The competing purposes of antibiotic documentation

To understand how best to implement antibiotic indication documentation and engage stakeholders, it is important to understand the myriad and often competing purposes of documentation. Generally, documentation—both related and unrelated to antibiotics—serves many purposes. These vary from the immediately patient relevant (eg, improving interclinician or patient/clinician communication, enabling prospective audit and feedback by pharmacists) to those with delayed clinical relevance (eg, enhancing retrospective research), and those largely for administrative purposes (eg, billing, resource distribution). In some countries, such as the USA where 35 of the studies included in the scoping review occurred, documentation also serves for legal protection (ie, defensive medicine).5 For antimicrobial stewardship in particular, there is another useful purpose of documentation to consider: improved mindfulness. Specifically, the act of documentation itself may trigger a mindful moment that forces clinicians to pause and move from system 1 (fast) to system 2 (slow) thinking.6 Slower, deliberative thinking has been shown to serve as an antibiotic self-timeout which reminds clinicians to reconsider antibiotic necessity and appropriateness.7 Similarly, accountable justification, which requires clinicians to explicitly justify an antibiotic treatment plan, has been shown to reduce inappropriate antibiotic prescribing.8

Because documentation purposes and thus priorities vary, changes in antibiotic indication documentation can be difficult to implement. For example, Saini and colleagues found that notable barriers to documenting antibiotic indications included competing priorities and time burden.9–11 In our practice, we find that documentation that is not directly and immediately clinically useful is often perceived by clinicians as an irrelevant administrative burden. Likewise, documentation interventions that hinder billing or quality compliance (eg, public reporting) may encounter leadership resistance. Thus, to engage clinicians and improve rates of antibiotic indication documentation, antimicrobial stewardship programmes must reframe the need for documentation in terms of clinical relevance, while not running counter to administrative needs.9

One way to improve clinician buy-in is by focusing on the salient, patient-relevant reasons for improving antibiotic documentation. Take, as one example, care transitions. Nowhere is the need for documentation more acute than during transitions of care, where adverse events are common, often harmful, and where accurate documentation can serve as a critical tool for improving communication.12 Because other forms of communication (eg, notes, verbal) are less common and reliable during care transitions (eg, to retail pharmacies), accurate documentation of antibiotic indications can prevent medication errors.13 When antibiotic prescriptions sent to pharmacies do not include the indication within the patient instructions, the pharmacist is left inferring the indication, reducing their ability to confirm prescription accuracy and prevent errors.13 Antibiotic stewards who wish to improve antibiotic indication documentation at discharge for one purpose (ie, optimising discharge antibiotic prescribing) could therefore have more successful clinical buy-in if they instead framed documentation in terms of improving communication and reducing harmful errors (eg, ‘bug-drug’ mismatch).

What do we mean by good documentation?

As we pause to consider best practices for implementation, it is important to first define what we mean by antibiotic indication documentation. Critically, this includes where documentation should occur. As clinicians ourselves, we recognise that clinicians often document antibiotic indications not once—but many times—including in clinician notes, interclinician handoff tools, inpatient antibiotic orders, outpatient antibiotic prescriptions and patient instructions. This complexity is not accounted for in most studies of antibiotic indications: in Saini and colleagues’ paper only 19 of the studies refer to documentation of the antibiotic indication as part of the prescription itself. For better or worse, many documentation interventions are operationalised in orders, including adjustments to the electronic health record (EHR) through order sets, drop-down menus and prompted indications.4 This type of EHR-driven documentation can serve as a nudge to improve prescribing, but come with unintended costs.14

Notably, all documentation interventions require additional clinician work. While demonstrably important, EHR-focused documentation interventions fail to account for documentation that may—or may not—already be occurring elsewhere. For example, the rise of EHR note templates can both hinder and benefit documentation. While note templates may remind clinicians to include antibiotic indications, they can lead to ‘note bloat’ and additions that are clinically irrelevant. The disjointed approach to documentation can lead to alert fatigue and annoy end-users who fail to see the value in including an indication in the order when it is already documented elsewhere.10 Similarly, too many documentation requirements can lead clinicians to over-ride or bypass mandatory fields.10 11 Finding the ‘sweet spot’ of antibiotic documentation to improve compliance and help stewardship without impeding clinician workflow is critical.

Best practices for implementation of antibiotic indications

We have summarised some high-level best practices for increasing antibiotic indication documentation in table 1. Critically, improving antibiotic indications requires an easy, user-centred, well-designed process. When not designed and implemented with end-user engagement, clinicians may manipulate or game clinical decision support tools to bypass restrictions and take short cuts. For example, in a study by Baysari and colleagues, the investigators created a drop-down system for restricted antibiotics that was meant to improve antibiotic-prescribing habits.10 The drop-down included only appropriate indications and clinicians were required to obtain stewardship approval if the planned use was for an indication other than in the drop-down. Rather than waste time by calling for approval, clinicians instead bypassed the approval process and gamed the system by selecting ‘appropriate’ indications, even when that indication was untrue.10 Because of this gaming, the intervention had no effect on antibiotic prescribing or documentation habits.10 When interviewed about these workarounds, clinicians reported they wanted to save time and minimise disruptions to their workflow.10

In their scoping review, Saini and colleagues mapped barriers and facilitators using the Capability, Opportunity, Motivation (COM-B) model for behaviour change. They acknowledge opportunity in the form of incentives, rewards and EHR improvements may overcome barriers to documentation such as lack of time and logistical challenges.4 Long term, we need better systems to allow open interactions between all members of the patient’s care team and improve interclinician communication around antibiotic indications. For example, in Taiwan, cloud-based technology allows clinicians, nurses and pharmacists to access patients’ prescription history and has resulted in enhanced medication safety.15 A similar system exists in the USA for opioids and controlled substances; however, this is a state-dependent rather than a nationwide database. As technology improves, having interoperable systems to allow communication of medication information—including antibiotic indication, dose and duration—may become more feasible.

Additionally, Saini and colleagues noted strategies to reduce alert fatigue and increase ease of documentation could help with clinicians’ capability and motivation to provide an indication on the antibiotic prescription.4 For example, automating or prepopulating indications based on documentation elsewhere (eg, in notes) could reduce the burden on clinicians and prevent gaming. Or the reverse—prepopulating notes based on antibiotic orders or prescriptions may better fit into workflow where orders are often written before notes. This latter solution would also provide an opportunity to incorporate ‘nudges’ to prod clinicians towards better prescribing.

In summary, we operate in an imperfect system where antibiotic indication documentation serves many purposes and occurs in many forms which can impede implementation. Myriad strategies to improve documentation have been attempted with mixed results. Until we have a system that can facilitate communication across all pertinent individuals, we must work to optimise what we have. This includes leveraging technology and behaviour change strategies to help documentation serve multiple purposes simultaneously. Doing all that, without increasing burden on clinicians, is the key to any successful antimicrobial stewardship—or documentation—intervention.