Abstract
Purpose: Following updates to the Infectious Diseases Society of America (IDSA) practice guidelines for the Diagnosis and Treatment of Adults with Community-acquired Pneumonia in 2019, Hartford HealthCare implemented changes to the community acquired pneumonia (CAP) order-set in August 2020 to reflect criteria for the prescribing of broad-spectrum antimicrobial therapy. The objective of the study was to evaluate changes in broad-spectrum antibiotic days of therapy (DOT) following these order-set updates with accompanying provider education. Methods: This was a multi-center, quasi-experimental, retrospective study of patients with a diagnosis of CAP from September 1, 2019 to October 31, 2019 (pre-intervention) and September 1, 2020 to October 31, 2020 (post-intervention). Patients were identified using ICD-10 codes (A48.1, J10.00-J18.9) indicating lower respiratory tract infection. Data collected included demographics, labs and vitals, radiographic, microbiological, and antibiotic data. The primary outcome was change in broad-spectrum antibiotic DOT, specifically anti-pseudomonal β-lactams and anti-MRSA antibiotics. Secondary outcomes included guideline-concordance of initial antibiotics, utilization of an order-set to prescribe antibiotics, and length of stay (LOS). Results: A total of 331 and 352 patients were included in the pre- and post-intervention cohorts, respectively. There were no differences in order-set usage (10% vs 11.3%, P = .642) between the pre- and post-intervention cohort, respectively. The overall duration of broad-spectrum therapy was a median of 2 days (IQR 0-8 days) in the pre-intervention period and 0 days (IQR 0-4 days) in the post-intervention period (P < .001). Patients in whom the order-set was used in the post-intervention period were more likely to have guideline-concordant regimens ([36/40] 90% vs [190/312] 60.9%; P = .003). Hospital LOS was shorter in the post-intervention cohort (4.8 days [2.9-7.2 days] vs 5.3 days [IQR 3.5-8.5 days], P = .002). Conclusion: Implementation of an updated CAP order-set with accompanying provider education was associated with reduced use of broad-spectrum antibiotics. Opportunities to improve compliance and thus further increase guideline-concordant therapy require investigation.
Keywords: anti-infectives, clinical pathways, education, infectious diseases, decision support systems, information systems and technology, medication process, physician prescribing
Background
Community acquired pneumonia (CAP) is among the most common reasons for emergency department (ED) visits in the United States. The 2017 National Hospital Ambulatory Medical Care Survey reported an average of 1286 pneumonia-related ED visits per 100 000 patients. 1 Historically, the Infectious Diseases Society of America (IDSA) recognized a subset of patients admitted from the community with pneumonia that were deemed to be at risk for multi-drug resistant (MDR) infections, formally referred to as healthcare-associated pneumonia (HCAP). This included patients who were hospitalized for >48 hours, received intravenous (IV) antibiotics within the previous 90 days, had residency in a nursing home or extended care facility, home infusion therapy, chronic dialysis within 30 days, home wound care, or a family member with a MDR pathogen. 2 However, more recent studies suggest that these risk factors may not accurately indicate which patients develop infection with an MDR organism.3-7
The updated 2019 Infectious Diseases Society of America (IDSA) Diagnosis and Treatment of Adults with Community-acquired Pneumonia guidelines formally abandon the prior categorization and recommend empiric coverage for MDR infections (ie, methicillin-resistant Staphylococcus aureus [MRSA] and Pseudomonas aeruginosa) only in adults with validated risk factors, such as prior respiratory isolation of MRSA or P. aeruginosa, or recent hospitalization and parenteral antibiotics within 90 days in patients with severe CAP. 8 Despite the growing evidence that MDR organisms rarely cause CAP, a large percentage of patients admitted with this infection receive broad-spectrum antibiotics, including anti-pseudomonal β-lactams (ie, cefepime, piperacillin/tazobactam, ceftazidime) and anti-MRSA antibiotics (ie, vancomycin, linezolid).9,10
Following guideline updates, the Hartford HealthCare Antimicrobial Stewardship Council revised the pre-existing CAP electronic order-set to reflect these new recommendations. We sought to evaluate the impact of this order set and accompanying education on antibiotic utilization across our system.
