COVID-19 prevention in the UCSF campus community

September 18, 2020

Why UCSF is Safe

The risk of COVID-19 infection in UCSF’s campus research facilities is very low, for a number of reasons, including:

  • UCSF has a low population density on site. On a typical weekday, with laboratory occupancy capped at 25% to 50% capacity, there are 1,000 to 2,000 laboratory staff and trainees spread across a million square feet of research space in more than 10 major research buildings. Most faculty and administrative staff work remotely.
  • There is a very low prevalence of asymptomatic infection in the UCSF population.
  • There is very high compliance among researchers with safety measures such as masking and physical distancing.
  • Daily symptom screening with the Conversa app prevents the great majority of symptomatic employees from coming to work.
  • Anyone who reports symptoms is immediately tested by UCSF Occupational Health Services (OHS), and anyone who obtains a positive test off campus is required to contact OHS or Student Health (link).
  • Rapid testing turnaround times (fewer than 24 hours) and effective contact tracing prevent transmission to other onsite workers.

The result of these factors is an extremely low disease transmission rate. This rate is the key metric for determining workplace safety, and it is far lower at UCSF than at an undergraduate campus. Indeed, disease transmission is considerably lower within UCSF than it is in the surrounding city community.

Overview of the Contact Tracing Process

Information about OHS COVID-19 health and safety procedures can be found here. Contact tracing begins with a detailed interview of the person who has tested positive for COVID-19, using established guidelines. The interview leads to the following key pieces of information:

Infectious period. Based on the timing of symptom onset and other metrics, contact tracing determines the time period when the person who tested positive was most likely to be infectious. This important information focuses attention on the limited time window when contacts with other people need to be analyzed. In many cases, the person who tested positive was not at work on campus during this period, greatly reducing the risk to coworkers.

Source of infection. Ideally, contact tracing determines the source of the infection. About 200 UCSF employees tested positive for COVID-19 in the first 6 months of the pandemic. Contact tracing indicates that most infections were acquired from the community outside UCSF (65%) or were associated with patient care (25%). About 7% were acquired from coworkers or from work-related travel. The remaining cases are new and are still under investigation. Workplace transmission is therefore very low.

Identification of close contacts. If the person who tested positive was at work on campus during the infectious period, contact tracing then identifies other people with whom a significant interaction occurred. Based on established guidelines in place at the majority of institutions and workplaces around the country, contact tracing focuses primarily on “close contacts:” coworkers who interacted closely with the person with COVID-19 (within 6 feet for over 15 minutes). These and other contacts are then categorized according to risk level.

Close contacts are divided into three levels (see section 1 here) according to mask use. A high-risk interaction is defined as a close contact in which neither person wore a mask. A medium-risk interaction occurs when one of the two people wore a mask. These medium- and high-risk contacts are considered serious enough to warrant testing and a 14-day quarantine. If both people wore a mask during the interaction, then the risk is considered low and quarantine is not needed, although testing is an option.

Data from OHS supports the emphasis on close contacts. There have been extremely few workplace transmissions at UCSF. The great majority of these transmissions occurred in high-risk close contacts: that is, when both people interacted closely for more than 15 minutes without wearing masks, typically in break rooms or carpools. Transmission in lower risk contacts has been negligible. In other words, risk plummets when one or both parties in a conversation or in close contact wear a mask. These simple facts strongly support two important conclusions: (1) high-risk close contacts are the primary concern in contact tracing; and (2) masks and physical distancing are the best measures for reducing workplace transmission (for more detailed analysis of the value of these measures in infection control, see here, here and here).

 Minimal risk contacts. Any employee who has interacted with a coworker with COVID-19 during the infectious period but does not fit in the “close contact” categories is considered minimal risk. No testing is required and work can continue. This policy might seem unsafe to some, but studies within UCSF and at other institutions indicate that transmission is extremely unlikely for minimal-risk contacts. Mask-wearing, physical distancing, and handwashing are the most effective mechanisms for preventing transmission through minimal-risk contacts.

There is general agreement among infectious disease specialists that close person-to-person contact is the primary means of transmission. Our contact tracing results, as well as other evidence, does not support significant transmission over long distances or via contaminated surfaces (see here). In any case, masks and handwashing eliminate all minimal risk routes of transmission.

