.
Has clinical effectiveness been established?

5.
Are costs and outcomes measured accurately?

8.
Is there an incremental analysis of costs and consequences? + -+ + +

9.
Were sensitivity analyses conducted to investigate uncertainty in estimates of cost or consequences?

10.
Do study results include all issues of concern to users?

11.
Are the results generalizable to the setting of interest in the review?