COVID risk in large workplaces is akin to cyber risk. It is a question of when, not if. By constantly monitoring relevant data, instituting appropriate precautions, and creating a resilient work environment companies can reduce the impact of COVID.

Auto manufacturers shuttered North American operations in March, sending home roughly 150,000 workers, as US auto retail sales plummeted to 50% of normal levels. Over the intervening two months, shelter-in-place restrictions have flattened the growth of COVID cases across the US and a flurry of research has increased our understanding of the transmission dynamics. During this time, the auto industry retrofitted factories to enable social distancing and developed employee policies including protective equipment requirements, health screenings and temperature checks.

In mid-May, as auto plants prepared to reopen, One Concern compared COVID risk in three automotive assembly plants, each with an annual output of approximately 300,000 vehicles: (1) Tesla in Fremont, California, (2) Mercedes-Benz in Tuscaloosa, Alabama, and (3) Ford in Chicago, Illinois. The Mercedes factory was the first to reopen on April 27th, but within a few weeks, the plant was forced to suspend production because of disruption at supplier sites in Mexico.

At the same time, a public battle was occurring between Elon Musk, the CEO of Tesla, and the local Alameda County government. California Governor Gavin Newsom announced that some manufacturing could resume on May 8th, but Alameda County remained under strict shelter in place orders.

The auto industry illustrates the complex interplay between local transmission dynamics, government policy, and supply chain dynamics required for industries to bounce back.

This analysis was designed to capture the impact of local transmission dynamics in order to assess the risk of returning to work. At the time of the analysis, the situation was as follows:

TESLA: The San Francisco Bay Area was early and aggressive with shelter-in-place restrictions. Despite being a large metropolitan area with early cases, the case rates in the local area were low. The local population is healthier that average, hospitals had plenty of capacity, and new cases were decreasing. Local authorities were reticent to let the factory open and will enforce shutdowns if case rates increase or hospital capacity diminishes.

MERCEDES-BENZ: Case rates in the local area around the Tuscaloosa plant remained low, but case rates in the state of Alabama were trending upward. The population has higher-rate of underlying health conditions, including diabetes, cardiovascular disease, and respiratory disease than the other plants. Modified restrictions were currently in place for the community, but the state let industrial facilities remain open to the extent possible.

FORD: The case rates and death rates in the Chicago area were some of the highest in the country. There was significant risk of community transmission, as social distancing guidelines had not been well enforced.

We used a compartmental epidemiologic (SEIR) model coupled with binomial probabilities to analyze the potential impact of factory openings on COVID in factory workers. We assumed that those with active infection would not be permitted to enter the factory, so based on the limited literature on population prevalence rates and rates of asymptomatic infections, we conservatively estimated the number of asymptomatic and pre-symptomatic cases was equal to the reported number of confirmed infections over the past two weeks. Based on industry anecdotes, we assumed that the workforce would be reduced to allow for social distancing, but even with a reduced capacity assumption, it was clear that the Ford plant was almost certainly going to have COVID cases in their factory.

Probability of Having Least one Asymptomatic Infected Worker in the Factory


Fig 1: Probability of having at least one infected, asymptomatic worker in automotive factory

The results of our analysis were confirmed, when on Wednesday May 20th, only two days after opening Ford Motor closed its Chicago Assembly plant twice in less than 24 hours after two workers tested positive, despite having passed temperature checks and health screenings. In the intervening time, additional workers at Ford and other auto assembly plants have tested positive for COVID. This underscores the importance of creating a work environment that minimizes the impact of having COVID positive employees on the factory floor.

COVID risk in large workplaces is akin to cyber risk. It is a question of when, not if. By constantly monitoring relevant data, instituting appropriate precautions, and creating a resilient work environment companies can reduce the impact of COVID.

Due to the high rates of asymptomatic and pre-clinical transmission, for every COVID case that gets identified in a workplace, it is likely that there is another unidentified infection. For this reason, it is critical to reduce the number of contacts employees have and the probability of transmission given a contact. Additionally, employers should be making plans to implement robust testing and contract tracing strategies, as well as, proactively modeling employee risk and monitoring the changing conditions on a daily basis.

To demonstrate the impact of social distancing and personal hygiene, we modeled the average case rate after one week with extensive social distancing and hygiene precautions versus business as usual. For the social distancing scenario, we reduced the number of contacts an employee has during a day, including household and commute to 20 and assumed that physical barriers are erected, factories are reconfigured to reduce contact between teams, and personal protective equipment and hygiene standards are enforced.

The Impact of Social Distancing of Case Rates over the Course of One Week


Fig 2: The impact of social distancing to COVID-19 case rates

As the analysis indicates, draconian measures reduce case rates, but they also impact production. As case rates rise and fall, operations should be altered and factories reconfigured to maximize the productivity to risk ratio. Human capital is what makes businesses run and the value of a business is often directly proportional to staffing and interactions between people. The only COVID risk-free building is one with no people, so companies will have to manage the ongoing tension between productivity and risk.

Each factory is unique and maximizing productivity requires continually synthesizing a variety of data sources including case counts, testing statistics, mobility data, employee demographics, and economic productivity information in order to adapt operations for the ever-changing conditions.

Additionally, the technology around testing and contract tracing is changing quickly, so although testing protocols may not make sense at the current price points and test sensitivity thresholds, factories should be planning for routine on-site testing and developing testing strategies that include both active infection and antibody testing.
Personal mobility data gathered in the workplace can help to quickly identify and mitigate areas that are particularly risk-prone and to facilitate contact tracing in the event of infection. This data coupled with machine learning algorithms can help optimize operations or reduce risk as conditions change.

Workforce capacity planning must include strategies to account for two-week absences of groups of potentially exposed employees, so that factories do not lose all of their employees with certain skill sets simultaneously.

Most pandemic viruses become endemic, so there is a high probability that COVID is here to stay. In order to address this, businesses should be planning for human capital resilience for the long-term.

Dr. Maura Sullivan is an adviser to One Concern. She specializes in risk quantification and emerging technology. Maura was the co-founder and COO of Fathom5, developing secure analytics for industrial applications. Before that, she was the Chief of Strategy and Innovation at the US Department of the Navy, responsible for institutional adoption of AI and emerging technology. Maura started her career at the global catastrophe risk company, RMS, leading the development of models and software for managing complex life and health risks for the financial and insurance markets. She was a White House Fellow, has a Ph.D. in epidemiology from Emory University, and a B.S and M.S. in earth systems from Stanford University focused in energy engineering and climate modeling.

About One Concern

One Concern is a Resilience-as-a-Service solution that brings disaster science together with machine learning for better decision making. With operations in the US and Japan, the company quantifies resilience from catastrophic perils, empowering leaders to measure, mitigate, and monetize risk so disasters aren’t so disastrous.