Every employer of CDL drivers who operate vehicles requiring a CDL has to run a random drug and alcohol testing program under 49 CFR Part 382. Not having one, or not actually running it, counts against you in a New Entrant audit and in compliance reviews.
The 2026 minimum rates
For 2026, FMCSA set the minimum annual random testing rates at:
- 50% of the average number of driver positions for drug testing
- 10% of the average number of driver positions for alcohol testing
These are minimums. You can test more, but not less. Example: a fleet averaging 10 drivers needs at least 5 random drug tests and 1 random alcohol test during the year.
Who has to be in the pool
Every driver who operates a CDL-required vehicle has to be in a random pool, including the owner if the owner drives. Owner-operators with no other drivers can't run a random program alone. They have to join a consortium (C/TPA) that runs a pool for them.
How selection works
- Scientifically valid method. Selection has to use a random number table or computer-based generator matched to driver IDs, not names drawn from a hat (382.305).
- Equal chance every time. Each driver has the same chance of being picked in every selection, even if they were picked last time.
- Spread through the year. Tests have to be reasonably spread across the year. Testing everyone in December doesn't count.
- Unannounced. Drivers can't know when a random selection is coming.
After a driver is selected
- The driver goes to the collection site immediately after being notified. Any delay must be because of the job itself, like finishing a load.
- Random alcohol tests can only happen just before, during, or just after the driver performs safety-sensitive work.
- If a selected driver is unavailable (on leave, for example), document why and select an alternate before the selection period ends.
What you have to keep
- The selection list and how it was generated
- Dates of selection, notification and testing
- Test results, in a separate confidential file
- Documentation for any selected driver who wasn't tested
- Annual summary data (the MIS report), which FMCSA can ask you to submit
Where small fleets get burned
- Running a pool on paper and losing track of who was selected
- A consortium dropping a driver from the pool after a lapsed payment
- Owner-operators who think they're exempt because they "only drive themselves"
- No record of why a selected driver was never tested