A Deeper Dive Into Antibiotic Stewardship Needs: A Multihospital Survey.

In a 2016 survey of 46 Michigan hospitals, we identified four key needs for antibiotic stewardship: clinically-relevant antibiotic data, monitoring compliance, syndrome-specific interventions, and discharge stewardship. A stewardship initiative now addresses these needs within the Michigan Hospital Medicine Safety Consortium.

Antibiotic stewardship programs improve outcomes by reducing adverse events, such as Clostridioides difficile infection and antibiotic resistance. Thus, The Joint Commission (TJC) began requiring that US hospitals have stewardship programs for accreditation in 2017 [1]. Because TJC focused on the Centers for Disease Control and Prevention’s (CDC) core elements for antibiotic stewardship, the requirement resulted in a substantial increase in their use [2]. Because multiple aspects of stewardship can improve care, the core elements allow flexibility in how hospitals implement stewardship [3]. This flexibility has allowed most hospitals to meet core elements and TJC requirements [3]. Regardless, many small, nonacademic hospitals have limited ability to execute the more resource-intensive strategies that may be more impactful elements of stewardship [2, 3].

One way for hospitals to overcome barriers to stewardship implementation is by joining or establishing regional collaborative quality initiatives. These regional organizations allow hospitals to pool resources, benchmark data, and share successful strategies [4-7]. The Michigan Hospital Medicine Safety Consortium (HMS) is one such statewide initiative. Previously, HMS has improved care related to venous thromboembolism and peripherally inserted central catheters [8, 9]. However, it was unclear what gaps in hospital-based stewardship could be fulfilled through such a statewide collaborative given the diversity of hospitals (eg, multiple payers, owners, etc) and existing stewardship interventions. Furthermore, HMS offered an ideal platform to assess statewide variation in antibiotic stewardship practices given this history of collaboration and engagement in patient safety. Thus, we surveyed the 46 hospitals participating in HMS immediately before the start of TJC’s stewardship requirement to assess current practices and needs related to antibiotic stewardship.

METHODS

Data Collection

Between September 30, 2016 and November 7, 2016, surveys were e-mailed to 46 hospitals voluntarily participating in HMS, which includes 50% of the noncritical access, nonfederal hospitals in Michigan. The data abstractor (typically a nurse in quality) at each hospital completed the survey and was responsible for working with local individuals to ensure accuracy.

Survey Measures and Data Analysis

The survey assessed antibiotic stewardship program characteristics including those related to the 2014 CDC core elements of “Action,” “Tracking,” “Reporting,” and “Education” (survey in Appendix) [10]. Descriptive statistics are reported for responses. Because small hospitals may have limited capacity to individualize interventions, we compared presence of facility-specific treatment recommendations and syndrome-specific interventions between hospitals with ≤200 vs >200 beds, using 2-sided Fisher’s exact, χ 2, or t tests, as appropriate (StataSE version 14).

RESULTS

Survey response rate was 100% (46 of 46 hospitals). With the exception of 2 small community hospitals, all hospitals reported having an antibiotic stewardship program (96%, 44 of 46). Most stewardship programs reported collaborating with CDC-defined “key support” groups, although partnering with information technology (67%, 31 of 46) and nursing (37%, 17 of 46) was less common (Table 1). All programs used 1 or more CDC-recommended Action elements, including most broad stewardship interventions; antibiotic timeouts were an exception with only 13% (6 of 46) of hospitals reporting their use. Although 80% of hospitals reported having a policy related to antibiotic documentation, only 17% (8 of 46) required documentation of intended duration and only half (52%, 24 of 46) reported monitoring compliance with documentation. It is interesting to note that only 17% (8 of 46) of hospitals had a process to review outpatient antibiotic orders before discharge.

Table 1.

Antibiotic Stewardship Program Interventions and Policies in the Michigan Hospital Medicine Safety Consortium (N = 46 Hospitals)

