Appropriateness of Surgical Antimicrobial Prophylaxis Practices in Australia

Key Points Question What are the current surgical antimicrobial prophylaxis prescribing practices in Australia, and what factors are associated with their appropriateness? Findings This quality improvement study of 9351 surgical episodes found high rates of inappropriate procedural and postprocedural antimicrobial use across various hospital, patient, and surgical factors. The most common reason for inappropriate procedural use was incorrect timing, while duration greater than 24 hours was the most common reason for inappropriate postprocedural use. Meaning These findings suggest that the identified hospital, patient, and surgical factors should be considered as targets for development of tailored interventions to ensure appropriateness of surgical antimicrobial prophylaxis prescriptions.


Introduction
Surgical antimicrobial prophylaxis (SAP) refers to the administration of antimicrobials for the prevention of surgical site infections (SSIs).2][3][4][5] Australian point prevalence data from the Hospital National Antimicrobial Prescribing Survey (Hospital NAPS) [4][5][6][7] found that 40.3% of surgical prophylaxis prescriptions were classified as inappropriate and 45.2% as noncompliant with Australian national Therapeutic Guidelines. 8,9However, point prevalence methods do not capture the complexity of antimicrobial use in surgery, which should be assessed preoperatively and intraoperatively (or procedurally), and post procedurally.Some procedures (eg, clean procedures such as routine sterile dermatological surgery) do not require SAP. 9,102][13] Current Australian national guidelines 9 advocate for single-dose SAP but acknowledge that if postoperative doses are still considered despite lack of quality evidence, then SAP should not continue beyond 24 hours.Surgical prophylaxis audits have been recommended within the antimicrobial stewardship (AMS) component of the National Safety and Quality Health Service Standards 14, 15 and the national Clinical Care Standard for AMS 16 in Australia since 2015.In response to the needs of hospital AMS programs, the Surgical NAPS was developed to collect surgery-specific data, including surgery details and timing of antimicrobials, for benchmarking and targeted feedback of SAP prescribing. 17is national online auditing platform was implemented to describe SAP prescribing in Australian hospitals.The Surgical NAPS aimed to delineate hospital, surgical, and patient factors associated with appropriate SAP prescribing.

Methods Online Audit Platform Design
The Surgical NAPS online auditing platform has facilitated this multicenter, national, quality improvement study.The platform was codesigned through stakeholder consultation from a range of specialties (ie, infectious diseases, AMS, infection control, anesthesia, surgery) and collaboration with the statewide center responsible for surveillance of SSIs (Victorian Healthcare Associated Infection Surveillance System) and supplemented by a systematic literature review.5][6][7] User feedback facilitated survey refinement prior to development and pilot of the webbased survey.
Surgical NAPS data collection and analysis was approved by Melbourne Health Human Research and Ethics Committee.Informed consent was implied when participants agreed to the terms and conditions of the Surgical NAPS online audit platform prior to data submission.This also permitted the use of deidentified data for research purposes.This report was developed in accordance with the Standards for Quality Improvement Reporting Excellence (SQUIRE) reporting guideline. 18e Surgical NAPS list of surgical procedures and groups was developed from the Royal Australasian College of Surgeons' morbidity audit and logbook tool procedure list 19 and the Fellow of the Royal Australian and New Zealand College of Obstetricians and Gynaecologists' logbook procedure list and classification list. 20When multiple surgical procedure groups were documented for a patient, only the primary procedure system group was included.Definitions of procedural and postprocedural doses are outlined in eTable 1 in the Supplement.
The survey captures patient demographic characteristics (age, sex), clinical information (allergy status, antimicrobial choice, timing, duration) and procedure-related factors (surgical procedure, incision time documentation).Hospital demographics captured include location (state or territory), funding type (public or private), Australian Institute of Health and Welfare peer groups, 21 and Australian Bureau of Statistics remoteness areas. 22The use of Australian Institute of Health and

