Compliance of clinical microbiology laboratories with recommendations for the diagnosis of bloodstream infections: Data from a nationwide survey in Italy

Abstract In 2014, the Italian Working Group for Infections in Critically Ill Patient of the Italian Association of Clinical Microbiologists updated the recommendations for the diagnostic workflow for bloodstream infections (BSI). Two years after publication, a nationwide survey was conducted to assess the compliance with the updated recommendations by clinical microbiology laboratories. A total of 168 microbiologists from 168 laboratories, serving 204 acute care hospitals and postacute care facilities, were interviewed during the period January–October 2016 using a questionnaire consisting of nineteen questions which assessed the level of adherence to various recommendations. The most critical issues were as follows: (a) The number of sets of blood cultures (BC) per 1,000 hospitalization days was acceptable in only 11% of laboratories; (b) the minority of laboratories (42%) was able to monitor whether BCs were over or under‐inoculated; (c) among the laboratories monitoring BC contamination (80%), the rate of contaminated samples was acceptable in only 12% of cases;(d) the Gram‐staining results were reported within 1 hr since BC positivity in less than 50% of laboratories. By contrast, most laboratories received vials within 2–4 hr from withdrawal (65%) and incubated vials as soon as they were received in the laboratory (95%). The study revealed that compliance with the recommendations is still partial. Further surveys will be needed to monitor the situation in the future.


| INTRODUC TI ON
The recent technological advances in diagnostic microbiology (e.g. the introduction of MALDI-TOF-based methods and of molecular biology-based syndromic panels) have revolutionized the workflow of clinical microbiology laboratories (Brooks, 2013;Laupland & Valiquette, 2013;Opota, Corxatto, & Prod'hom G., Greub G., 2015). Concerning blood cultures (BCs), the average reporting time, can be significantly shortened providing clinicians with earlier information on infecting pathogens and their susceptibility profile that allow a more rapid revision or confirmation of the empirical therapy (Cohen et al., 2015;Liesenfeld, Lehman, Hunfeld, & Kost, 2014;Livermore & Wain, 2013;Maurer, Christner, Hentschke, & Rohde, 2017). Two years after publication, we conducted a survey to evaluate the adherence to the recommendations by Italian clinical microbiology laboratories.

| Participants and data collection
A total of 168 microbiologists (from 168 laboratories) were interviewed. Altogether the laboratories served 204 hospitals and postacute care facilities (some laboratories acted as hubs for several hospitals). The data were collected from January to October 2016.

| Survey
Each participant received a questionnaire with 19 questions and was given assistance in answering from the bioMérieux Italia Company specialists. All interviewed were bioMérieux customers who used the BACT/ALERT 3D BC monitoring system (bioMérieux). The questionnaires were collected and processed anonymously.
Each question had four possible answers. For each choice, a score (ranging from 0 to 3) was assigned to grade the level of adherence of the assessed behavior to the updated recommendations (Table A1, Figure 1).  19. What is your BC contamination rate? Do you produce cumulative reports as support?
Results were merged to calculate an average questionnaire score per center and an average answer score, for each question, intercenter.

| RE SULTS
Overall, 168 microbiologists from 168 laboratories were inter- laboratories was as follows: 68 in northern Italy, 59 in central Italy, and 41 in southern Italy. Figure 1 shows the average answer score for each question.
Question no. 16 was excluded from the evaluation because of the low number of responses. Questions no. 7, 11, 13, and 19 yielded the lowest average scores. Among these, question no. 7 (no. of sets of blood cultures carried out for 1,000 days of hospitalization) yielded the lowest score. Only the 58% of laboratories (98/168) F I G U R E 1 Average of the results for each question in all the centres interviewd. The value at the end of the bars indicates the average answered scored for each question (score ranging from 0 to 3, zero = no one answered)

