Matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry (MS) is an emerging technology for rapid identification of bacterial and fungal isolates. In comparison to conventional methods, this technology isM ass spectrometry (MS) has been traditionally utilized for chemical analysis, although its utility was limited to lowmolecular-weight organic compounds (1). Matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) MS expanded the field and allowed for the analysis of biological molecules with no theoretical upper limit of mass (2). Previously employed to determine the mass of peptides and proteins, this emerging technology has been adapted for rapid identification of bacterial and fungal isolates in the clinical microbiology laboratory. In an increasing number of settings, MALDI-TOF MS has replaced traditional identification methods, including microscopy and determination of phenotypic characteristics, which typically require multiple steps (3, 4).The performance of MALDI-TOF MS for the identification of microorganisms was examined in numerous studies and shown to be accurate and reliable (5-8). In comparison to conventional methods, this technology is much less labor intensive and can provide accurate and reliable results in minutes from a single isolated colony. The only reagents utilized are the target slides that contain sample spots used for the identification of microorganisms, an organic matrix solution, formic acid for yeast isolates, pipette tips, and disposable loops or toothpicks for sample application. Therefore, reagent rental programs are usually not available from the two manufacturers of MALDI-TOF MS, so the instrument is purchased, which is a large capital cost to the laboratory. In addition, fairly expensive annual maintenance contracts of $25,000 to $30,000 are needed to ensure limited downtime of this technically complex equipment. Causes of downtime are multifaceted and consist of various issues, including time required for remote access to software for manipulation or tuning, and hardware maintenance/repair, reducing the time the instrument is available on a given day for identification of isolates. Sufficient protocols need to be in place for times when the MALDI-TOF MS will be unavailable. Although the initial instrumentation price is high and maintenance expenses are significant, the cost of identifying an isolate can be very low. Based on this, use of MALDI-TOF MS has the potential to improve laboratory efficiency, reduce turnaround times, and lower costs (9).We compared the cost of performing the bioMérieux Vitek MALDI-TOF MS (Durham, NC, USA) with that of conventional microbiological methods to determine the amount saved by the laboratory after converting to the new technology. This study ex-

Introduction: During the coronavirus disease 2019 (COVID-19) outbreak, novel approaches to diabetes care have been employed. Care in both the inpatient and outpatient setting has transformed considerably. Driven by the need to reduce the use of personal protective equipment and exposure for patients and providers alike, we transitioned inpatient diabetes management services to largely ''virtual'' or remotely provided care at our hospital. Methods: Implementation of a diabetes co-management service under the direction of the University of North Carolina division of endocrinology was initiated in July 2019. In response to the COVID-19 pandemic, the diabetes service was largely transitioned to a virtual care model in March 2020. Automatic consults for COVID-19 patients were implemented. Glycemic outcomes from before and after transition to virtual care were evaluated. Results: Data over a 15-week period suggest that using virtual care for diabetes management in the hospital is feasible and can provide similar outcomes to traditional face-to-face care. Conclusion: Automatic consults for COVID-19 patients ensure that patients with serious illness receive specialized diabetes care. Transitioning to virtual care models does not limit the glycemic outcomes of inpatient diabetes care and should be employed to reduce patient and provider exposure in the setting of COVID-19. These findings may have implications for reducing nosocomial infection in less challenging times and might address shortage of health care providers, especially in the remote areas.