Data Collection and Collation
The Surgical NAPS audit was conducted annually between January 1, 2016, to June 30, 2018.Survey participation was voluntary and could be completed prospectively or retrospectively.Hospitals could adopt a convenience sampling method to audit either a targeted surgical specialty or all procedures conducted during a specific period.
Trained auditors collected data according to a standardized method and data collection form 17 (eFigure 1 in the Supplement).Auditors were primarily pharmacists, nurses, and infectious disease physicians who were provided with structured education and online training to ensure consistency of methods.All data from registered sites were entered in the Surgical NAPS online audit platform.
Ongoing support and advice from a central clinical support team was also available by phone and email to guide auditors, specifically with appropriateness assessments.
Appropriateness was assessed by surgical episode and by each prescription given for the surgery.Three categories of antimicrobial prescriptions were defined for each surgical episode: procedural, postprocedural, and existing (eTable 1 in the Supplement).Guideline compliance was assessed against the national guidelines 8 or local site-based guidelines available at the time of assessment.A summary of the guideline's recommendations for the general principles of SAP prescribing is included in eTable 2 in the Supplement.Appropriateness was a composite measure based on antibiotic choice, timing of administration, dose, duration, and repeat dosing applying the standardized Appropriateness Assessment Guide (eFigure 2 in the Supplement).For example, incorrect dosing may refer to a strength that is not consistent with the guidelines or is not clinically appropriate for that specific patient in regard to his or her body weight or renal function.
The purpose of this study was to assess the quality of SAP prescriptions in terms of appropriateness.Surgical episodes and prescriptions in which no antimicrobials were prescribed, existing antimicrobial therapies only, or prescriptions deemed not assessable were therefore excluded from the statistical analysis.Our statistical analysis also excluded repeat doses that were indicated but not given and postprocedural prescriptions with a treatment indication or an indication that was not assessable (Figure 1).
As a sensitivity analysis, crude overall appropriateness of the surgical episode was calculated and compared when antimicrobials were and were not prescribed for either procedural or postprocedural SAP.
For univariable and multivariable analysis, exclusions were doses deemed not assessable or subvariables with missing, low, and/or disproportionate numbers.The top 10 prescribed antimicrobials accounted for more than 90% of the data, and the remaining antimicrobials were classified as "other."

Statistical Analysis
Data are presented descriptively, with categorical data presented as frequencies and percentages.
Antimicrobial doses were stratified into procedural and postprocedural surgical prophylaxis.
Logistic regression models were used to identify hospital, patient, and surgical factors associated with appropriateness.Model fitting was started with the maximal model, including all relevant factors.Model selection was performed using a likelihood ratio test, and model fit assessed by residual plots.Mixed-effects logistic models fit with unique hospital identifiers as random intercepts provided the best fit to the data and were selected.
Crude estimates of appropriateness were adjusted for factors included in the model by calculating estimated marginal means, presented as adjusted appropriateness (AA) with 95% confidence intervals.These means are generated by estimating the outcome (ie, appropriateness) from the model and calculating an equal-weighted average across all subgroups.Here, the aim is to adjust for any biases caused by varying subgroup sizes in our sample.

Results
[25] In total, data on 12 982 surgical episodes were reviewed, including 6872 female patients (52.9%), 6069 male patients (46.8%), and 41 surgical episodes (0.3%) classified as other; median (range) patient age was 56.5 (0-105) years.Figure 1 provides an overview of included and excluded surgical episodes and antimicrobial prescriptions.There were 3631 surgical episodes excluded from statistical analysis.The remaining 9351 surgical episodes were included in statistical analysis (Figure 1).
A Subanalysis compared acute hospital peer groups based on their funding type (ie, public or private groups A, B, and C).Significant differences in appropriateness were identified in only 1 comparison, between public acute group B and private acute group B (χ 2 = 4.03; P = .04).

Postprocedural Prescriptions
A total of 4523 postprocedural surgical prophylaxis prescriptions (97.2%) were included for statistical analysis (132 prescriptions [2.8%] were not assessable).eTable 4 in the Supplement shows the crude and adjusted appropriateness values for all the hospital, patient, and surgical variables that were included in the multivariable model.

Reasons for Inappropriateness
For procedural prescriptions, 11.7% (1252) were prescribed when procedural SAP was not deemed to be required.For the 9488 procedural prescriptions that were clinically indicated, 33 For the remaining 1077 postprocedural prescriptions (38.5%), multiple reasons for inappropriateness could be applied to each prescription (1312 total reasons).Ninety-eight health care facilities contributed to this data subset.Surgical procedure groups with the largest proportion of contributing health care facilities were orthopedic (66.4%) and plastic and reconstructive (26.5%) procedures.Comparatively, dentoalveolar (2.0%) and ophthalmology (6.1%) surgery had the lowest contribution of different health care facilities.
The reasons for inappropriateness per surgical procedure group are presented in Table 2.

