New research, published in the American Journal of Public Health, reveals a reduction in flu cases when access to paid sick days is made available in the workplace. As you might imagine, it would be difficult to quantify the impact of a policy that hasn’t yet been implemented, so the researchers simulated their population of interest—1.2 million “agents” in Allegheny County, in this case—to evaluate the transmission patterns of influenza in workplaces under different scenarios (see below for details of their methods, which may be of interest to HIA researchers).
In a baseline simulation, access to paid sick days was inequitably distributed, as it is in reality. Data from the Bureau of Labor Statistics in the United States shows that only 53% of employees in small workplaces (workplaces with fewer than 50 employees) compared with 85% in large workplaces (workplaces with 500 or more employees) have access to paid sick days. A larger percentage of employees with access to paid sick days stayed home than did employees without paid sick days—both for an average of 1.7 days when sick. They then compared results to the estimated number of flu cases that might result under two alternative scenarios: (1) all employees had access to paid sick days (a universal paid sick days policy); and (2) all employees had access to one or two days when they could stay home from work and be paid to recover from the flu (a “flu days” intervention).
They defined flu days as additional days layered on top of existing paid sick days policies. Flu days were conceptualized as an intervention educating employees to stay home for an additional day over and above what they might anyway. Currently, employees stay home for 1.7 days on average with the flu. With a flu day, they would stay home for 2.7 days on average. In comparison, universal access to paid sick days increases the proportion of people staying home, but (according to their assumptions) does not increase the duration for which employees stay home. In short, universal access to paid sick days increases the probability of workers staying home when ill, and flu days increase the time they spend at home when they are infectious.
This study found that universal access to paid sick days would reduce flu cases in the workplace by 5.86 percent and a “flu days” intervention would reduce cases by 25.33 percent. Together, universal paid sick days and flu days would equitably and effectively reduce influenza infections in workplaces. The universal paid sick days scenario was estimated to be more effective for small workplaces while “flu days” would lead to fewer flu cases in larger workplaces.
It is as important to test an intervention’s impact on health equity as it is to test its effectiveness. Universal access to paid sick days would enhance equity in the workplace by leveling the playing field in terms of access to resources. Additional interventions that promote employees staying away for more than 1.7 days on average could then be layered on to effectively reduce disease further.
The US Centers for Disease Control and Prevention recommends that people with influenza stay home for 24 hours after their fever has resolved. However, not everyone is able to follow these guidelines--many more workers in small workplaces than in large ones lack access to paid sick days and hence find it difficult to stay home when ill. These simulations show that allowing all workers access to paid sick days would reduce illness due to workplace transmission—fewer workers get the flu over the course of the season if employees are able to stay home and keep the virus from being transmitted to their co-workers. These findings make a strong case for paid sick days. Future research will be examining the economic impacts of workplace policies.
More on the simulation methods
Agent-based modeling was used to simulate the population of Allegheny County so that they could assess the impact of a paid sick days policy. This is a quantitative technique that is becoming increasingly popular among health behavior researchers, and may be useful to HIA researchers as well. In this case, authors of the study used the Framework for Reconstructing Epidemic Dynamics (FRED), a platform developed at the Public Health Dynamics Laboratory, University of Pittsburgh, to simulate the 1.2 million population of the county. Other simpler platforms such as NetLogo have been developed to enable researchers and practitioners to get started with simulating a population of interest.
A “synthetic population” of Allegheny County was developed by RTI International based on data from the American Community Survey, LandScan USA, and the census; populations for counties and states in the United States and are freely available. This synthetic population ensured that the model reflected reality in ways that were important to this study and to infectious disease spread. For example, agents are assigned to schools or workplaces based on location, size of schools/workplaces, and commuting patterns; and each agent had characteristics including age, sex, race, employment status, household income, and household location. Health information such as health status on each day, infectiousness, and susceptibility were associated with each agent in FRED.
During each simulated day, children went to school and working adults to work. They interacted with other agents who shared the same social activity locations, and returned home at the end of the day to interact with others at home. Each weekday, this routine repeated. Agents had a probability of disease transmission during interactions with others based on parameters from published studies.