Keeping school bus drivers sober and safe on the road is, in part, the responsibility of the Federal Motor Carrier Safety Administration, which regulates commercial drivers. It used to be that employers were required to randomly test 50% of their commercial drivers for drugs each year, along with 10% for alcohol.

In 2016, however, right in the midst of the opioid crisis, the agency reduced that requirement. Now, employers only have to test 25% of their drivers for drugs and 10% for alcohol. The reason for the reduction is that employer surveys indicated that less than 1% of drivers were failing the controlled substance tests.

But still, the problem of drunken or drugged school bus drivers is real, and it’s alarming. Stateline, an initiative of The Pew Charitable Trusts, recently dug into court filings, news reports and police records to find out how common the problem is.

How common? Unfortunately, we don’t know

Stateline found at least 118 drivers across the country who were arrested or issued citations for being impaired while operating a school bus. About a third of the drivers were impaired by drugs.

Over a third of the 118 incidents involved a crash. These crashes involved injuries to nearly three dozen students, some severely enough to need a trip to the ER.

Drivers are also being caught through random testing. Stateline learned that at least 260 school bus drivers from five states had either failed random tests or refused to take them since 2015.

However, Stateline also found that no one is keeping track of this data in any systematic way. No state or federal agency is tracking incidents involving impaired school bus drivers, and that many states couldn’t even compile the information upon request.

For example, Stateline found that local school districts know about individual incidents, but no one is keeping track of the data at a statewide or national level. Stateline asked 268 agencies from all 50 states and Washington, D.C., about how many school bus drivers had been caught impaired or failing random drug and alcohol tests. Only 11% of the agencies came up with any data at all.

Of those that could, the data was often wrong. For example, some agencies found no cases in their databases even though Stateline had identified at least one. In other cases, the data was incomplete or wrong. Most agencies don’t know how many school bus drivers are failing random drug and alcohol tests, and most couldn’t even query the total number of accidents involving impaired school bus drivers.

Even one incident involving an impaired school bus driver is too many. What can we do to better track these incidents and prevent them?