Hospital-Wide Protocol Significantly Improved Appropriate Management of Patients with Staphylococcus aureus Bloodstream Infection

Background:Staphylococcus aureus bloodstream infection (SA-BSI) causes morbidity and mortality. We established a management protocol for patients with SA-BSI aimed at improving quality of care and patient outcomes. Methods: A retrospective pre–post intervention study was conducted at Maharaj Nakorn Chiang Mai Hospital from 1 October 2019 to 30 September 2020 in the pre-intervention period and from 1 November 2020 to 31 October 2021 in the post-intervention period. Results: Of the 169 patients enrolled, 88 were in the pre-intervention and 81 were in the post-intervention periods. There were similar demographic characteristics between the two periods. In the post-intervention period, evaluations for metastatic infections were performed more frequently, e.g., echocardiography (70.5% vs. 91.4%, p = 0.001). The appropriateness of antibiotic prescription was higher in the post-intervention period (42% vs. 81.5%, p < 0.001). The factors associated with the appropriateness of antibiotic prescription were ID consultation (OR 15.5; 95% CI = 5.9–40.8, p < 0.001), being in the post-intervention period (OR 9.4; 95% CI: 3.5–25.1, p < 0.001), and thorough investigations for metastatic infection foci (OR 7.2; 95% CI 2.1–25.2, p = 0.002). However, the 90-day mortality was not different (34.1% and 27.2% in the pre- and post-intervention periods, respectively). The factors associated with mortality from the multivariate analysis were the presence of alteration of consciousness (OR 11.24; 95% CI: 3.96–31.92, p < 0.001), having a malignancy (OR 6.64; 95% CI: 1.83–24.00, p = 0.004), hypoalbuminemia (OR 5.23; 95% CI: 1.71–16.02, p = 0.004), and having a respiratory tract infection (OR 5.07; 95% CI: 1.53–16.84, p = 0.008). Source control was the only factor that reduced the risk of death (OR 0.08; 95% CI: 0.01–0.53, p = 0.009). Conclusion: One-third of patients died. Hospital-wide protocol implementation significantly improved the quality of care. However, the mortality rate did not decrease.

However, studies have been performed in different hospital settings, and the treatment protocol showed some variations. ID specialists must be involved in the development of treatment protocols. However, ID consultations were not always mandated in the protocols 2 of 12 of some studies [6][7][8][9][10]. The value of adherence to treatment protocols, ID consultations, or a combination of both is difficult to determine.
Maharaj Nakorn Chiang Mai Hospital, an affiliated hospital of Chiang Mai University, is a 1400-bed tertiary care referral center in Northern Thailand. There have been 80-90 patients with SA-BSI each year since 2013 (data from the Diagnostic Laboratory, Maharaj Nakorn Chiang Mai Hospital, Chiang Mai, Thailand). The primary care team includes residents and attending physicians who are responsible for patient management and for consulting specialists if indicated. For example, with SA-BSI, the primary care team sometimes manages patients by themselves without ID consultation. To improve patient care, we therefore developed a hospital-wide protocol for the management of patients with SA-BSI.
The primary objective of this study was to determine the 90-day mortality rate of patients who had SA-BSI. The secondary objectives were to determine (1) the recurrent rate of SA-BSI; (2) the rate of appropriate management of SA-BSI, e.g., the rate of investigations for metastatic infections, the rate of follow-up blood culture at 72 h after receiving appropriate antibiotics; (3) the rate of appropriate of antibiotic prescription and factors associated with appropriateness; (4) risk factors for death.

