Diagnostic Stewardship for Healthcare-Associated Infections: Opportunities and Challenges to Safely Reduce Test Use

dehydrogenase; NAAT, nucleic acid ampli ﬁ cation test; HABSI, hospital-acquired bloodstream infection; CLABSI, central-line associated bloodstream infection; VRE, vancomycin-resistant Enterococcus ; VAP, ventilator-associated pneumonia; ATS, American Thoracic Society; MRSA, methicillin-resistant Staphylococcus aureus ; CFU,

mortality or other patient-centered adverse outcomes, despite reduced antimicrobial use. 3 Recognizing this, The Joint Commission now requires ASPs for hospital accreditation, and the CMS has proposed ASP standards in acute-care hospitals, critical-access hospitals, and long-term care facilities. 4 Diagnostic stewardship practices are increasingly common among hospitals, often classified as quality improvement or under the umbrella of antimicrobial stewardship. Examples include targeted staff education with regard to test ordering, interpretation, or proper specimen collection, as well as laboratory "prior authorization" policies designed to limit tests. In the near future, the CMS may begin to require diagnostic stewardship in the form of an approved clinical-decision support system, to receive full payment for advanced diagnostic imaging tests (through the Appropriate Use Criteria program established under the Protecting Access to Medicare Act of 2014, pending final approval by the CMS). 5 Diagnostic stewardship has a potentially important role in HAI surveillance. The Centers for Disease Control and Prevention (CDC), through the National Healthcare Safety Network (NHSN), monitors >70% of all US hospitals for several hospital-related infections including SSI, CLABSI, CAUTI, ventilator-associated pneumonia (VAP) (now more broadly characterized as a possible ventilator-associated pneumonia, or PVAP), and healthcarefacility-onset Clostridium difficile infection (HO CDI). 4 Surveillance-based definitions, such as those developed by the NHSN for HAI events, are pragmatically designed for surveillance purposes and are not intended for use in the clinical evaluation and care of patients. For example, current NHSN surveillance definitions for HO CDI require only a positive test for C. difficile from an unformed stool specimen on or after hospital day 4, irrespective of patient symptoms, clinical condition, alternative diagnoses, or multistep testing laboratory algorithms, whereas clinical practice guidelines require clinical indications of disease and advocate that testing of asymptomatic patients is not clinically useful. 6,7 Many surveillance definitions cannot necessarily be used to distinguish true infections from false-positive tests.
Overuse of tests is predicted to increase false positives that trigger needless downstream cost and treatment that may cause harm for the patient. Conversely, test underuse risks missed diagnoses and potential harm related to untreated conditions. As with antimicrobial utilization, we hypothesize that there exists a state of optimal test use for HAIs in at-risk patients.
HAI rates based on surveillance definitions may over-diagnose CAUTI, CLABSI, HO CDI, hospital-acquired pneumonia (HAP), and VAP, estimated up to 37%, 8 30%, 9 15%-53%, 10,11 47%, 12 and 58%-68%, 13,14 respectively. Furthermore, the results of new, highly sensitive molecular diagnostics that detect minute amounts of a microbial target, such as nucleic acid amplification testing (NAAT) for C. difficile toxin gene(s), may identify colonized rather than clinically infected patients. This misattribution of colonized patients can artificially increase HAI rates. 11 Diagnostic stewardship is defined as coordinated systems or user-based interventions designed to promote evidence-based utilization of diagnostic tests, with the primary goals of improving value and care quality and safely reducing cost. It has the potential to address falsely inflated HAI rates due to overtesting. 15,16 Diagnostic stewardship has been described recently by Morgan et al 15 to occur in three stages: preanalytic (test-related decision making and specimen collection), analytic (relating to laboratory practices including protocolized or reflex test algorithms), and postanalytic (eg, selective reporting of antimicrobial susceptibility data to encourage the use of narrower spectrum agents).
Diagnostic stewardship has been shown to effectively reduce a variety of unnecessary general inpatient medicine tests, from excessive or redundant daily inpatient labs to diagnostic imaging. 16,17 Diagnostic stewardship strategies are varied and include user-based approaches (eg, auditing, price display, and provider feedback) and systems-based approaches (eg, modifications to the computerized physician order entry (CPOE) system requiring selection of an indication for testing and inappropriate specimen rejection).

