Discussion

Mounting evidence supports the assertion that increased use of microbiology laboratory resources influences patient care and healthcare expenditures. ^{Reference Ohki, Fukui, Morishita and Iwata4} During the study period, 23 CLABSIs were identified. For comparison, 9 CLABSIs were reported during the same period in 2019 (internal data). Possible explanatory factors include rapid modification of infection control practices such as placing intravenous (IV) pumps outside patient rooms, utilizing IV line extenders, and inconsistent use of alcohol-impregnated caps for needleless ports.

Clinicians were more likely to document blood-culture contamination from COVID-19 samples. Isolation potentially results in less patient-facing time, inappropriate blood-culture collection, and suboptimal antiseptic technique, which all increase the likelihood of contamination. ^{Reference LeRose, Sandhu, Polistico, Ellsworth, Cranis and Jabbo3,Reference Esquer Garrigos, Wingler and Svoronos5} In our COVID-19 cohort, Black race, ESRD, and obesity were significantly associated with blood-culture contamination. However, whether these observations represent factors intrinsic to this patient population remains unknown. Interestingly, using the standard microbiologic definition, our microbiology surveillance data did not reveal an increase in the blood-culture contamination rates during the study period compared to the same time in 2019 (data not shown). This finding may coincide with later COVID-19 hospitalizations in Texas compared to other areas at the beginning of the pandemic (June to July versus April to June 2020, respectively), resulting in detection bias. ^{Reference Trout and Chen6} However, using a clinical definition for blood-culture contamination enhances the generalizability of our findings.

Exacerbated by the pandemic, our health system, like many in the United States, experienced high workforce turnover and staffing shortages. Internal review of nursing turnover during the study period revealed rates of 9.3%–10.2%, with a benchmark national rate not to exceed 10.7%. High nursing turnover has been correlated with increases in CLABSI. ^{Reference Scheier, Kuster, Dunic, Falk, Sax and Schreiber7} There are several reasons for this, and high patient-to-nurse ratio is a major factor. ^{Reference Robert, Fridkin and Blumberg8}

This study had several limitations. First, we did not analyze COVID-19 diagnosis in relation to blood-culture collection time; thus, misclassification bias cannot be excluded. We likely underestimated the effect of isolation status on blood-culture collection. We expect that the observed ratio is likely biased toward the null. Second, blood cultures with an “unknown” contamination status were not included in the analysis. Thus, it is unclear how these data would have affected our results. Third, we did not conduct assessments on blood-culture collection techniques, nurse-to-patient ratios, or nursing staff expertise. Therefore, our statements regarding these factors are hypothetical, but they are supported by prepandemic descriptive studies. ^{Reference Trout and Chen6–Reference Robert, Fridkin and Blumberg8} Fourth, healthcare worker burnout is an important factor that was not assessed and could have affected outcomes. ^{Reference Van Wert, Gandhi and Gupta9} Finally, differing COVID-19 policies and PPE availability during the pandemic’s first wave make these results difficult to extrapolate to non–pandemic-related scenarios.

In our study, patients with COVID-19 were more likely to experience CLABSI and blood-culture contamination during the first wave of the pandemic. In our COVID-19 cohort, malignancy was significantly associated with CLABSI, and Black race, ESRD, and obesity was significantly associated with blood-culture contamination. These findings should be considered in future infection control policy development for preparedness and response to events that compromise healthcare systems.