Demographic Characteristics
One-hundred and sixty-nine patients who had SA-BSI and met the inclusion criteria were enrolled (Figure 1). The pre-intervention and post-intervention period included 88 and 81 cases, respectively. Patients' demographic characteristics were generally similar in both periods. In the pre-intervention period, fifty-one patients (58%) were male, and the median age was 64.5 years (IQR 56, 71.5). In the post-intervention period, fifty-one patients (63%) were male, and the median age was 63 years (IQR 50, 72). These characteristics were similar in both periods. Most patients were admitted to the general internal medicine unit. The three most common underlying diseases were hypertension, end-stage renal disease, and diabetes mellitus, respectively ( Table 1). development of treatment protocols. However, ID consultations were not always mandated in the protocols of some studies [6][7][8][9][10]. The value of adherence to treatment protocols, ID consultations, or a combination of both is difficult to determine.
Maharaj Nakorn Chiang Mai Hospital, an affiliated hospital of Chiang Mai University, is a 1400-bed tertiary care referral center in Northern Thailand. There have been 80-90 patients with SA-BSI each year since 2013 (data from the Diagnostic Laboratory, Maharaj Nakorn Chiang Mai Hospital, Chiang Mai, Thailand). The primary care team includes residents and attending physicians who are responsible for patient management and for consulting specialists if indicated. For example, with SA-BSI, the primary care team sometimes manages patients by themselves without ID consultation. To improve patient care, we therefore developed a hospital-wide protocol for the management of patients with SA-BSI.
The primary objective of this study was to determine the 90-day mortality rate of patients who had SA-BSI. The secondary objectives were to determine (1) the recurrent rate of SA-BSI; (2) the rate of appropriate management of SA-BSI, e.g., the rate of investigations for metastatic infections, the rate of follow-up blood culture at 72 h after receiving appropriate antibiotics; (3) the rate of appropriate of antibiotic prescription and factors associated with appropriateness; (4) risk factors for death.

Demographic Characteristics
One-hundred and sixty-nine patients who had SA-BSI and met the inclusion criteria were enrolled ( Figure 1). The pre-intervention and post-intervention period included 88 and 81 cases, respectively. Patients' demographic characteristics were generally similar in both periods. In the pre-intervention period, fifty-one patients (58%) were male, and the median age was 64.5 years (IQR 56, 71.5). In the post-intervention period, fifty-one patients (63%) were male, and the median age was 63 years (IQR 50, 72). These characteristics were similar in both periods. Most patients were admitted to the general internal medicine unit. The three most common underlying diseases were hypertension, end-stage renal disease, and diabetes mellitus, respectively (Table 1).

Clinical Characteristics
Overall, fifty-eight patients (34.3%) had concurrent sites of infection. The common concurrent sites were skin and soft tissue (36 patients, 21.3%), bone and joint (35 patients, 20.7%), and respiratory tract (30 patients, 17.7%) ( Table 2). Clinical characteristics were similar in both periods, except that the proportion of patients with shock was higher in the pre-intervention period (36.4% vs. 22.2%, p = 0.044). Laboratory data were similar in both periods, except for alanine aminotransferase, which was higher in the pre-intervention Antibiotics 2022, 11, 827 4 of 12 period (p = 0.023). The proportion of methicillin-susceptible strains was higher in the pre-intervention period (98.9% vs. 91.4%, p = 0.029).

Process Measures after Implementation of SA-BSI Treatment Protocol
The appropriateness of the management of patients who had SA-BSI increased significantly in the post-intervention period ( Table 3) in terms of both nonpharmacologic and antibiotic management.
Third, the infectious disease consultation rate was similar in both periods (50% vs. 56.8%, p = 0.377). The ID consultation rate for complicated SA-BSI in the post-intervention period was 73.7%; for optional ID consultation for uncomplicated SA-BSI in the post-intervention period, it was 41.9%.
Forth, source control was no different between the pre-and post-intervention periods (34.1 vs. 43.2%).

Pharmacologic Management Antibiotic Prescription
Antibiotics prescribed for the treatment of SA-BSI included cloxacillin, cefazolin, piperacillin/tazobactam, meropenem, imipenem/cilastatin, and vancomycin, as shown in Table 3. Cloxacillin was more likely to be prescribed, ceftriaxone was less likely to be prescribed, and no patients received piperacillin/tazobactam in the post-intervention period. However, rather than looking at specific drugs in detail, we aimed to evaluate the appropriateness of antibiotic prescription, as described in Section "Appropriateness of Antibiotic Prescription".