Discussion
The Surgical NAPS captures real-world prescribing behaviors for SAP across a broad range of procedures, hospital peer groups, and locations from both public and private hospitals.Our analysis has identified key targets for AMS programs, particularly related to timing and duration of SAP.An important finding of this study is that overall rates of appropriateness were low in the surveyed population.Comparatively, when SAP was not prescribed, high rates of appropriateness were demonstrated.Thus, patients not receiving SAP at all were considered a low priority target for SAP optimization.
Our data reveal that there are some interesting differences in rates of appropriateness across surgical specialties in Australian hospitals.Importantly, we did not demonstrate a significant difference between public and private hospital peer groups.
a Inappropriate prescriptions when surgical antimicrobial prophylaxis was indicated.
b Rationale for the reasons for inappropriateness are described in eFigure 2 in the Supplement: Surgical National Antimicrobial Prescribing Survey Appropriateness Assessment Guide.
c Timing was only assessable for procedural prescriptions.
interpretation of comparisons requires caution.Our development and application of appropriateness assessments is, to our knowledge, novel in the current literature as it accounts for circumstances in which guideline-noncompliant prescriptions may in fact be appropriate in relation to the prescription's context.
Orthopedic surgery was the most commonly audited surgical procedure group, accounting for 31.9% of all prescriptions.Our findings are consistent with the literature, in which orthopedic surgery has demonstrated low SAP guideline compliance 34,37,39,40 ranging from 24.1% 37 to 44.4%. 34 have identified that all surgical procedure groups demonstrated low AAs across procedural and postprocedural prescriptions.We believe low AAs may have the potential to adversely affect patient care and outcomes.All surgical procedure groups require ongoing support and AMS interventions tailored to their common reasons for inappropriateness to optimize SAP prescribing.
A Chinese prospective multicenter study 40 assessed the quality of SAP for 14 525 clean and clean-contaminated elective surgical procedures (orthopedic, vascular, gynecologic, and intestinal) and excluded emergency and contaminated and/or dirty procedures.Quality of SAP prescriptions was measured as adherence to the Chinese national guidelines, similar to our methods.Orthopedic surgery was also used as the reference group and gynecological surgery demonstrated significantly high rates of inappropriateness in comparison (odds ratio, 1.60; 95% CI, 1.37-1.88;P < .001). 40mparatively, the Surgical NAPS data included 156 hospital sites and emergency and contaminated or dirty procedures.These observations were similar to our results for postprocedural prescriptions, but not to our results for procedural prescriptions.a Inappropriate prescriptions when surgical antimicrobial prophylaxis was indicated.
b Rationale for the reasons for inappropriateness are described in eFigure 2 in the Supplement: Surgical National Antimicrobial Prescribing Survey Appropriateness Assessment Guide.
c Duration and frequency were only assessable for postprocedural prescriptions.
][68][69] Adoption of clinical decision support systems may improve evidence-based antimicrobial use. 69 We suppoimplementation of clinical decision support systems with a variety of interventions such as feedback 58 from the Surgical NAPS audits to synergistically improve SAP prescribing, feedback, and subsequent outcomes for patients and the health care system.

Limitations
This study has some limitations.First, participation in the Surgical NAPS is voluntary; thus, our results may not be generalizable across all hospitals.Second, auditor flexibility to perform convenience sampling introduced respondent bias due to the variation in data collected.A mixed-effects regression model was used to account for the intrahospital correlations; however, the effects of an unbalanced survey design may persist.Third, variability of appropriate assessments is probable, as one auditor's interpretation may have differed from another.To minimize this, an assessment rubric and support from the central clinical support team were available.After completion of the first Surgical NAPS report in 2016, 17 developers completed a small validation study and demonstrated a 6.7% disagreement rate when comparing assessments by local auditors with those conducted by the NAPS support team.This was deemed acceptable for this type of self-auditing by nonexperts; however, larger validation studies may be warranted.
In addition, there were limited data on clinical outcomes, as this component of the current survey was not mandatory.We support the introduction of mandatory outcome fields and the need for outcome data to provide greater insight and meaning for surgeons.We believe these data are required to drive change in prescribing behaviors.However, collection of data such as SSIs after admission is complex and would require additional systems and resources.

Conclusions
The Surgical NAPS data have identified high rates of inappropriateness for procedural and postprocedural SAP.Low and wide variation in appropriateness was noted across hospital and surgical factors, in particular surgical procedure groups.Reasons for inappropriateness varied according to the type of SAP, highlighting the need for targeted AMS interventions to address timing for procedural SAP and duration for postprocedural SAP.The Surgical NAPS data set is unique and extensive and continues to grow each year, with more and more Australian surgical centers participating.Ongoing analysis will continue to provide support and direction for AMS interventions, guideline development, and hospital policy.

Figure 2 .
Figure 2. Appropriateness of 10 740 Procedural Prescriptions per Surgical Procedure Groups

Table 1 .
39ported a 9.4% overall rate of appropriateness, while Hohmann et al39reported an overall compliance rate of 70.7%.The Surgical NAPS data demonstrated SAP appropriateness rates of 53.6% for procedural prescriptions and 36.9% for postprocedural prescriptions, with 48.7% as the combined overall rate of appropriateness.Definitions of appropriateness, compliance, and concordance vary in the literature; thus, Reasons for Inappropriateness of Procedural Prescriptions per Surgical Procedure Group

Table 2 .
Reasons for Inappropriateness of Postprocedural Prescriptions per Surgical Procedure Group