Methods
This was a multi-center, quasi-experimental, retrospective study of patients admitted to 5 Hartford HealthCare hospitals with a diagnosis of CAP from September 1, 2019 to October 31, 2019 (pre-intervention period) and September 1, 2020 to October 31, 2020 (post-intervention period). Patients were identified by ICD-10 codes indicating lower respiratory tract infection (A48.1, J10.00-J18.9) and included if ≥18 years old. Patients were excluded if they were admitted from an outside hospital, received active treatment for another infection immediately prior to or during admission, had a diagnosis of hospital-acquired or ventilator-associated pneumonia, had a history of cystic fibrosis, or had a positive SARS-COV-2 PCR during hospital admission. Data collected included baseline demographics, pre-specified labs and vitals on admission, radiographic, microbiological, and antibiotic data.
CAP order-sets existed within the electronic medical record during both the pre- and post-intervention periods. Following updates to the order-set, clinicians were guided through a new clinical decision pathway to select the most appropriate, narrow-spectrum antibiotics based on specific patient characteristics. The order-set prompted users to first stratify patients into a “non-severe CAP” or “severe CAP” pathway based on whether the patient was admitted to or expected to be admitted to the ICU. ICU admission was used as a surrogate for the qualifying severity criteria used in the last 2 guideline iterations. The order-set recommended a combination of ceftriaxone and azithromycin for any patient without the presence of risk factors, including those with suspected aspiration pneumonia. In the non-severe CAP pathway, levofloxacin monotherapy was available as an alternative to ceftriaxone and azithromycin in the event of cephalosporin allergy, and the option to replace azithromycin with doxycycline in combination with ceftriaxone was recommended if there was a contraindication to macrolide or fluoroquinolone therapy. Alternatively, within the severe CAP pathway, the alternative to first line therapy in patients without risk factors was ceftriaxone plus levofloxacin in keeping with the guideline’s avoidance of fluoroquinolone monotherapy or azithromycin replacement with doxycycline due to lack of data in this subset of patients. In order to be eligible for broad-spectrum therapy with anti-MRSA or anti-pseudomonal therapy in non-severe CAP, providers had to acknowledge prior respiratory isolation of the respective pathogen in the selected patient. For severe CAP, such agents could be selected in the presence of the same risk factor or recent hospitalization with IV antibiotics. Respiratory cultures were pre-selected within the order-set and several other diagnostic tools were available for selection, including blood cultures, routine labs, chest X-ray, and medications for symptomatic relief. MRSA nasal PCR was pre-selected when an anti-MRSA agent was ordered, but otherwise selectable as an opt-in choice.
In the weeks prior to implementation of the update, education on the order-set and guideline update was provided to prescribers in the form of live educational sessions by the antimicrobial stewardship team and a newsletter distributed to provider groups, including hospital medicine and emergency medicine. Additional support and recommendations for compliance with guideline recommendations were provided by the antimicrobial stewardship team and pharmacy staff during the course of standard practice throughout the study period. The updated CAP order-set was implemented across Hartford HealthCare hospitals and the now obsolete HCAP order set was removed in August, 2020.
The primary outcome was broad-spectrum antibiotic days of therapy (DOT), which was defined as a β-lactam antibiotic with P. aeruginosa activity or any drug with activity against MRSA. Secondary outcomes included utilization of an order-set to prescribe antibiotics, total inpatient antibiotic DOT, guideline-concordance of initial antibiotics for the treatment of CAP in the post-intervention period, incidence of Clostridioides difficile infection (CDI), length of stay (LOS), and in-hospital mortality. Guideline-concordance of initial antibiotics was assessed via chart review. Normally distributed continuous data were analyzed using a Student’s t-test while continuous data that were not normally distributed were analyzed using a Mann–Whitney U test. For categorical data, a Pearson chi-square test was utilized and results reported as percentages with P-values. All analyzes were conducted using SigmaPlot Version 14 (Systat Software Inc.) using an alpha level of 0.05 such that results yielding P < .05 were considered statistically significant.