Appropriateness of Antibiotic Prescription
When looking at each component of appropriateness of antibiotic prescription, we found the following: right drug (72.7% vs. 92.6%, p = 0.001), right dose (64.8% vs. 90.1%, p < 0.001), right route (85.2% vs. 96.3%, p = 0.014), and right duration (52.3% vs. 90.1%, p < 0.001) which were higher in the post-intervention period. The appropriateness of antibiotic prescription which included all components was also higher in the post-intervention period (42% vs. 81.5%, p < 0.001). The number of patients who received intravenous antibiotics for more than 14 days was significantly higher in the post-intervention period (60.2% vs. 79%, p = 0.007).
As the primary objective was to determine overall mortality, we compared overall mortality between the two periods. The mortality rate was 34.1% and 27.2% in the pre-and post-intervention periods, respectively (p = 0.329) ( Table 3). The patients who died were older, were more likely to be admitted to the medical intensive care unit, and were more likely to have hypertension, dyslipidemia, malignancy, cardiovascular disease, respiratory tract infections, and hypoalbuminemia. On the other hand, patients who survived were more likely to have skin and soft tissue infections and source control (especially drainage), received intravenous antibiotics for at least 14 days, and received appropriate antibiotic prescriptions ( Table 4).
Recurrent SA-BSI was rare (three patients, 1.8%), and thus no further analysis was conducted relating to this outcome.  Data are presented as the count (%), unless otherwise specified. cu.mm., cubic millimeter; mg/dL, milligrams per deciliter; IU/L, international unit per liter; mg/L, milligrams per liter.

Post Hoc Analysis
This study was not planned to evaluate the impact of ID consultation in addition to the implementation of a hospital-wide protocol. However, as the rate of ID consultation was not different between the two periods and the overall consultation rate was 53.3%, we further analyzed the impact of ID consultation on the management of patients with SA-BSI.
However, ID specialists were less frequently consulted when a patient was admitted to the intensive care unit (OR 0.3; 95% CI 0.1-0.8, p = 0.022).