Question	Hospitals, N (%)
Stewardship Collaboration (“Key Support” Groups Under “Accountability” CDC Core Element)
Infection Prevention and Healthcare Epidemiology	41 (89%)
Microbiology (Laboratory)	39 (85%)
Infectious Diseases Physician	37 (80%)
NonInfectious Diseases Physician	36 (78%)
Quality Improvement	35 (76%)
Information Technology	31 (67%)
Nursing	17 (37%)
Stewardship Interventions (“Action” CDC Core Element)	46 (100 %)
Policies that Support Optimal Antibiotic Use
Any Policy Related to Antibiotic Documentation	37 (80%)
Require Documentation of Antibiotic Dose	36 (78%)
Require Documentation of Antibiotic Indication	21 (46%)
Require Documentation of Intended Antibiotic Duration	8 (17%)
Broad Stewardship Interventions
Restricted Formulary for Some Antibioticsa	42 (91%)
Prospective Audit and Feedback for Some Antibiotics	41 (89%)
Preprescription Approval (Prior Authorization) for Some Antibiotics	36 (78%)
“Antibiotic Timeout” to Review Antibiotics at 48–72 hours	6 (13%)
Pharmacy (or Electronic Medical Record) Interventions
Antibiotic Dose Adjustments for Organ Dysfunction	45 (98%)
Automatic Changes from Intravenous to Oral in Certain Situations	37 (80%)
Antibiotic Dose Optimization for Organisms with Reduced Susceptibility	34 (74%)
Alerts for Unnecessarily Duplicative Therapy	31 (67%)
Time-Sensitive Automatic Stop Orders	24 (52%)
Facility-Specific Treatment Recommendation for Any Infection	45 (98%)
Any Syndrome-Specific Intervention	32 (70%)
Other Stewardship Interventionsa
Order Sets that Incorporate Facility-Specific Treatment Recommendations	42 (91%)
Order Sets Contain Decision-Support	28 (61%)
Review of Outpatient Antibiotic Orders Before Discharge	8 (17%)
Antibiotic Monitoring (“Tracking” CDC Core Element)	46 (100%)
Monitor Facility-Wide Antibiotic Prescribing	38 (83%)
By Direct Expenditure (purchasing costs)	30 (65%)
By Count of Antibiotic Administered (day of therapy)	19 (41%)
By Individual Antibiotics	16 (35%)
By Unit Level	6 (13%)
By Service level (eg, hospitalists, surgeons)	7 (15%)
By Diagnosis (eg, diagnosis code)	3 (6%)
By Grams of Antibiotics Used (defined daily dose)	12 (26%)
Submit to CDC’s National Healthcare Safety Network Antimicrobial Use and Resistance Module	8 (17%)
Antibiotic Use Process Measures
Monitor Compliance with their Antibiotic Documentation Policy	24 (52%)
Monitor Adherence to Facility-Specific Treatment Recommendations	18 (39%)
Outcome Measures
Produce a Hospital-Specific Antibiogram	43 (94%)
Unit-Specific Antibiogram	18 (40%)
Antibiotic Feedback (“Reporting” CDC Core Element)	44 (96%)
Distribute Antibiogram to Prescribers	40 (87%)
Share Facility-Specific Reports on Antibiotic Use With Providers	31 (67%)
Provide Providers Personalized Communication on Improving Antibiotic Use	23 (50%)
“Education” CDC Core Element	34 (74%)
Provide Education to Clinicians on Antibiotic Prescribing	34 (74%)
Microbiology/Laboratory Interventionsa
Review of (Any) Cultures to Ensure Optimal Antibiotic Use	41 (89%)
Blood Cultures	41 (89%)
Urine Cultures	34 (74%)
Wound Cultures	33 (72%)
Sputum Cultures	32 (70%)
Microbiology Results Provide Comments to Guide Therapy	36 (78%)
Microbiology Lab Uses Rapid Diagnostic Tests	31 (67%)
MALDI-TOF	17 (37%)
Verigene or BioFire	9 (20%)
PNA FISH	8 (17%)
Other	7 (15%)
Review of Rapid Diagnostic Tests to Ensure Optimal Antibiotic Use	16 (35%)

Abbreviations: CDC, Centers for Disease Control and Prevention; MALDI-TOF, matrix-assisted laser desorption ionization time-of-flight mass spectrometry; PNA-FISH, peptide nucleic acid fluorescence in situ hybridization.

aNot specifically delineated as a “core element” in 2014 CDC Core Elements.

Although most hospitals monitored facility-wide antibiotic use (83%, 38 of 46), few were able to do so by individual antibiotic, unit, service, or diagnosis, and sharing of data with prescribers was limited (Table 1). Two thirds (67%, 31 of 46) of hospitals used rapid diagnostic tests, but only 35% (16 of 46) reviewed test results to optimize antibiotic use.

Although 98% of hospitals reported facility-specific recommendations to treat specific syndromes, only 39% monitored adherence with those recommendations. Although most hospitals (70%) had at least 1 syndrome-specific intervention to optimize antibiotic use, syndromes targeted varied with the most common syndrome (pneumonia) targeted in only 54% of hospitals (eTable 1). Compared with larger hospitals (>200 beds), smaller hospitals (≤200 beds) were less likely to have facility-specific treatment recommendations for urinary tract infection or skin and soft tissue infection (eTable 1).

DISCUSSION

In a survey of 46 Michigan hospitals immediately before TJC stewardship requirement, we identified 4 areas of need for antibiotic stewardship, including the following: (1) clinically relevant antibiotic use data, (2) tracking and monitoring compliance, (3) syndrome-specific interventions, and (4) discharge stewardship.

Consistent with prior studies, we found that most hospitals monitored facility-wide antibiotic use [11]. However, few hospitals were able to monitor use by antibiotic type, service, unit, or diagnosis. Nationally, the CDC is attempting to provide antibiotic use data through the National Healthcare Safety Network’s (NHSN) antimicrobial use and resistance module. However, we found fewer than 1 in 5 hospitals were contributing data to NHSN, lower than in other studies [4, 11]. Although NHSN provides unit-specific data, there remains a missing gap related to service and syndrome-specific data, which we found most hospitals were unable to obtain on their own. Syndrome-specific data can be quite powerful—for example, a statewide collaborative in Colorado improved antibiotic use by providing hospitals with data on their treatment of urinary tract infection and skin and soft tissue infection [7]. A second need involved monitoring process measures related to stewardship. Although almost all hospitals met the CDC’s 2014 Action core element (ie, “implementing at least one recommended [stewardship] action”), a deeper dive found that adherence to interventions, policies, and recommendations were rarely monitored. Limited resources and the noted lack of collaboration with information technology may contribute. Because both process and outcome data are necessary to improve outcomes, limited data could severely impact the ability of programs to change prescribing.