IMPLEMENTATION CHALLENGES AND SAFETY CONCERNS
While reducing unnecessary tests for HAIs can have many potential benefits for the patient and hospital, test underutilization raises the possibility for serious infections going undiagnosed and untreated. For example, while excessive C. difficile testing may identify patients with colonization or resolving infections (which is not only a waste of resources but also leads to unnecessary treatment), more restricted testing might result in unrecognized and untreated CDI (resulting in harm to individual patients and greater risk of cross infection) or empiric treatment for CDI without testing as a workaround (resulting in unnecessary treatment in a subset of patients). A major objective for diagnostic stewardship for HAIs is to identify the "sweet spot" of test utilization that minimizes overdiagnosis and false positive results while maximizing appropriately indicated testing and true positive results. This spot likely will be infection and population (eg, disease prevalence) specific.
Because HAI-related tests pose unique risks associated with reduced testing, which outcomes should be tracked to monitor patient safety? General outcome measures, as in ASPs, could include length of stay, antimicrobial resistance rates, antimicrobial use, CDI rates, mortality, and readmission. Potential comorbid complications tailored to the HAI(s) in question are also an essential stopgap that should prompt reconsideration for testing. For instance, following the introduction of a "stewardship of culturing" aimed at reducing CAUTIs, Mullin et al 18 monitored overall rates of hospital-acquired (HABSI) infections, given the potential for complications of untreated urinary tract infection. However, outcome data in this and other HAI-related diagnostic stewardship studies were collected in aggregate and were not stratified to patients for whom the test was prevented and thus were at the highest risk for untreated infection. Ideally, prospective monitoring for HAIs should be performed for patients before and after diagnostic stewardship interventions to assess the direct patient-centered impact of these interventions in addition to aggregate data. These safety measures have largely been overlooked in the limited literature to date that has assessed diagnostic stewardship for HAIs, and incorporation in future studies presents significant logistical hurdles. Discordance between surveillance and clinical definitions for HAIs or those without a clear gold-standard clinical definition (eg, CDI) present challenges to evaluating safety when differentiating true positives remains elusive. Similar to ASPs, there is no one-size-fits-all approach to reducing unnecessary HAI tests among all institutions. Information technology and CPOE capabilities, population characteristics, local ordering practices, HAI incidence, and laboratory test performance characteristics should all be taken into account when developing a diagnostic stewardship approach. Institutional factors, such as laboratory and stewardship activity, hospital administration support, and barriers such as provider pushback, are additional factors to consider. As with any quality improvement effort, process measures are also vital to ensure that stewardship interventions are having their intended effects, such as testing rates (including tests that are rejected from processing) and rates of the target HAI. Table 1, incorporating the stages-of-testing concept of Morgan et al, 15 lists examples of diagnostic stewardship strategies for HAIs from the literature as well as other potential strategies that could be used to optimize test utilization. As in ASPs, engineered flexibility is key in the event that special circumstances require deviation from prescribed practices, the diagnostic stewardship strategy fails to achieve intended goals, or patient harm is detected.

CONCLUSIONS
Clinicians are faced with increasingly complex medical problems and varying test sensitivity and specificity that usually are not apparent to those ordering tests. Thus, understanding how to limit false positives without restricting appropriate testing has become a major challenge as well as an important opportunity for improving hospital infection control, infection prevention, and patient safety. As new diagnostic technologies proliferate, key metrics like clinical relevance and cost-effectiveness must be considered before such technologies are incorporated into clinical practice, and systems must be in place for stewardship of each new test before it is introduced into clinical practice. 32 Established testing recommendations (preferably from professional societies or governing medical bodies) are essential to developing a stewardship strategy; however specific, useful consensus guidelines for diagnostic testing for HAIs are often lacking. For instance, no clear consensus exists to guide the use of repeated blood culturing to minimize false-positive rates and maximize true positives, as in patients with repeated fevers and/or patients who are already on antibiotics. 33 Developing meaningful guidelines for diagnostic stewardship for HAIs requires quality evidence from thoughtfully conducted clinical studies. Much work remains to be done to determine the safety and efficacy of limiting providers' autonomy for HAI-related diagnostics. Outcomes and safety-oriented quality improvement research may help bridge the gap between clinical research and practice.
A combined diagnostic and antimicrobial stewardship model could promote better patient evaluations, test choices, interpretations of results, and decisions to prescribe antimicrobial therapy. 32 Expanding on the success of antimicrobial stewardship, diagnostic stewardship should take a multidisciplinary, collaborative approach to existing best practices for HAI prevention.