Results
A total of 1211 patients were screened for eligibility and 331 and 352 patients met inclusion criteria in the pre- and post-intervention cohorts, respectively. Patient demographics were largely evenly matched between groups (Table 1). Select exceptions include: COPD was more common in the pre-intervention cohort (43.8% vs 34.4%, P = .014), more patients in the pre-intervention cohort received parenteral antibiotics in the previous 90 days (26.6 vs 18.4, P = .011), and more patients in the post-intervention cohort had a positive chest X-ray on admission (76.1% vs 83%, P < .001) (Table 1). A total of 77 and 451 patients were excluded from the study in the pre-and post-intervention cohorts respectively. The primary reason for exclusion across the entire study group was SARS-COV-2 PCR positivity (n = 247), which occurred only in the post-intervention period due to the time frame of the study. Other reasons for study exclusion were treatment of a different infection during or immediately prior to hospital admission (n = 174), admission from an outside hospital (n = 53), and pneumonia diagnosis after 48-hours of admission (n = 41).
Table 1.
Demographics, Labs and Vitals, and Admission Disposition.
| Pre-intervention cohort, n = 331 | Post-intervention cohort, n = 352 | P-value | |
|---|---|---|---|
| Age (years), median (IQR) | 70 (58-80) | 71 (60-82) | .142 |
| Female sex, n (%) | 153 (44.7) | 163 (47.7) | .476 |
| Race, n (%) | |||
| White | 249 (75.2) | 262 (74.4) | .880 |
| Black | 29 (8.7) | 41 (11.6) | .264 |
| Asian | 3 (0.9) | 3 (0.9) | .738 |
| Other | 47 (14.2) | 44 (12.5) | .589 |
| Unknown | 1 (0.3) | 2 (0.6) | .957 |
| Ethnicity, n (%) | |||
| Hispanic | 40 (12.1) | 47 (13.4) | .703 |
| Charlson co-morbidity index, n (%) | |||
| Myocardial infarction | 46 (13.8) | 60 (17) | .303 |
| Heart failure | 127 (38.7) | 140 (39.8) | .766 |
| Peripheral vascular disease | 9 (2.7) | 16 (4.5) | .286 |
| Cerebrovascular attack/trans-ischemic attack | 4 (1.2) | 4 (1.1) | .788 |
| Chronic obstructive pulmonary disease | 145 (43.8) | 121 (34.4) | .014 |
| Connective tissue disease | 5 (1.5) | 3 (0.9) | .658 |
| Peptic ulcer disease | 3 (0.9) | 8 (2.3) | .265 |
| Mild liver disease | 10 (3) | 16 (4.5) | .401 |
| Dementia | 22 (6.6) | 47 (13.4) | .005 |
| Diabetes mellitus | 116 (35) | 117 (33.2) | .677 |
| Hemiplegia | 0 | 3 (0.9) | .269 |
| Chronic kidney disease | 64 (19.3) | 77 (21.9) | .468 |
| Diabetes with end organ dysfunction | 59 (17.8) | 63 (17.9) | .940 |
| Hematologic malignancy, solid tumor | 63 (19) | 39 (11) | .005 |
| Moderate-severe liver disease | 3 (0.9) | 2 (0.6) | .945 |
| Metastatic cancer | 17 (5.1) | 12 (3.4) | .353 |
| AIDS | 1 (0.3) | 4 (1.1) | .407 |
| Total Score, median (IQR) | 2 (1-4) | 2 (1-4) | .470 |
| Vitals and labs upon admission | |||
| WBC, median (IQR) | 12 (8.4-15.3) | 11.4 (8.3-15.6) | .603 |
| Temperature (degrees F), median (IQR) | 98.2 (97.7-99.1) | 98.3 (97.6-98.8) | .36 |
| Heart rate (bpm), median (IQR) | 91 (77-104) | 86 (75-100) | .037 |
| Resp rate (breaths/min), median (IQR) | 20 (18-22) | 18.5 (18-20) | .076 |
| SpO2 (%), median (IQR) | 95 (94-97) | 96 (94-98) | <.001 |
| Procalcitonin, median (IQR) | 0.195 (0.1-0.518) | 0.170 (0.090-0.475) | .698 |
| n (%) | 231 (70) | 247 (70.1) | .972 |
| MRSA PCR positive | 9 (2.7) | 7 (2.0) | .558 |
| n (%) | 62 (18.7) | 72 (20.5) | .638 |
| Site, n (%) | |||
| Hartford hospital | 123 (37.1) | 123 (34.9) | .601 |
| Hospital of central Connecticut | 50 (15.1) | 55 (15.6) | .935 |
| Midstate medical center | 55 (16.2) | 78 (22.2) | .083 |
| Backus hospital | 92 (27.8) | 69 (19.6) | .015 |
| Windham hospital | 11 (3.3) | 27 (7.7) | .021 |
