As part of clinical quality improvement activities at the San Francisco Veterans Affairs Healthcare System, we analyzed positive SARS-COV-2 RT-PCR testing (nasopharyngeal or oropharyngeal swab). These included both symptomatic and asymptomatic employees (the latter tested due to exposure or in routine surveillance monitoring). We limited analysis to cases March 2020 through January 2021 whose initial positive testing was performed at the San Francisco Veterans Affairs Healthcare System laboratory using the same Abbott RealTime SARS-COV-2 assay providing directly comparable Ct values. The Abbott assay is United States Food and Drug Administration Emergency Use Authorization-approved as a qualitative test and was calibrated in-house using standard reference materials [3]. We excluded cases initially diagnosed with COVID-19 by use of other assays at our site or elsewhere.
We reviewed our employee health unit’s COVID-19 database, using narrative information to identify whether infection was attributable to residing with someone whose COVID-19 preceded that of the employee. The database narrative text, in all cases, had been entered by one of four trained clinicians (two physicians and two advanced practice nurses) at the time of intake by interviewing the cases and was supplemented as appropriate by a nurse who carried out contact tracing. The interview specifically focused on potential sources of exposure, including any household contacts who were symptomatic or diagnosed with SARS-COV-2. The designation of a household source required that the narrative indicated onset of symptoms or diagnosis preceding the onset in the employee case in a time period consistent with the COVID-19 incubation period (2 to 14 days). A single researcher (author AC) categorized the employees on this basis, with adjudication when questions arose by another (author SD) using supplemental information garnered through the return-to-work process for ill employees.
We also determined if the employee was symptomatic at initial testing. We defined symptomatic as log documentation of the employee reporting any of a standard check list of COVID-19 related complaints we routinely used to categorize any employees as a “person under investigation” (PUI). These included: a body temperature of 100 degrees Fahrenheit or greater, malaise or muscle ache, upper or lower respiratory complaints, loss of smell or taste, other neurological symptoms, gastrointestinal upset, or dermal or ocular complaints.
Because we anticipated that symptomatic status at the time of initial test positivity might be associated with a higher estimated viral load, we wanted to assess any confounding association between such symptomatic status and a household source of exposure. To do so, we tested the cross-tabulation of symptomatic status vs. household exposure (both defined dichotomously) using the Chi square. We defined low Ct (indicative of higher viral load) as a value below the lowest quartile cutoff among the non-household contact cases: a Ct of 9.32 (estimated viral load approximating 4.6 × 106 copies per ml). We chose the lowest quartile rather than the median value of the referent HCWs (those without a household source) to reflect the lower threshold “tail” of these non-normally distributed data. Our laboratory standard curve established a linear correlation of lower Ct to higher viral load with this lower quartile cut-off being multiple orders of magnitude above the reliable limit of detection (the positive/negative test report used clinically). Logistic regression tested the association of household exposure alone or adjusted for symptomatic case status with low Ct (RStudio 2021; PBC, Boston, MA).
