Reporting that is based on an anonymous or confidential source has no special protections. The statement may still be defamatory, and the publisher of the reporting would have the standard defences available.
The elements of defamation
The elements of defamation are (Grant v. Torstar Corp., 2009 SCC 61, para. 28):
(1) that the impugned words were defamatory, in the sense that they would tend to lower the plaintiff’s reputation in the eyes of a reasonable person; (2) that the words in fact referred to the plaintiff; and (3) that the words were published, meaning that they were communicated to at least one person other than the plaintiff.
I pause to emphasize a terminological point: across the common law the truth of a statement is a separate question from whether it is defamatory (in the sense that it is disparaging or casts someone in a negative light). However, an important difference between Canada and the U.S. is that in the U.S., to make out a successful defamation action, a plaintiff needs to prove that the defamatory statement is also false. In Canada, the plaintiff does not need to plead or prove that the impugned words are false. Instead, in Canada, if the above three elements are proved, falsity and damages are presumed. In Canada, truth is a defence for the defendant to prove on a balance of probabilities.
In your hypothetical, the words are clearly defamatory, in that they tend to lower the reputation of Andy in the eyes of a reasonable person. They also clearly refer to Andy.
The source published the defamatory words when they communicated to the news outlet. And the news outlet published the defamatory words when they communicated to the world. See Grant, at para. 76: "someone who repeats a defamatory statement is no less liable than the person who originated it."
In Canada, there is no heightened standard for making out a defamation claim against the media or in respect of a public official or figure. Instead, there are a various defences that news media (or anyone, really) can advance.
After Andy makes out the prima facie case of defamation, the burden shifts to the defendant to make out a defence on the balance of probabilities. See Grant v. Torstar, 2009 SCC 61, para. 29:
If the plaintiff proves the required elements, the onus then shifts to the defendant to advance a defence in order to escape liability.
The news outlet's defences
The defences that the news outlet could attempt seem to be:
- fair comment (for matters of opinion) (see WIC Radio Ltd. v. Simpson, 2008 SCC 40), or
- responsible communication on matters of public interest (Grant v. Torstar, 2009 SCC 61).
First, for any of the factual claims, if the defendant proves they are true on a balance of probabilities, then there is no liability in defamation.
Fair comment (for opinion)
The claims that Andy is a "racist, biggoted transphobe" would have to be defended under fair comment. Fair comment is a defence that applies to expressions of opinion (not fact) that are defamatory. The elements of a fair comment defence are (see WIC Radio Ltd. v. Simpson, 2008 SCC 40):
the comment must be on a matter of public interest;
the comment must be based on fact;
the comment, though it can include inferences of fact, must be
recognisable as comment;
the comment must satisfy the following objective test: could any
[person] honestly express that opinion on the proved facts?
even though the comment satisfies the objective test the defence can
be defeated if the plaintiff proves that the defendant was
[subjectively] actuated by express malice.
Responsible communication on matters of public interest
If not true and not fair comment, the claim that Andy has been "sexually creepy with multiple women" would likely have to be defended on the basis of responsible communication on matters of public interest. (I am assuming you were just paraphrasing, and the reporting would include actual factual claims, rather than just an expression of opinion of creepiness. If you did really mean the reporting would only report creepiness, then these aspects would have to be defended under fair comment instead.) To establish the defence of responsible communication on matters of public interest, the defendant must show:
- that the communication was on a matter of public interest, and
- that the communication was done responsibly.
Regarding what it means to be of public interest, see Grant, para. 105:
To be of public interest, the subject matter “must be shown to be one inviting public attention, or about which the public has some substantial concern because it affects the welfare of citizens, or one to which considerable public notoriety or controversy has attached." ... Public interest may be a function of the prominence of the person referred to in the communication, but mere curiosity or prurient interest is not enough. Some segment of the public must have a genuine stake in knowing about the matter published.
Whether the communication was done responsibly involves consideration of all the circumstances, including:
the seriousness of the allegation;
the public importance of the matter;
the urgency of the matter;
the status and reliability of the source;
whether the plaintiff's side of the story was sought and accurately reported;
whether the inclusion of the defamatory statement was justifiable;
whether the defamatory statement’s public interest lay in the fact that it was made rather than its truth (“reportage”); and
any other relevant circumstances.
Relying on an anonymous source
One factor central to your hypothetical is the status and reliability of the source. See Grant, at paras. 114–15:
 Some sources of information are more worthy of belief than others. The less trustworthy the source, the greater the need to use other sources to verify the allegations. This applies as much to documentary sources as to people; for example, an “interim progress report” of an internal inquiry has been found to be an insufficiently authoritative source in the circumstances. Consistent with the logic of the repetition rule, the fact that someone has already published a defamatory statement does not give another person licence to repeat it. As already explained, this principle is especially vital when defamatory statements can be reproduced electronically with the speed of a few keystrokes. At the same time, the fact that the defendant’s source had an axe to grind does not necessarily deprive the defendant of protection, provided other reasonable steps were taken.
 It may be responsible to rely on confidential sources, depending on the circumstances; a defendant may properly be unwilling or unable to reveal a source in order to advance the defence. On the other hand, it is not difficult to see how publishing slurs from unidentified “sources” could, depending on the circumstances, be irresponsible.
In your hypothetical, you say the source is anonymous. I understand this to mean only that the news outlet has chosen to keep the identity of the source anonymous in its reporting but that the source is known to the news outlet. As the Supreme Court said in Grant, this could be a factor weighing against responsible publication. The ultimate success of the defence would depend on what else the news outlet did as part of responsible journalism before publishing the impugned statements.
If you meant that the source was anonymous even to the news outlet, then the news outlet had no way to verify the identity and relationship of the source to any element of this story. This would greatly undercuts the responsibility of their reporting. No one factor is determinative, but this would strongly weigh against a successful defence. Of course, if the anonymous source was just the starting point, and the news outlet did independent research to confirm the facts it reports, that would be a wholly different circumstance.
I understand you are only asking about the news outlet's liability, but as explained above, the anonymous source also published the statement to the news outlet, so would be a viable defendant. It is also typical to name the author or specific employee who communicated the statement as a defendant. The identity of these parties might not be known until after some discovery, but one can apply to the court to add a defendant after partial discovery.
If this were a federal claim (it isn't), then the journalistic sources protection regime under the Canada Evidence Act would apply and a court would not order disclosure of information about the source unless "the information cannot be produced in evidence by any other reasonable means" and "the public interest in the administration of justice outweighs the public interest in preserving the confidentiality of the journalistic source." This scheme has not been implemented in the provinces, so in a defamation suit which is brought under the common law, journalists could not rely on the CEA protections. Instead, journalists would have the burden of establishing a case-by-case privilege under the "Wigmore" criteria (see Globe and Mail v. Canada (Attorney General), 2010 SCC 41).