| Admission disposition, n (%) | |||
| ICU | 37 (11.2) | 42 (11.9) | .851 |
| Stepdown unit | 57 (17.2) | 43 (12.2) | .082 |
| Ward | 237 (71.6) | 267 (75.9) | .240 |
| Long term care on admission, n (%) | 30 (9.1) | 41 (11.6) | .327 |
| Parenteral antibiotics within 90 days, n (%) | 88 (26.6) | 65 (18.4) | .011 |
| Chest X-ray positive on admission, n (%) | 252 (76.1) | 292 (83) | <.001 |
Respiratory cultures were collected in 61.8% of patients in the pre-intervention cohort and 23.3% of patients in the post-intervention cohort. Of patients who had respiratory cultures collected, 30.3% (n = 62/204) and 39% (n = 32/82) ultimately isolated a pathogen in the pre- and post-intervention cohorts, respectively (Table 2). The most commonly isolated organisms in both pre- and post-intervention cohorts were MSSA (methicillin-susceptible Staphylococcus aureus), MRSA, and P. aeruginosa. In patients with respiratory cultures collected, the rate of MRSA isolation was 5.9% (n = 12/204) and 6.1% (n = 5/82) and the rate of P. aeruginosa isolation was 6.4% (n = 13/204) and 10.9% (n = 9/82) in the pre- and post-intervention cohorts, respectively. The next most commonly isolated organisms were E. coli (3.9%, n = 8/204) in the pre-intervention group and Streptococcus pneumoniae (6.1%, n = 5/82) in the post-intervention group (Table 3). In patients who had previously isolated P. aeruginosa from a respiratory culture (n = 13) within 1 year, 1 patient had a respiratory culture positive for P. aeruginosa, 6 patients were culture negative, and 5 patients did not have respiratory cultures collected. In patients who had previously isolated MRSA within 1 year (n = 11), 1 patient had a respiratory culture positive for MRSA, 3 patients were culture negative, and 6 patients did not have respiratory cultures collected.
Table 2.
Results of Microbiologic Testing.
| Pre-intervention cohort, n = 331 | Post-intervention cohort, n = 352 | P-value | |
|---|---|---|---|
| Prior P. aeruginosa isolated within 1 year, n (%) | 12 (3.6) | 1 (0.3) | .004 |
| Prior MRSA isolated within 1 year, n (%) | 10 (3) | 1 (0.3) | .011 |
| Respiratory culture collected, n (%) | 204 (61.8) | 82 (23.3) | <.001 |
| Any pathogen isolated, n (%) | 62/204 (30.3) | 32/82 (39) | .215 |
Table 3.
Most Common Causative Pathogens.
| Pre-intervention cohort, n = 204 | Post-intervention cohort, n = 82 | P-value | |
|---|---|---|---|
| P. aeruginosa, n (%) | 13 (6.4) | 9 (10.9) | .186 |
| MRSA, n (%) | 12 (5.9) | 5 (6.1) | .945 |
| MSSA, n (%) | 8 (3.9) | 9 (10.9) | .023 |
| Escherichia coli, n (%) | 8 (3.9) | 3 (3.7) | .011 |
| Streptococcus pneumoniae, n (%) | 1 (0.5) | 5 (6.1) | .003 |
There were no differences in order-set usage (33 [10%] vs 40 [11.3%], P = .642) between groups. Overall duration of broad-spectrum therapy was a median of 2 days (IQR 0-8 days) in the pre-intervention period and 0 days (IQR 0-4 days) in the post-intervention period (P < .001). Fewer patients received anti-MRSA therapy (30.1% vs 44.7%; P < .001) and anti-pseudomonal β-lactam therapy (27% vs 46.2%; P < .001) in the post-intervention cohort compared with the pre-intervention cohort, respectively. Following order-set implementation and provider education, there was increased use of ceftriaxone (66.2% vs 49.8%; P < .001) and azithromycin (55.4% vs 42.9%; P = .001) and decreased use of vancomycin (28.4% vs 39.2%; P = .003), cefepime (22.2% vs 38.1%; P < .001), and piperacillin/tazobactam (1.1% vs 5.1%; P = .005). There were no significant differences in prescribing of any other individual antibiotics during the study period, including doxycycline and levofloxacin (Table 4).
