Another Look at PHMSA Transportation Incident Database
4-2-24
Since January 13th, I have been trying to report on chemical transportation incident as reported to PHMSA on their Form 5800.1. I have not been able to access the PHMSA database every weekend due to PHMSA’s work on the site during the weekend, so I have tried accessing it on Friday or Monday. After publishing my last week’s post, I noticed that my weeks’ worth of data only contained data for three days. When I did yesterday’s post, I paid more attention to the dates in the database and selected a range that contained seven days’ worth of data.
While this provided a near term fix, I decided to take a more detailed look at the PHMSA database so that I could better determine what would be an appropriate ‘week’ to use for my periodic reporting on the incidents in the data base. So, I started by downloading all incidents from March 1st, 2024, through April 2nd.
One Month Overview
The graph below shows the daily number of incident reports recorded in the PHMSA database for the month of March 2024 as of 10:18 a.m. EDT on April 2nd 2024. According to the search page: “Incident Source Data as of 4/1/2024”. When I first started using the database, I assumed (always dangerous) that this meant that the data was complete as of this date.
The low points on the graph for March 3rd (18), March 10th (7), March 17th (6), and March 24th (7) are all on Sundays. This would be the day of the week when we would expect the fewest number of hazmat shipments in transit, so low numbers of incidents is not unexpected. The more concerning issue is the tail at the end the month, starting on March 27th and continuing through the 31st. With just two incidents reported in that time frame would seem to indicate, at the very least, that database was not fully up-to-date for the last four days of the month.
Previous Reporting
Short of waiting for May 2nd to roll around to test this supposition, it would seem that there is insufficient data to reach this conclusion. But there are three times during the month when I collected data that can be compared back to the one month data reported above. I will start with the data from my March 16th post where I used data from the ‘week’ starting March 5th, 2024.
The data in the ‘3-16-2024’ column shows the number of incidents for each day in the data I downloaded from the PHMSA database on March 16th. The third column shows the number of incidents from the April 2nd download from the same database. The bottom entry in each column is the sum of reported incidents for the week. The fourth column shows the 3-5 data as a percentage of the 4-2 data.
The below table shows the same data used in my post of March 23rd.
Finally, the table below shows the same data used in my post of April 1st.
This shows that there are only relatively minor changes being made to the most recent data.
Directly comparing the three sets of data is fraught with analytical dangers. The percentage data in each table is based upon a different period of elapsed time between the initial data collection and the final data point. One thing is apparent though, there is a time lag between the incident date and the time the data is entered into the database.
Form 5800.1 ICR
To better understand the potential reason for the apparent delays in the incident data showing up in the PHMSA database, we can look at the Supporting Document that PHMSA provided the OMB’s Office of Information and Regulatory Affairs (OIRA) the last time that PHMSA updated the information collection request where PHMSA justified the data collection supporting this data base.
First, the regulatory justification for the submission of Form 5800.1 is found at 49 CFR 171.16. The form is available on-line. The form can be filled out electronically and submitted by clicking on the ‘SUBMIT’ button on the bottom of the page. Or the form can be printed out and submitted by mail. Paragraph §171.16(a) requires submission of reports on “a Hazardous Materials Incident Report on DOT Form F 5800.1 (01/2004) within 30 days of discovery of the incident”.
The supporting document notes that PHMSA expects (based upon recent history) that there will be “22,608 incident reports are submitted to (2,888 paper [12.8%] and 19,720 electronic [87.2%]) each year”. The delays in getting the data from the paper reports into the database are easy to understand. What is less easy to understand is the delay in getting electronic data into the database.
The Supporting Document provides the following description of the use of the data collected on the Form F 5800.1:
“The hazardous materials transportation safety program relies on DOT Form F 5800.1 to gather basic information on incidents that occur during transportation. The prescribed form provides meaningful, accurate, and comprehensive information relative to causes and effects associated with hazardous material releases. The form provides the user with a variety of describing factors leading to the release of a hazardous material.”
There is, however, no information provided by the agency on how the data is processed. What is clear from the analysis of the data reported here, is that there is no direct connection of the electronic submission of the F 5800.1 to the inclusion of the information in the database. I would assume that PHMSA provides some sort of data review process before the information is added to the database.
Practical Consequences
For the purposes of my periodic reporting on transportation chemical incidents, what are the consequences of this delayed database updating? To answer this question, let us look at what I reported on March 16th, and what that report would have looked like if I had used the data from the April 2nd download for the same period.
On March 16th I reported:
Number of incidents – 137 (127 highway, 6 air, 4 rail)
Serious incidents – 2 (2 Bulk release, 0 injuries, 0 deaths, 0 major artery closed)
Largest container involved – DOT 117R100W railcar (Gasoline Includes Gasoline Mixed with Ethyl Alcohol, with not more than 10% Alcohol) about 1-gal leaked through damaged manway gasket.
Largest amount spilled – 330-gallons (Bisulfites, Aqueous Solutions, N.O.S.) from 330-gal plastic IBC due to forklift strike.
Using the expanded data available on April 2nd, the report would have looked like this:
Number of incidents – 519 (490 highway, 18 air, 11 rail)
Serious incidents – 5 (5 Bulk release, 0 injuries, 0 deaths, 0 major artery closed)
Largest container involved – DOT 117R100W railcar (Gasoline Includes Gasoline Mixed with Ethyl Alcohol, with not more than 10% Alcohol) about 1-gal leaked through damaged manway gasket.
Largest amount spilled – 330-gallons (Bisulfites, Aqueous Solutions, N.O.S.) from 330-gal plastic IBC due to forklift strike.
While the two descriptive report elements did not change in this example, the first two reporting items did change significantly.
Conclusion
Because of regulatory incident reporting requirements for PHMSA’s F 5800.1 form and PHMSA’s internal processing of that data before it gets into their database their may be significant delays before chemical transportation incidents are recorded in that database. I have not yet collected sufficient data to succinctly describe the length of that delay.
Since the purpose of these transportation incident blog posts is to provide a snapshot of chemical transportation safety, it is clear to me that I need to provide more time for the PHMSA data processing to take place before I report on the data. While I do not have a complete picture of how long it takes PHMSA to populate the database, it is clear that I am going to have to accommodate the 30-day reporting period into my selection criteria. So, starting with this weekend’s blog post I will be using a reporting week five-weeks in the past, so that blog post will cover incidents for the week starting March 2nd, 2024.