Concerning question; (question no.)
Possible answer  Gram-stain results on positive BCs as soon as possible (Clerc et al., 2013;GLIPaC, 2014;Thairu, Nasir, & Usman, 2014). Out of the 168 laboratories, 140 (83%) always reported Gram-stain results, 28 (17%) either did not report at all or occasionally (Figure 3). More relevant, however, is the evaluation of the timing of communication of the results of microscopic observation (

| D ISCUSS I ON
Blood cultures remains the gold standard for the diagnosis of BSIs.
As Miller et al. stated in their guidelines, "the diagnosis of bloodstream infections is one of the most critical functions of clinical microbiology laboratories" (Miller et al., 2018). Therefore, it is of fundamental importance for microbiologists, based on the available technological and human resources, to implement a diagnostic workflow capable of returning useful results to clinicians in the shortest possible timeframe to maximize impact on clinical decisions and patients outcomes (Serpa-Pinto & Cardoso, 2014;Seymour et al., 2017;Yealy et al., 2015). Monitoring suitable indicators can contribute to these purposes (Lamy, Ferroni, Henning, Cattoen, & Laudat, 2018). Therefore, the rules and indicators reported in our recommendations should not be perceived as a burden for the laboratory but rather as a guidance to improve the use of BCs for the benefit of patients (GLIPaC, 2014).

F I G U R E 2
Number of set received for each BC, expressed in percentage for each Hospital. A (1) = one set; B (2) = two sets; C (3) = three sets; D (4) = four sets F I G U R E 3 Hospital Adherence in Gram-stain reporting. "yes" means that the microbiologist always reported Gram-stain results; "no" that microbiologist never reported results; "sometimes" that microbiologist communicated results occasionally With the intention to monitor the adherence to our document by the clinical microbiology laboratories and to identify areas for improvement, we conducted a fact-finding survey in our country. From the data collected, several critical issues were detected, showing that adherence to the recommendations is still far from satisfactory. Some of these issues deserve a special attention. First, and probably the most important, is the deviation from the minimum required number of blood cultures ordered (the optimal is 103-108 per 1,000 hospital/days). This indicator, although not properly indicative of laboratory performance, can indicate correct/incorrect behaviors of clinicians in ordering BCs (EARSnet, 2012;Karch et al., 2015). A second critical issue is that a significant percentage of the laboratories perform the Gram-stain and communicate the results in times longer than those recommended. It is well demonstrated from the literature, how this delay may impact on patient outcomes (Clerc et al., 2013;Thairu et al., 2014). A third critical issue concerns contamination rates, which deserves more awareness and attention. The optimal value is <3%, but only 10% of laboratories were in this range. Reducing the number of contaminated BCs avoids useless or even misleading reports (Bates, Goldman, & Lee, 1991;Dawson, 2014;Gander, 2009;Jakko, Hilt, & Bosboomb, 2013;Snyder et al., 2012). This parameter, which reflects the quality of withdrawal practices in terms of asepsis conditions during the collection of BCs, is also useful to understand when and where it is necessary to organize training courses for medical and/or nursing staff on methods for BC collection, storage, and transport standards (Rupp, Cavalieri, Marolf, & Lyden, 2017;Snyder et al., 2012). Another critical issue related with the performance of laboratories was that very few laboratories monitor the volume of blood inoculated in BCs.
In an era of remarkable technology innovation in clinical microbiology, a drastic reduction of reporting times is possible (Arena et al., 2016;Özenci & Rossolini, 2019). In this perspective, it is noteworthy that most laboratories were unable to answer question no. 16, which had the purpose of evaluating the diffusion of rapid diagnostic systems for BCs. Therefore, it could be useful to repeat this survey in the future, focusing on this aspect. Microbiologists should also be encouraged to better apply the SOPs on BCs before the next survey, to verify whether a call for adherence to the procedures is actually effective in achieving greater compliance.

| CON CLUS IONS
In conclusion, optimal practices of BC sampling and processing require thorough understanding of several issues. Quality control programs, including software-based controls of pre-and postanalytical variables, should be strengthened to address the shortcomings described by numerous authors and also emerged in our study. We hope that the results of this first survey could encourage microbiologists to improve adherence to BC guidelines and recommendations.

ACK N OWLED G M ENT
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

E TH I C S S TATEM ENT
A written informed consent to publish the data was obtained from all survey participants.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data are provided in full in the results section of this paper.