Table 4.
Prescribing and Outcomes Data.
| Pre-intervention cohort, n = 331 | Post-intervention cohort, n = 352 | P-value | |
|---|---|---|---|
| Duration of all broad-spectrum therapy, days (median, IQR) | 2 (0-8) | 0 (0-4) | <.001 |
| Total duration of anti-MRSA therapy, days (median, IQR) | 0 (0-3) | 0 (0-2) | <.001 |
| Received MRSA antibiotic, n (%) | 148 (44.7) | 106 (30.1) | <.001 |
| Total duration of anti-pseudomonal therapy, days (median, IQR) | 0 (0-4) | 0 (0-2) | <.001 |
| Received anti-pseudomonal β-lactam, n (%) | 153 (46.2) | 95 (27) | <.001 |
| C. diff infection, n (%) | 5 (1.5) | 1 (0.3) | .191 |
| CAP order-set use, n (%) | 33 (10) | 40 (11.3) | .642 |
| Hospital length of stay (days), median (IQR) | 5.3 (3.5-8.5) | 4.8 (2.9-7.2) | .002 |
| 30-day re-admission rate, n (%) | 77 (23.3) | 79 (22.4) | .853 |
| All-cause mortality, n (%) | 19 (5.7) | 19 (5.3) | .978 |
| Total antibiotic DOT (days), median (IQR) | 5 (4-7) | 5 (3-7) | .003 |
| Appropriate use per 2019 IDSA Guidelines, n (%) | - | 226 (64.2) | - |
| Individual Antibiotic Prescribing, n (%) | |||
| Ceftriaxone | 165 (49.8) | 233 (66.2) | <.001 |
| Azithromycin | 142 (42.9) | 195 (55.4) | .001 |
| Doxycycline | 37 (11.2) | 31 (8.8) | .365 |
| Levofloxacin | 16 (4.8) | 17 (4.8) | .860 |
| Vancomycin | 131 (39.6) | 100 (28.4) | .003 |
| Cefepime | 126 (38.1) | 78 (22.2) | <.001 |
| Piperacillin-tazobactam | 17 (5.1) | 4 (1.1) | .005 |
| Ceftazidime | 2 (0.6) | 1 (0.3) | .957 |
The overall rate of guideline-concordant prescribing in the post-intervention group was 64.2% (n = 226/352). Patients in whom the order-set was used in the post-intervention period were more likely to have guideline-concordant regimens ([36/40] 90% vs [190/312] 60.9%; P = .003). Total inpatient antibiotic DOT was a median of 5 days in both groups, however was shorter in the post-intervention group as evidenced by an IQR of 3 to 7 days compared with 4 to 7 days in the pre-intervention group (P = .003). Hospital LOS was shorter in the intervention cohort (4.8 days [IQR 2.9-7.2 days] vs 5.3 days [IQR 3.5-8.5 days]; P = .002). There were no differences in rates of CDI (5 [1.5%] vs 1 [0.3%]; P = .191), or in-hospital mortality (19 [5.7%] vs 19 [5.3%]; P = 0.191) between the pre- and post-intervention cohorts, respectively (Table 5).
Table 5.
Guideline Concordant Prescribing With Order-Set Use.
| Did not use order-set, n = 312 | Used order-set, n = 40 | P-value | |
|---|---|---|---|
| Initial regimen concordant with 2019 CAP guidelines a , n (%) | 190 (60.9) | 36 (90) | .003 |
Post-intervention period only.
