Ohio train derailment: The public health angle

On February 3, a train carrying toxic and combustible chemicals derailed in East Palestine, Ohio. Of the 150 cars on the train, 50 were involved in the accident. Of those, several carried toxic chemicals on board.

Knowledge “translation” and risk communication for residents have been, at best, suboptimal. This is how us, public health experts, are thinking about this disaster.

Who’s at risk

For any health emergency, understanding what the health risks are, to whom, and when are absolutely essential. 

For chemicals, there are two main points of interest:

1. Intensity (amount of exposure)

2. Time (duration of exposure)

These are highly dependent on the timeframe of the emergency. In this case, pre-burn, burn, and post-burn—and where people were during those times. 

Pre-burn

Public health officials were most worried about chemicals that leaked out of the rail cars, and specifically vinyl chloride—a clear, combustible substance that is used to create PVC piping. It’s a gas that spreads through the air.

The best way to know if the air is safe is to measure the amount in the air, and this is commonly referred to in units of measurements of “parts per million,” or “ppm”. The table below shows the potential health problems from different short-term exposure levels.

For example, if air samples were over 250 ppm, we would expect people to experience things like headaches and respiratory irritation, but the symptoms would go away after the exposure went away. But if the concentrations in air samples went over 1,200 ppm for an hour, then there is the possibility of irreversible health effects.  

This is really helpful to understand because the EPA is sharing their air monitoring data here. But I encourage the EPA to “translate” this data for residents to understand their associated risks. This is absolutely critical. (It’s a bad sign if it took us days to decipher!)

Most measurements in the pre-burn period were taken outside of the 1-mile radius. (This is likely due to safety concerns about explosion risk in those first hours.)

Future efforts should include construct an air dispersion model to help reconstruct what happened within that 1-mile buffer of the derailment, so the residents living close to the derailment can understand what potential exposures they might have had. 

Burn

After the immediate derailment, it appears the on-the-ground decision-makers were grappling with essentially two options:

1. Don’t do anything. There were reports of concern about the tankers exploding before cleanup could begin.

2. Burn the chemicals off. Burning vinyl chloride gets rid of the vinyl chloride but creates other chemicals and particulates in its wake. 

The decision logic in the early days has not been made public. But, ultimately, officials chose #2: evacuate the residents, drain the substances from the cars into a trench, burn the chemicals, and then let the residents come back.

Post-burn

Other gasses, like phosgene and hydrogen chloride are created when vinyl chloride is burned. This would be an immediate threat but would subside over time (hence the evacuation during the burn). 

Also of concern are particles from burning chlorinated chemicals (e.g., doxins) that can become persistent in the environment. In other words, we don’t just have the gases in the air to think about, we have particles now, too. (The EPA data show an increase in airborne particle concentrations post-burn.) The gases eventually disperse in the air, and so will much of the particulates, but particles also settle on surfaces.

That means ongoing testing should not be limited to the air in people’s homes and businesses, but should also include surfaces.

One concerning sign is that residents continue to report health symptoms, like headaches. One way to understand why is to track symptoms in a harmonized, systematic manner. The CDC started this surveillance last Monday, ideally to answer: 

• What are people experiencing? 

• Are there significant geographic patterns? 

• Do the hot spots correspond with the train derailment site? 

The EPA is continuing to test outdoor air, indoor air samples and surface samples in homes. The last data publicly available, though, was from February 17. It’s impossible for a non-expert to read.

In addition, the Ohio Department of Natural Resources recently reported more than 43,000 dead fish within 7 miles of the crash. This tells us there were immediate impacts to waterways, but it’s not clear the extent to which waterways—and how far from the site—continue to be impacted by that initial release. EPA is reporting that the drinking water is safe at this time. 

Looking to past events

A past event produced a very similar environmental situation 10 years ago when a train derailed in Paulsboro, New Jersey. On November 30, 2013, four rail cars fell into a creek. One tank broke and released 23,000 gallons of vinyl chloride. 

The case study found: 

• 1 in 10 residents sought care at a hospital for symptoms. 

• Those living closer to the derailment site reported symptoms at higher frequencies than those living farther away.

• Most of the vinyl chloride was released from the tank within the first hour.

• Air monitoring detectors were used within 90 minutes of derailment. The detectors read vinyl chloride concentration of ~1,400 ppm. 

• Distance mattered. Maximum concentrations were 0.2 miles from the derailment, and concentrations were as high as 250 ppm as far as 0.8 miles from the derailment site. Thus, the 1-mile range in the current disaster is reasonable. 

While this information is instructive, it’s important to note that the events are unique, too. The burning in Ohio, for example, changes the chemistry of the substance. This wasn’t done in New Jersey. Also, there were multiple chemicals in Ohio, whereas the New Jersey derailment was limited to vinyl chloride.

What to do?

In the aftermath of the 2013 incident, the NJ Public Health Department asked residents what could have gone better. The strongest recommendation: Improve health and emergency communication: What do we know? What do we not know? And how are we finding answers? 

It doesn’t seem like these lessons were applied immediately after the current event. But the EPA seems to have done a better job once they went “all in” on the response. Knowledge translation can always be improved, though.

Bottom line

The public health implications of this disaster are not fully known. Ongoing data collection, monitoring, transparency, resident engagement, and communication will be absolutely essential in understanding and translating health risks associated with this emergency. 

Love, YLE and Dr. Joseph Allen

---

Dr. Joseph Allen is an Associate Professor at Harvard whose expertise lies on probing how air quality affects human health and cognition.

“Your Local Epidemiologist (YLE)” is written by Dr. Katelyn Jetelina, MPH PhD—an epidemiologist, data scientist, wife, and mom of two little girls. During the day she works at a nonpartisan health policy think tank and is a senior scientific consultant to a number of organizations, including the CDC. At night she writes this newsletter. Her main goal is to “translate” the ever-evolving public health science so that people will be well equipped to make evidence-based decisions. This newsletter is free thanks to the generous support of fellow YLE community members. To support this effort, subscribe below: