Real-Time Automated Surveillance for Ventilator Associated Events Using Streaming Electronic Health Data

Abstract

Those with multiple VAP episodes >7 days apart with a different causative organism were counted separately. Results. Fifty-six patients (67 episodes) with VAP in 992 admissions were identified. Ten had ≥2 episodes. In 11 episodes, ≥2 isolates were found from a respiratory sample; 78 isolates were identified in total. The cohort median age was 61 (interquartile range [IQR]: 48-70) years, with 43 (76.8%) males. Fourteen (24.6%) patients had diabetes , 10 (17.5%) had chronic kidney diseases, 17 (29.8%) had congestive heart disease , 9 (15.8%) had COPD, and 5 (8.8%) had malignancy. Among isolated bacteria, Acinetobacter baumannii (ACB) was highly resistant to meropenem, levofloxacin, and amikacin (Table). The 7-day mortality was 13% (n = 7) and 31-day mortality was 43.8% (n = 21). ACB cases had higher 31-day mortality (18 [56.2%] vs. 4 [25%]; P = 0.041) and longer ICU stay (16 days [IQR: 10-27] vs. 9 [3-15]; P = 0.024; deceased excluded) than non-ACB. Colistin was used in 23 (41.1%) cases as empiric therapy and 25 (44.6%) as definitive therapy. Conclusion. High resistance rates and worse clinical outcomes were found in VAP cases due to ACB in ICU in Vietnam. Further study is warranted for appropriate treatment and infection control measures. Table: Antimicrobial resistance of bacterial isolates* in ventilator-associated pneumonia, n (%) Acinetobacter baumannii (n = 37) Klebsiella pneumoniae (n = 11) Pseudomonas aeruginosa (n = 15) Colistin 0 0 0 *Include Stenotrophomonas maltophilia (n = 6), Serratia marcescens (n = 3), Enterobacter cloacae (n = 2), Elizabethkingia meningoseptica (n = 2). Disclosures. All authors: No reported disclosures. Background. Ventilator-associated pneumonia (VAP) is a significant cause of hospital-acquired infection. Many institutions use National Healthcare Safety Network (NHSN) definitions for VAP surveillance. One criterion of the NHSN definition for possible VAP (PVAP) is a positive culture obtained via non-bronchoscopic bronchoalveolar lavage (NB-BAL). NB-BAL specimens are not routinely assessed for oropharyngeal contamination prior to quantitative culture. Thus, we hypothesized that NB-BALs can yield poor quality specimens that may contribute to the misdiag-nosis of PVAP. Methods. From May 2016 to January 2017, we performed background quality assessments for NB-BAL specimens collected from patients on mechanical ventilation for >3 days. Thereafter, we retrospectively reviewed NB-BAL quality, culture results, and contribution to NHSN-defined PVAPs. Quality assessments included number of white blood cells (WBC) or squamous epithelial cells (SEC) per low-power field (lpf). Specimens were deemed acceptable if they had ≤10 SEC/lpf by "standard" criteria and zero SEC/lpf by "strict" criteria. All specimens were cultured regardless of quality assessment results, which were not revealed to ordering clinicians. Results. Of 117 NB-BAL specimens, 8 (7%) did not pass standard quality assessment and an additional 47 (40%) did not pass strict quality assessment. Most samples (82%) were purulent (>25 WBC/lpf). Overall, 56 (48%) of samples resulted in significant growth of at least one species of bacteria (>10 4 CFU/mL). Of the 8 samples that did not pass standard assessment, 7 (87%) resulted in significant bacterial growth. Four PVAPs were diagnosed on the basis of NB-BAL specimens. Of these, all were acceptable by standard criteria, but one failed by strict criteria. Conclusion. Approximately 50% of our NB-BAL specimens had evidence of oropharyngeal contamination on quality assessment, including one specimen that contributed to a NHSN-reported PVAP. While limited by small sample size and short study duration, our data suggest that the quality of NB-BAL specimens may affect the diagnosis and surveillance of VAP. Disclosures. All authors: No reported disclosures.