Discussion
In this retrospective, quasi-experimental, multi-center study of patients with CAP, prescriber education surrounding order-set updates following IDSA CAP guideline updates improved broad-spectrum antibiotic usage across the Hartford HealthCare system. It is likely that the reduction in broad-spectrum DOT was a result of improved guideline-concordant prescribing after the intervention, given that the primary focus of the education and order-set was adherence to guideline recommendations. Provider education has proven to be effective in improving guideline-concordant prescribing. In one retrospective analysis, a 1-day educational session on management of pediatric pneumonia significantly improved the selection of guideline-concordant antibiotics in hospitalized patients and antibiotic duration in non-hospitalized patients. 11 While no other studies have assessed broad-spectrum DOT as a result of interventions involving order-sets and education for CAP, such strategies have consistently been shown to improve empiric antibiotic prescribing in common infectious diseases.12,13
CAP order-set usage was low across the study population with 9% and 11.3% usage for antibiotic prescribing in the pre- and post-intervention cohorts, respectively. The low utilization rate is likely due to several factors. Firstly, due to the nature of CAP, initial antibiotics are typically ordered while patients are in the ED as 1-time orders. When patients are later admitted, admitting providers may be more likely to continue antibiotic orders using individual “à la carte” orders rather than seek out an order-set to re-order. Additionally, the presence of medical residents at Hartford HealthCare sites introduces a challenge for consistent utilization of order-sets. These trainees rotate through various health systems in the geographic region, thus it may be difficult to expect compliance with each system’s clinical pathways. It is unknown how much antibiotic ordering within the study was done by trainees versus permanent clinicians, however, this is a challenge that may be encountered at other academic medical centers where trainees rotate through multiple hospitals.
Overall, the rate of isolation of a pathogen in respiratory culture was low in our study. Only 41.8% (n = 286/683) of patients in the study had a respiratory culture collected, and of those, 32.8% (n = 94/286) grew a microorganism in culture. Although MDR organisms were demonstrated to be some of the most common pathogens isolated, the overall rate was low (5%-11%) and was consistent with prior studies of this disease state. In comparison, 1 retrospective study of 531 patients comparing HCAP and CAP reported an overall incidence of MRSA of 11% and P. aeruginosa of 9% in patients with respiratory cultures. 4 In a separate retrospective study comprised of 2167 CAP patients, the incidence of MDR infections was 3% of all patients included (not just those with cultures), though pathogens were isolated in only 14% of patients. 10 It is likely that the rate of MRSA and P. aeruginosa seen in our study is reflective of the inclusion of patients previously considered to have HCAP into the new CAP definitions. In our analysis, the rate of anti-MRSA and anti-pseudomonal β-lactam prescribing was 30.1% and 27% in the post-intervention cohort respectively, which far exceeded the rate of isolation of MRSA or P. aeruginosa.
Current IDSA guidelines consider prior isolation of an MDR organism within 1 year and hospitalization with parenteral antibiotics within 90 days to be risk factors for MDR organisms in CAP. In our analysis, few patients had prior isolation of MDR pathogens documented in our electronic medical record within the preceding year. The rate of prior isolation of P. aeruginosa from a respiratory culture was 3.6% and 0.3% and the rate of MRSA isolation was 3% and 0.3% in the pre- and post-intervention groups, respectively. In this study, 21% of patients received parenteral antibiotics in the previous 90 days, indicating a substantial portion of patients with prior healthcare exposures overall. Additionally, patients presenting from a long-term care facility were also included this study, which was previously considered a risk factor for MDR HCAP, but only 10.4% of patients came from such a facility. Of patients who ultimately isolated P. aeruginosa during the study (n = 22), only 2 met criteria for receipt of anti-pseudomonal β-lactam on admission based on risk factors. Alternatively, of patients who ultimately isolated MRSA during the study (n = 17), only 1 met criteria for receipt of anti-MRSA antibiotic on admission based on risk factors. This data suggests that despite updates to the risk criteria in the guidelines, their specificity remains poor. Further investigation is warranted to identify more specific risk factors for MDR organisms in patients with CAP within this health system, as suggested by the 2019 IDSA guidelines.
These data suggest that an order-set, when used for prescribing of antibiotics for CAP, improves prescribing practices as the rate of concordance with IDSA guidelines in the post-intervention period was 90% when the CAP order-set was used and 60.9% when it was not. This is consistent with Seitz et al, 14 who demonstrated that use of order-sets for common infectious diseases in the ED was low (21%), but significantly associated with appropriate empiric antibiotic selection (86.4% vs 33.8%; p < .001). It is likely that with increased utilization of the order-set, overall guideline concordance of antibiotic regimens would continue to improve. Exploration of the current barriers and continued education surrounding the potential benefits of order-set use is needed. One strategy that has been suggested to increase use of order-sets in the ED is insertion of guideline-concordant antibiotics for common infections into other more widely used admission order sets. 13 A potential limitation to this strategy may be the inability to include other valuable components to aid in clinical decision-making, such as pre-selected orders for respiratory cultures and MRSA nares swabs within larger order panels.