Third, hospitals had limited use of syndrome-specific interventions, despite evidence that they are more effective at engaging prescribers, and more sustainable, than broad stewardship interventions [12]. Pneumonia and urinary tract infection account for approximately half of all inpatient antibiotic use; however, only approximately half of the hospitals had interventions in place for these conditions. It is notable that hospitals with ≤200 beds were less likely to report having recommendations for treating urinary tract infection or skin and soft tissue infection.

Finally, we found that few hospitals reviewed antibiotics before discharge. Likewise, in 2015, the Veterans Health Administration (VHA) found that 24% of VHA hospitals reported “never” reviewing antibiotics at discharge and only 5% evaluated appropriate antibiotic duration at discharge [13]. Although medication reconciliation at discharge has long been an established element of hospital care, antibiotic stewardship at discharge is a relatively novel area—one that is now recommended in both the CDC’s recently updated 2019 core elements (within Action and Tracking) [14] and by TJC [3]. Because antibiotics prescribed at discharge account for up to half of antibiotic use related to hospitalization for common infections, interventions at discharge (or lack thereof) could have broad implications for stewardship and patient outcomes.

Our study has limitations. We relied on self-reporting and were limited to a single state. However, HMS represents a group of diverse hospitals across multiple payers, sizes, and locations. We were also limited by survey methodology on program details, including additional information on program infrastructure (eg, full time equivalent [FTE] support). Although the survey took place immediately before TJC requirements took effect, hospitals may have subsequently increased stewardship efforts—the CDC has found increased use of their core elements over time [2]. Finally, further research is needed to understand whether addressing the 4 needs we identified may improve antibiotic use and outcomes. Study strengths include an outstanding response rate with no missing data. Also, we provide additional detail related to adherence monitoring, microbiology, and discharge interventions that builds on prior national surveys.

The results from our study already have implications in Michigan that may be applicable to hospitals—especially those with limited resources—nationwide. Specifically, the results of this survey helped inform an antimicrobial stewardship initiative across HMS hospitals beginning in 2017. First, through HMS, we collect clinically relevant data on a sample of patients hospitalized at member hospitals with the 2 most frequent drivers of inpatient antibiotic use, pneumonia, or urinary tract infection. Data collection includes process (antibiotic) and outcome data. Second, data are used to help hospitals track and monitor compliance with guidelines related to antibiotic use. With this detailed data collection process, we are able to assess antibiotic appropriateness and provide clinically relevant metrics not previously available to HMS hospitals. This, for example, has enabled feedback to individual providers. Third, because approximately half of hospitals lacked syndrome-specific interventions, we provide syndrome-specific guidelines and suggest syndrome-specific interventions via quarterly in-person meetings and an online toolkit (https://mi-hms.org/resources/hms-quality-initiative-toolkits/hms-antimicrobial-toolkit). Finally, we provide a mechanism for monitoring, tracking, and improving antibiotic use at discharge. The results of this multiyear effort are underway, but preliminary results are promising [15] and may inform similar stewardship initiatives elsewhere. Specifically, other initiatives should focus on providing concrete, actionable, and benchmarked hospital data.

CONCLUSIONS

In summary, we provide a picture of antibiotic stewardship practices in Michigan hospitals immediately before TJC standard for stewardship. Although almost all hospitals had an antibiotic stewardship program, we identified 4 key areas of need: (1) clinically relevant antibiotic data, (2) tracking and monitoring compliance, (3) syndrome-specific interventions, and (4) antibiotic use at hospital discharge. Quality collaboratives may play a key role in helping hospitals—especially those with limited resources—address these needs to optimize antibiotic use.

Supplementary Data

Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

eTable 1. Syndrome Specific Stewardship Interventions (N = 46 Hospitals).

Click here for additional data file.

	Yes	No
Infectious Disease Clinician (without Antimicrobial Stewardship Team (AST) support)	○	○
Clinicians (NonInfectious Disease)	○	○
Infection Prevention and Healthcare Epidemiology	○	○
Quality Improvement	○	○
Microbiology (Laboratory)	○	○
Information Technology (IT)	○	○
Nursing	○	○
Other	○	○

	Yes	No
Dose	○	○
Intended Duration of Use	○	○
Indication	○	○

	Yes	No
Automatic changes from intravenous to oral antibiotic therapy in appropriate situations	○	○
Dose adjustments in cases of organ dysfunction	○	○
Dose optimization to optimize the treatment of organisms with reduced susceptibility	○	○
Automatic alerts in situations where therapy might be unnecessarily duplicative	○	○
Time-sensitive automatic stop orders for specified prescriptions	○	○
Other	○	○