Background: Sepsis, a life-threatening infection-induced inflammatory condition, has significant global health impacts. Timely detection is crucial for improving patient outcomes as sepsis can rapidly progress to severe forms. The application of machine learning (ML) and deep learning (DL) to predict sepsis using electronic health records (EHRs) has gained considerable attention for timely intervention. Methods: PubMed, IEEE Xplore, Google Scholar, and Scopus were searched for relevant studies. All studies that used ML/DL to detect or early-predict the onset of sepsis in the adult population using EHRs were considered. Data were extracted and analyzed from all studies that met the criteria and were also evaluated for their quality. Results: This systematic review examined 1942 articles, selecting 42 studies while adhering to strict criteria. The chosen studies were predominantly retrospective (n = 38) and spanned diverse geographic settings, with a focus on the United States. Different datasets, sepsis definitions, and prevalence rates were employed, necessitating data augmentation. Heterogeneous parameter utilization, diverse model distribution, and varying quality assessments were observed. Longitudinal data enabled early sepsis prediction, and quality criteria fulfillment varied, with inconsistent funding–article quality correlation. Conclusions: This systematic review underscores the significance of ML/DL methods for sepsis detection and early prediction through EHR data.

Bacterial culture and biochemical testing (CBtest) have been the cornerstone of pathogen identification in the diagnostic microbiology laboratory. With the advent of Sanger sequencing and later, next-generation sequencing, 16S rRNA next-generation sequencing (16SNGS) has been proposed to be a plausible platform for this purpose. Nevertheless, usage of the 16SNGS platform has both advantages and limitations. In addition, transition from the traditional methods of CBtest to 16SNGS requires procurement of costly equipment, timely and sustainable maintenance of these platforms, specific facility infrastructure and technical expertise. All these factors pose a challenge for middle-income countries, more so for countries in the lower middle-income range. In this review, we describe the basis for CBtest and 16SNGS, and discuss the limitations, challenges, advantages and future potential of using 16SNGS for bacterial pathogen identification in diagnostic microbiology laboratories of middle-income countries.

This paper investigates the role of Group II Secretory Phospholipase A2 (sPLA2-IIA) as a biomarker for the diagnosis of sepsis and bacterial infection in adults. Sepsis and bacterial infection are common problems encountered by patients in the hospital and often carry adverse outcomes if not managed early. Two independent reviewers conducted a comprehensive search using Ovid MEDLINE published from years 1993 to 2016 and SCOPUS published from year 1985 to 2017 to screen for relevant studies. The main inclusion criteria included adult subjects, patients with suspected or confirmed signs of infection and relevant outcomes which looked into the role of sPLA2-IIA in detecting the presence of sepsis and bacterial infection in the subjects. Four studies met the inclusion criteria. SPLA2-IIA was found to be effective in detecting the presence of sepsis and bacterial infection in adults. The levels of serum sPLA2-IIA also correlated well with the presence of sepsis and bacterial infection. This systematic review highlights the role of sPLA2-IIA as a reliable tool to diagnose sepsis and bacterial infection in adult patients. Nonetheless, further studies should be done in the future to provide more compelling evidence on its application in the clinical setting.

Antimicrobial resistance remains a significant public health issue, and to a greater extent, caused by the misuse of antimicrobials. Monitoring and benchmarking antimicrobial use is critical for the antimicrobial stewardship team to enhance prudent use of antimicrobial and curb antimicrobial resistance in healthcare settings. Employing a comprehensive and established tool, this study investigated the trends and compliance of antimicrobial prescribing in a tertiary care teaching hospital in Malaysia to identify potential target areas for quality improvement. A point prevalence survey method following the National Antimicrobial Prescribing Survey (NAPS) was used to collect detailed data on antimicrobial prescribing and assessed a set of quality indicators associated with antimicrobial use. The paper-based survey was conducted across 37 adult wards, which included all adult in-patients on the day of the survey to form the study population. Of 478 patients surveyed, 234 (49%) patients received at least one antimicrobial agent, with 357 antimicrobial prescriptions. The highest prevalence of antimicrobial use was within the ICU (80%). Agents used were mainly amoxicillin/β-lactamase inhibitor (14.8%), piperacillin/β-lactamase inhibitor (10.6%) and third-generation cephalosporin (ceftriaxone, 9.5%). Intravenous administration was ordered in 62.7% of prescriptions. Many antimicrobials were prescribed empirically (65.5%) and commonly prescribed for pneumonia (19.6%). The indications for antimicrobials were documented in the patients’ notes for 80% of the prescriptions; however, the rate of review/stop date recorded must be improved (33.3%). One-half of surgical antimicrobial prophylaxis was administered for more than 24 h. From 280 assessable prescriptions, 141 (50.4%) were compliant with guidelines. Treating specialties, administration route, class of antimicrobial, and the number of prescriptions per patient were contributing factors associated with compliance. On multivariate analysis, administering non-oral routes of antimicrobial administration, and single antimicrobial prescription prescribed per patient was independently associated with non-compliance. NAPS can produce robust baseline information and identifying targets for improvement in antimicrobial prescribing in reference to current AMS initiatives within the tertiary care teaching hospital. The findings underscore the necessity to expand the AMS efforts towards reinforcing compliance, documentation, improving surgical prophylaxis prescribing practices, and updating local antibiotic guidelines.