Decreased hospital length of stay to 4.8 days from 5.3 days (P = .002) in the post-intervention group was a somewhat unexpected finding given that the intervention of an updated order-set and accompanying prescriber education did not explicitly include suggestions for earlier discharge. Shorter LOS has been demonstrated previously, however, with decreased use of broad-spectrum agents. 10 One contributing factor may have been the introduction of the Coronavirus Disease 19 (COVID-19) pandemic between the pre-and post-intervention periods. The time frame of the post-intervention period coincided with the beginning of a surge in cases in Connecticut, where the study was conducted. 15 An increase in hospital admissions and strain on resources may explain an increase in effort for bed turnover and expedited discharge. Notably, nearly twice the number of patients were screened for eligibility in the post-intervention period (803 vs 408) due to identification of lower respiratory tract infection by ICD-10 code during the specified time frame, but 247 patients were ultimately excluded due to SARS-COV-2 PCR positivity in the post-intervention period.
A number of limitations may have impacted the outcomes of this study. Firstly, the time frame of the pre-intervention period surrounded the release of the 2019 CAP guidelines (October 1, 2019). Consequently, a significant portion of patients in the pre-intervention period could not be assessed for guideline-compliance as they were under the purview of a previous iteration of guidelines. This limitation may also have contributed to the findings of a reduction in broad-spectrum DOT if more patients qualified for broad-spectrum therapy during this timeframe underneath the previous guideline recommendations. Assessment of the efficacy of the education piece of the intervention was significantly limited by several factors. In addition to the pre-intervention time frame prohibiting assessment of baseline of guideline-concordant prescribing (as detailed above), no effort was made to track prescribing of the providers who attended the educational sessions. Additionally, given the widespread distribution of the educational newsletter, it was not possible to track which providers acknowledged and read the newsletter and thus, we were unable to compare groups with and without this piece to quantify efficacy. Separate from methodological limitations, the occurrence of the COVID-19 pandemic during the post-intervention period may have impacted a number of factors. Since COVID-19 is a respiratory illness, this may have confounded the initial identification of CAP patients by treating providers during the post-intervention period. During the study period, however, there were no significant restrictions on SARS-COV-2 PCR tests, making identification and diagnosis fairly rapid (within 24 hours). Additionally, the COVID-19 pandemic may have impacted the types of patients who sought medical care due to concerns with healthcare contact. There were less patients who received parenteral antibiotics within 90 days in the post-intervention group (18.4% vs 26.6%, P = .011), demonstrating a lower rate of prior healthcare exposure in this group. Despite this, baseline characteristics among the patients were well-balanced between the groups with regard to age, gender, and comorbidities, with differences seen only in rates of underlying COPD, dementia, and malignancy. Other limitations that may have impacted the outcomes of this study include retrospective design, which is prone to bias and the inability to collect data such as prior microbiologic and admissions data from other health systems outside of Hartford HealthCare.
Despite low order-set utilization, when used, the updated order-set was associated with improved guideline-concordant prescribing. While the rate of broad-spectrum antibiotic prescribing was significantly reduced, further improvements could still be made. Future efforts should be made to identify barriers to increased order-set usage and identify locally-validated risk factors for MDR organisms in CAP patients to aid in patient risk stratification within this health system. Antimicrobial stewardship programs should prioritize order-set updates and accompanying education with the release of new guidelines to keep providers informed and improve antibiotic prescribing.
Acknowledgments
Ying Han, PharmD Candidate, University of Connecticut School of Pharmacy, Storrs, CT.
Footnotes
Declaration of Conflicting Interests: The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding: The author(s) received no financial support for the research, authorship, and/or publication of this article.