Low Prevalence of Infections in the UCSF Workplace

There have been just over 200 positive tests among UCSF employees since the pandemic began (out of almost 6,000 employees tested) (PDF). Almost all of these cases arose from testing of symptomatic employees. These numbers are based on the entire UCSF community, including UCSF Health facilities that are operating at full capacity and are considerably more crowded than UCSF research buildings. Almost all employees who have tested positive for COVID-19 work at UCSF Health. Just a small number (~10) have been in research laboratories (on the two major campuses).

OHS estimates that just 10 UCSF employees have become infected by a coworker in the workplace through Sept. 18, 2020. As described above, these transmissions occurred almost entirely among the high-risk close contacts (both parties not wearing a mask) in UCSF Health facilities, where close contact with co-workers or patients can be more difficult to avoid.

A conservative estimate of workplace transmission, based primarily on studies in UCSF Health facilities, is that there are 34 workplace infections per 1 million person-days at UCSF. In contrast, there have been 60 cases per 1 million person-days in the city of San Francisco. Given that (1) workplace transmission is likely to be lower in UCSF research buildings, and (2) most workplace infections come from outside UCSF, the data clearly indicate that working within UCSF buildings is safer than living in San Francisco.

Asymptomatic COVID Cases in the UCSF Workplace

Since the pandemic began, about 6,000 UCSF employees have been tested. In the first few months of the pandemic, testing focused on symptomatic employees. In the past two months, however, new programs have been implemented for testing hundreds of asymptomatic employees, beginning with a large group of UCSF Health employees and now extending to increasing numbers of UCSF campus personnel.

More than 1,000 asymptomatic UCSF employees have been tested thus far. The fraction of positive results in these tests was extremely low (approximately 0.8%), indicating that the prevalence of asymptomatic infection is very low in the UCSF workplace. Given the low rate of workplace transmission among symptomatic workers as described above, and given the effectiveness of masks and other measures, the risk of infection from asymptomatic employees is likely to be negligible.

UCSF continues to conduct baseline testing of several groups of asymptomatic employees and trainees, including all new hires, new students arriving on campus, new tenants in campus housing, employees in our child care facilities, and a large random cohort of health care workers in UCSF Health facilities.

UCSF is also initiating a new testing program called the COVID-19 Asymptomatic Random Employee Screening (CARES) Program. Using random testing of large numbers of employees working in specific buildings, this program will monitor the general prevalence of asymptomatic infection among our employees and trainees. This program is designed to detect an increase in COVID-19 prevalence from 1% (currently) to 2%, triggering broader testing in specific buildings when there are more than 2-3 related COVID-19-positive cases in a specific area.

Mass surveillance testing

Many undergraduate institutions are using large-scale testing to manage the spread of infections on their campuses. Modeling by epidemiologists indicates that mass testing may have some value in controlling outbreaks at undergraduate colleges with a large, interacting student population (as described here and here; online modeling site here). In undergraduate populations, the background level of asymptomatic positives is high, compliance with safety measures is poor, and the rate of transmission (reproductive number) is high: 1.5 to 2.5 (meaning that every person with COVID-19 infects another 1.5 to 2.5 people). The goal of mass surveillance testing in these populations is, in theory, to reduce the reproductive number below 1, thereby preventing outbreaks. However, this approach can fail due to unsafe student behavior (link). It is clear from experiences at many universities that personal behavior is a critical factor in reducing the rate of transmission.  

Mass surveillance testing is unlikely to be effective as an infection control measure at an academic medical center like UCSF, where there is no large undergraduate population. The UCSF on-campus population complies with health and safety guidelines, the prevalence of infections is extremely low, and the workplace reproductive number is extremely low (much less than 1 and likely below 0.2). The same modeling approaches that support mass testing at undergraduate institutions indicate that surveillance testing has little impact in a population like that at UCSF: our reproductive number is so low that frequent testing would reduce that number only slightly and have a minor effect on infection rates. Furthermore, when the reproductive number is low and infection is rare, the number of false positives (typically 1-2%) will exceed the number of true positives, resulting in unnecessary anxiety and quarantine until tests are repeated. In our workplace at UCSF, safety measures like wearing masks and maintaining physical distancing have a far greater impact than testing.