Commensal bacteria from the skin and mucosal surfaces are routinely isolated from patient samples and considered contaminants. The majority of these isolates are catalase-positive Gram-positive rods from multiple genera routinely classified as diphtheroids. These organisms can be seen upon Gram staining of clinical specimens or can be isolated as the predominant or pure species in culture, raising a priori suspicion of a possible involvement in infection. With the development and adoption of matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS), suspicious isolates are now routinely identified to the species level. In this study, we performed a retrospective data review (2012 to 2015) and utilized site-specific laboratory criteria and chart reviews to identify species within the diphtheroid classification representative of true infection versus contamination. Our data set included 762 isolates from 13 genera constituting 41 bacterial species. Only 18% represented true infection, and 82% were deemed contaminants. Clinically significant isolates were identified in anaerobic wounds (18%), aerobic wounds (30%), blood (5.5%), urine (22%), cerebrospinal fluid (24%), ophthalmologic cultures (8%), and sterile sites (20%). Organisms deemed clinically significant included multiple Actinomyces species in wounds, Propionibacterium species in joints and cerebrospinal fluid associated with central nervous system hardware, Corynebacterium kroppenstedtii (100%) in breast, and Corynebacterium striatum in multiple sites. Novel findings include clinically significant urinary tract infections by Actinomyces neuii (21%) and Corynebacterium aurimucosum (21%). Taken together, these findings indicate that species-level identification of diphtheroids isolated with a priori suspicion of infection is essential to accurately determine whether an isolate belongs to a species associated with specific types of infection. C linical microbiology specimens frequently grow variable levels of commensal bacteria from the skin and mucosal surfaces in addition to true pathogens (1). The majority of these isolates are aerobic, asporogenic, irregularly shaped, non-partially-acid-fast, catalase-positive, Gram-positive rods from multiple genera with nondistinct colony morphology, routinely classified as diphtheroids (2). The term diphtheroid and coryneform are interchangeable, and for a comprehensive review of bacterial genera classified as coryneform bacteria, see two excellent reviews by Bernard et al. (2,3). In brief, the medically relevant genera whose morphological and biochemical descriptions fit in the diphtheroid classification include Arcanobacterium, Arthrobacter, Brevibacterium, Cellulomonas, Cellulosimicrobium, Corynebacterium (non-diphtheriae), Curtobacterium, Dermabacter, Exiguobacterium, Helcobacillus, Janibacter, Knoellia, Leifsonia, Microbacterium, Pseudoclavibacter, and Trueperella (2-4).Additionally, certain species within other genera share some but not all features of diphtheroids, are part of t...

Burkholderia cepacia has recently been recognized as an important pathogen in chronic lung disease in patients with cystic fibrosis (CF). Because of the social, psychological, and medical implications of the isolation of B. cepacia from CF patients, accurate identification of this organism is essential. We compared the accuracies of four commercial systems developed for the identification of nonfermenting, gram-negative bacilli with that of conventional biochemical testing for 150 nonfermenters including 58 isolates of B. cepacia recovered from respiratory secretions from CF patients. The accuracies of the four systems for identifying all nonfermenters ranged from 57 to 80%, with the RapID NF Plus system being most accurate. The accuracies of these systems for identifying B. cepacia ranged from 43 to 86%, with the Remel system being most accurate. Depending on the commercial system, from two to seven isolates were misidentified as B. cepacia. The relatively poor performance of the commercial systems requires that identification of certain nonfermenters be confirmed by conventional biochemical testing. These organisms include B. cepacia, Burkholderia sp. other than B. cepacia, and infrequently encountered environmental species (Pseudomonas and Flavobacterium species). In addition, conventional biochemical testing should be done if a commercial system fails to assign an identification to an organism. Confirmatory testing should preferably be performed by a reference laboratory with experience in working with organisms isolated from CF patients.

Over the past two decades, validation of choice models has focused on predictive validity rather than parameter bias. In real-world validation of choice models, true parameter values are unknown, so examination of parameter bias is not possible. In contrast, the main focus of this study is parameter bias in simulated scanner-panel choice data with known parameter values. Study of parameter bias enables the assessment of a fundamental issue not addressed in the choice modeling literature—the extent to which the logit choice model is capable of distinguishing unobserved effects that give rise to persistence in observed choices (e.g., heterogeneity and state dependence). Although econometric theory provides some information about the causes of bias, the extent of such bias in typical scanner data applications remains unclear. The authors present an extensive simulation study that provides information on the extent of bias resulting from the misspecification of four unobserved effects that receive frequent attention in the literature—choice set effects, heterogeneity in preferences and market response, state dependence, and serial correlation. The authors outline implications for model builders and managers. In general, the potential for parameter bias in choice model applications appears to be high. Overall, a logit model with choice set effects and the Guadagni–Little loyalty variable produces the most valid parameter estimates.