Premises:
Industrial accidents are more common when some people in safety-sensitive jobs have drinking problems than when none do.
Even after treatment, people with a history are more likely to have drinking problems again

Conclusion: To reduce the risk of accidents bar anyone who has ever been treated for a drinking problem from holding a safety-sensitive job.

We need to weaken it.

Options (B), (C) and (D) seem to be causing trouble.

B.Many accidents in the workplace are the result of errors by employees who do not hold safety-sensitive jobs.

"Many accidents" are caused by employees who do not hold safety sensitive jobs. But we are given that Industrial accidents are MORE common when some people in safety-sensitive jobs have drinking problems than when none do. So the accidents become MORE common in our context. It doesn't matter how many take place in other circumstances.

C.Workers who would permanently lose their jos if they sought treatment for a drinking problem try instead to conceal their problem and continue working for as long as possible.
Our conclusion says that we should bar anyone "who has ever been treated". It doesn't say bar anyone who has a drinking problem (which might be far more difficult to identify anyway). If there is a policy in which workers who take treatment are barred from their jobs, workers with drinking problem may decide to not seek treatment at all. In that case, the risk of accidents will not reduce. Hence this is the answer.

D.People who hld safety-sentive jobs are subject to stresses that can exacerbate any personal problems they may have, including drinking problems.

Safety sensitive jobs could worsen the drinking problem. So that is more of a reason to not hire people with drinking problem (mind you, not those who have sought treatment) on such jobs. It doesn't weaken our conclusion.

Answer: C