Study selection
A total of 529 studies were retrieved from the databases. From the total, 478 studies were excluded based on titles and abstracts and 12 studies were excluded based on full-text articles (Figure 1). Primary reasons for exclusion were: non-ICU, non-pharmacist-led intervention, non-SUP-related medications, cannot extract ICU data separately, reviews, case reports, and duplicate literature (Figure 1). We included 8 studies from 9 articles in the narrative synthesis.[14, 23-30] All studies were cohort studies, of which 6 (75.0%) were retrospective and the other 2 (25.0%) were prospective. Observation periods ranged from 2 weeks to 6 months. All studies assessed appropriateness during ICU hospitalization. In addition, 4 (50.0%) studies assessed appropriateness at ICU transfer and hospital discharge at the same time (Table 1).
Participant characteristics
Most studies included adult patients (6, 75.0%) and the other 2 (25.0%) did not specify the study population (Table 2). Regarding the type of ICU, two (25.0%) studies included patients in medical and surgical ICUs, two (25.0%) studies only included patients in medical ICU, and the other 4 (50.0%) studies did not specify the ICU category. Five (62.5%) studies included all patients admitted to the ICU, while 3 (37.5%) studies only focused on patients who received AST. Inclusion criteria varied between studies but most of them (5, 62.5%) excluded patients having an additional indication for AST (e.g., active GIB, active peptic ulcer disease, and Zöllinger-Ellison syndrome) or they were not indicated for SUP pharmacotherapy regardless of risk factors (e.g., total gastrectomy). [14, 25, 27-29]
Risk of bias within studies
The NOS quality stars ranged between 5 and 7, and the average score was 5.88 for cohort studies (Table 3). Only 1 (12.5%) cohort study was regarded as high quality (NOS ≥7 points).
Intervention content and delivery
Pharmacist interventions mainly included 4 aspects: 1) clarify indications for SUP pharmacotherapy; 2) education and awareness campaign; 3) reviewed patients on SUP pharmacotherapy during rounds; 4) adjustments of drug use (Table 4).
Four (50%) studies clarified the indication for the initiation and discontinuation of SUP pharmacotherapy by developing locally SUP pharmacotherapy guidelines/protocol or algorithm.[14, 23, 27, 29] Four (50%) studies provided the medical staff with an educational intervention and/or supplied a pocket card of SUP pharmacotherapy indications for reference.[24, 25, 27, 28]
In 3 (37.5%) studies, pharmacists reviewed each patient on SUP pharmacotherapy during medical ICU rounds.[24, 25, 28] In 5 (62.5%) studies, pharmacists made appropriate changes on SUP pharmacotherapy, in which 2 (25.0%) studies gave the pharmacist prescriptive authority to make such changes (i.e. initiate, continue, discontinue, or modify the route of medication administration) for SUP pharmacotherapy only.[14, 23, 24, 27, 28]
Effects on inappropriate use of SUP pharmacotherapy
To clarify the definition of ‘inappropriate’, we first clarified the indication of SUP pharmacotherapy in all studies. Based on the most recent published guidelines and the latest evidence at the time of the study’s initiation, the indications for and cessation of SUP pharmacotherapy were different in each study (Appendix Table 2,3). For the initiation of SUP pharmacotherapy, it involved 12 major risk factors (to meet one) and 14 minor risk factors (to meet two or more). The most common major risk factors were mechanical ventilation for >48 hours and coagulopathy which were used by 7 (87.5%) studies. The common minor risk factors were high-dose glucocorticoid use and severe sepsis or septic shock which were used by 5 (62.5%) studies and 4 (50.0%) studies. For the cessation of SUP pharmacotherapy, four (50.0%) studies specified that SUP pharmacotherapy should be ceased when there is no ongoing indication.[14, 23, 25, 27] Two (25.0%) studies specified that SUP pharmacotherapy should be ceased when patients are discharged from ICU.[24, 27] One (12.5%) study specified that SUP pharmacotherapy should be ceased when patients received enteral feeding.[23] Three (37.5%) studies did not specify the cessation of SUP pharmacotherapy.[26, 28, 29]
Between pre- and post- intervention groups, the assessment time of appropriateness varied from studies (Table 5). Seven studies comprised the incidence of inappropriate SUP initiation during ICU hospitalization, of which 5 (71.4%) studies found a significant intervention effect.[14, 24, 25, 27, 29] Four studies comprised the incidence of inappropriate continuation of SUP pharmacotherapy at ICU transfer, of which 2 (50.0%) studies found a significant intervention effect.[14, 28] Five studies included the incidence of inappropriate continuation of SUP pharmacotherapy at hospital discharge, of which 3 (60.0%) studies found a significant intervention effect.[14, 23, 28]
Effects on complications and economic outcomes
Four studies identified the complications related to SUP pharmacotherapy (Table 6). There was no significant difference in the incidence of Clostridioides difficile-associated disease, pneumonia or hospital-acquired pneumonia, gastrointestinal bleeding, and thrombocytopenia between pre- and post- intervention groups.
Four (50%) studies explored the economic benefits of pharmacist-led interventions improving SUP pharmacotherapy (Table 7). [14, 23, 24, 29] Anstey 2019 determined the extrapolated direct savings to all Australian intensive care units from reduced SUP pharmacotherapy were $1.61 million/year, and indirect savings from the reduction in complications were $12.86 million/year nationally.[23] Masood 2018 clarified the pharmacist-led interventions could reduce the cost of medications for inappropriate SUP pharmacotherapy during the study period from $2,433.00 to $239.80.[6] Buckley 2015 and Coursol 2005 identified the cost of the drugs for SUP per patient and clarified that the pharmacist-led intervention reduced it from $30.52±51.45 to $8.91±11.03 and $8.74 to $6.68.[9, 12]