Discussion
This study demonstrated that the implementation of a treatment protocol for SA-BSI improved the quality of care. The radiologic imaging rate, follow-up blood culture rate, and appropriateness of antibiotic prescription were significantly increased in the post-intervention period, which is concordant with previous reports [5,[7][8][9][11][12][13][14]. This intervention raised awareness among physicians at the hospital of SA-BSI. A clear and concise protocol may guide non-ID physicians to properly manage patients with SA-BSI.
Despite the improvement in the appropriateness of antibiotic prescription in terms of the right drug, right dose, right route, and right duration and thorough investigations to determine whether the patients had metastatic infections, including echocardiography and radiologic imaging (i.e., MRI of the spine and CT scans of the abdomen), we failed to demonstrate a significant reduction in mortality (34% in pre-intervention vs. 27% in post-intervention, p-value = 0.329) or recurrent infection.
Most studies that implemented protocols to treat patients with SA-BSI show improved quality of care and decreased mortality if the protocol was adhered to [5][6][7]9,[11][12][13]. The majority of these studies incorporated ID consultation into the protocol. Their results show that ID specialists play an important role and reduce mortality in patients with SA-BSI. However, some studies did not show a significant reduction in mortality rate, even when there was an improvement in antibiotic appropriateness and source control [8,14]. These studies implemented and monitored adherence to the protocol but did not mandate consultations with ID specialists in the post-intervention period, similar to our study. In the authors' view, along with previous studies [5,[7][8][9][11][12][13][14], a hospital-wide protocol that includes ID consultation is recommended.
This study was designed to guide non-ID physicians to manage patients with uncomplicated SA-BSI properly with optional ID consultation. On the other hand, ID consultation is recommended for patients with complicated SA-BSI. However, ID consultation was adhered to in complicated SA-BSI in only 74% of cases in the post-intervention period. The appropriateness of antibiotic prescription was associated with ID consultation, being in the post-intervention period (hospital-wide protocol implementation), and thorough investigation for metastatic infections (with indirectly reflected appropriate treatment duration and source control). Interestingly, ID specialists were less involved in the care of patients that developed SA-BSI in the ICU. The reasons for this need to be further explored.
In this study, the factors associated with mortality were related to host factors and severe infection. These factors included alteration of consciousness at first presentation, having a malignancy as an underlying disease, hypoalbuminemia (albumin ≤2.5 mg/dL) at first presentation, and having a concurrent respiratory infection. The alteration of consciousness may represent an acute brain dysfunction from sepsis [15]. Patients who had a malignancy had a higher risk for infection, subsequent complications, and death [16]. Low albumin may represent a greater severity of illness [17]. Those with a concurrent respiratory tract infection had a high prevalence of septic shock [18]. On the other hand, drainage source of infection can reduce mortality. Patients who underwent source control had a lower mortality rate than patients who did not [19].
In this study, one-fifth and one-fourth of patients in the pre-and post-intervention periods, respectively, were exposed to antibiotics within 3 months prior to the occurrence of SA-BSI; however, this factor was not associated with treatment outcome. In addition, previous studies have reported that prior exposure to third-generation cephalosporins and fluoroquinolones was associated with MRSA infection [20][21][22][23][24]. We also found a higher proportion of methicillin-resistant strains in patients who had been exposed to antibiotics compared with those who had not (13.2%; 5 of 38 vs. 2.3%; 3 of 131, p-value = 0.015). However, nine patients who had been exposed to vancomycin had methicillin-susceptible S. aureus (MSSA) BSI.
The strength of this study lies in the fact that the treatment protocol may be applied to a hospital where infectious disease specialists are not available. Although this study did not demonstrate a death reduction, the patient management improved significantly in terms of the appropriateness of antibiotics, follow-up blood culture, and attempt to discover metastatic infections.
This study had several limitations. First, the mortality rate appeared to be 34% and 27% in the pre-and post-intervention periods, which were different from the number used for the sample size calculation (40% and 20% for the pre-and post-intervention periods). If this was the case, a larger sample size may be required to detect the difference in the mortality rate. Second, this study was not designed to measure real-time protocol adherence in the post-intervention period, as this was one of the programs for quality improvement and patient safety. However, we performed the process measure including protocol adherence after implementing the protocol as presented in this study. Measuring protocol adherence helped to determine the room for improvement.
One may be concerned about the impact of the COVID-19 pandemic on this study's results. The first COVID-19 case at the Maharaj Nakorn Chiang Mai Hospital was documented in the third week of January 2020, and the hospital policy then reduced the number of admissions for nonemergency conditions, including nonemergency surgeries and procedures, in March 2020. However, we continued to service patients who required hospitalization. In addition, the number of patients with SA-BSI was similar to the hospital's annual reports since 2013 (data from the Diagnostic Laboratory, Maharaj Nakorn Chiang Mai Hospital, Chiang Mai, Thailand). Only one patient with COVID-19 in the post-intervention period had SA-BSI, who had an ID consultation and survived. Therefore, if COVID-19 did affect our results, its effects were minimal and similar in both periods. and a two-sided alpha of 0.05, at least 79 cases per group was required. Therefore, the sample size needed to be 158 cases or more.

Conclusions
A treatment protocol involving thorough investigations for metastatic infections, follow-up blood culture at 72 h after treatment, and appropriate antibiotic prescription significantly improved the quality of care for patients with SA-BSI. However, the mortality rate did not decrease. A hospital-wide protocol with recommended ID consultation may help to improve patient outcomes.
Author Contributions: Conceptualization, R.C.; methodology, R.C.; formal analysis, S.Y.; resources, K.K. and P.K.; data curation, K.K.; writing-original draft preparation, K.K.; writing-review and editing, R.C.; visualization, K.K. and P.K.; supervision, R.C. All authors have read and agreed to the published version of the manuscript. Informed Consent Statement: Patient consent was waived as the implementation of the treatment protocol was part of a quality improvement program for patient safety. The data were retrospectively collected.