ORCID iD: Jessica L. Colmerauer, 
https://orcid.org/0000-0001-8931-0512
References
- 1. Rui P, Kang K. National Hospital ambulatory medical care survey: 2017 emergency department summary tables. National Center for Health Statistics. CDC. Accessed February 28, 2021. https://www.cdc.gov/nchs/data/nhamcs/web_tables/2017_ed_web_tables-508.pdf [Google Scholar]
- 2. American Thoracic Society; Infectious Diseases Society of America. Guidelines for the management of adults with hospital-acquired, ventilator-associated, and healthcare-associated pneumonia. Am J Respir Crit Care Med. 2005;171(4):388-416. doi: 10.1164/rccm.200405-644ST [DOI] [PubMed] [Google Scholar]
- 3. Chalmers JD, Rother C, Salih W, Ewig S. Healthcare-associated pneumonia does not accurately identify potentially resistant pathogens: a systematic review and meta-analysis. Clin Infect Dis. 2014;58:330-339. doi: 10.1093/cid/cit734 [DOI] [PubMed] [Google Scholar]
- 4. Gross AE, Van Schooneveld TC, Olsen KM, et al. Epidemiology and predictors of multidrug-resistant community-acquired and health care-associated pneumonia. Antimicrob Agents Chemother. 2014;58:5262-5268. doi: 10.1128/AAC.02582-14 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Zilberberg MD, Shorr AF. Healthcare-associated pneumonia: the state of evidence to date. Curr Opin Pulm Med. 2011;17:142-147. doi: 10.1097/MCP.0b013e328343eb33 [DOI] [PubMed] [Google Scholar]
- 6. Jones BE, Jones MM, Huttner B, et al. Trends in antibiotic use and nosocomial pathogens in hospitalized veterans with pneumonia at 128 medical centers, 2006-2010. Clin Infect Dis. 2015;61:1403-1410. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Vallés J, Martin-Loeches I, Torres A, et al. Epidemiology, antibiotic therapy and clinical outcomes of healthcare-associated pneumonia in critically ill patients: a Spanish cohort study. Intensive Care Med. 2014;40:572-581. doi: 10.1007/s00134-014-3239-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Metlay JP, Waterer GW, Long AC, et al. Diagnosis and treatment of adults with community-acquired pneumonia. An official clinical practice guideline of the american thoracic society and infectious diseases society of america. Am J Respir Crit Care Med. 2019;200(7):e45-e67. doi: 10.1164/rccm.201908-1581ST [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Jones BE, Ying J, Stevens V, et al. Empirical anti-MRSA vs standard antibiotic therapy and risk of 30-day mortality in patients hospitalized for pneumonia. JAMA Intern Med. 2020;180(4):552-560. doi: 10.1001/jamainternmed.2019.7495 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Webb BJ, Sorensen J, Jephson A, Mecham I, Dean NC. Broad-spectrum antibiotic use and poor outcomes in community-onset pneumonia: a cohort study. Eur Respir J. 2019;54(1):1900057. doi: 10.1183/13993003.00057-2019 [DOI] [PubMed] [Google Scholar]
- 11. Di Pietro P, Della Casa Alberighi O, Silvestri M, et al. Monitoring adherence to guidelines of antibiotic use in pediatric pneumonia: the Marea study. Ital J Pediatr. 2017;43:113. doi: 10.1186/s13052-017-0432-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Fowkes CT, Gee C, Bluemink T, Cole D, Falkner BL, Hamour AA. Audit of physicians’ adherence to a preprinted order set for community-acquired pneumonia. Can J Hosp Pharm. 2010;63(4):289-294. doi: 10.4212/cjhp.v63i4.932 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Chan A, Kapur A, Langford B, Downing M. 1046. Evaluating the outcomes of embedding antimicrobial stewardship order sets in the general medicine admission electronic order set: a retrospective study. Open Forum Infect Dis. 2019;6(Suppl 2):S368-S369. doi: 10.1093/ofid/ofz360.910 [DOI] [Google Scholar]
- 14. Seitz RM, Wiley Z, Francois CF, Moran TP, Rupp JD, Sexton ME. Improved empiric antibiotic prescribing for common infectious disease diagnoses using order sets with built-in clinical decision support in the emergency department. Infect Control Hosp Epidemiol. Published online March 9 2021. doi: 10.1017/ice.2021.73 [DOI] [PubMed] [Google Scholar]
- 15. Connecticut Daily COVID-19 Update. Connecticut open data. Accessed April 26, 2021. https://data.ct.gov/stories/s/q5as-kyim