Early diagnosis of sepsis and bacterial infection is imperative as treatment relies on early antibiotic administration. There is a need to develop new biomarkers to detect patients with sepsis and bacterial infection as early as possible, thereby enabling prompt antibiotic treatment and improving the survival rate. Fifty-one adult patients with suspected bacterial sepsis on admission to the Emergency Department (ED) of a teaching hospital were included into the study. All relevant cultures and serology tests were performed. Serum levels for Group II Secretory Phospholipase A2 (sPLA2-IIA) and CD64 were subsequently analyzed. Sepsis was confirmed in 42 patients from a total of 51 recruited subjects. Twenty-one patients had culture-confirmed bacterial infections. Both biomarkers were shown to be good in distinguishing sepsis from non-sepsis groups. CD64 and sPLA2-IIA also demonstrated a strong correlation with early sepsis diagnosis in adults. The area under the curve (AUC) of both Receiver Operating Characteristic curves showed that sPLA2-IIA was better than CD64 (AUC = 0.93, 95% confidence interval (CI) = 0.83–0.97 and AUC = 0.88, 95% CI = 0.82–0.99, respectively). The optimum cutoff value was 2.13μg/l for sPLA2-IIA (sensitivity = 91%, specificity = 78%) and 45 antigen bound cell (abc) for CD64 (sensitivity = 81%, specificity = 89%). In diagnosing bacterial infections, sPLA2-IIA showed superiority over CD64 (AUC = 0.97, 95% CI = 0.85–0.96, and AUC = 0.95, 95% CI = 0.93–1.00, respectively). The optimum cutoff value for bacterial infection was 5.63μg/l for sPLA2-IIA (sensitivity = 94%, specificity = 94%) and 46abc for CD64 (sensitivity = 94%, specificity = 83%). sPLA2-IIA showed superior performance in sepsis and bacterial infection diagnosis compared to CD64. sPLA2-IIA appears to be an excellent biomarker for sepsis screening and for diagnosing bacterial infections, whereas CD64 could be used for screening bacterial infections. Both biomarkers either alone or in combination with other markers may assist in decision making for early antimicrobial administration. We recommend incorporating sPLA2-IIA and CD64 into the diagnostic algorithm of sepsis in ED.

The purpose of this meta-analysis was to compare the ability of the qSOFA in predicting short- (≤30 days or in-hospital mortality) and long-term (>30 days) mortality among patients outside the intensive care unit setting. Studies reporting on the qSOFA and mortality were searched using MEDLINE and SCOPUS. Studies were included if they involved patients presenting to the ED with suspected infection and usage of qSOFA score for mortality prognostication. Data on qSOFA scores and mortality rates were extracted from 36 studies. The overall pooled sensitivity and specificity for the qSOFA were 48% and 86% for short-term mortality and 32% and 92% for long-term mortality, respectively. Studies reporting on short-term mortality were heterogeneous (Odd ratio, OR = 5.6; 95% CI = 4.6–6.8; Higgins’s I² = 94%), while long-term mortality studies were homogenous (OR = 4.7; 95% CI = 3.5–6.1; Higgins’s I² = 0%). There was no publication bias for short-term mortality analysis. The qSOFA score showed poor sensitivity but moderate specificity for both short and long-term mortality, with similar performance in predicting both short- and long- term mortality. Geographical region was shown to have nominal significant (p = 0.05) influence on qSOFA short-term mortality prediction.