A protocol mandated test not done, lead to subject withdrawal. This means that data may not be usable for certain purposes. Obviously needs documented examination of the safety/well being and data integrity implications of such a lack of protocol compliance. Ideally the protocol should also cover what procedures are followed when a test is not done (including any remedial actions to ensure subject well being). CAPA to prevent recurrence at that site and throughout the trial at all centres. Interestingly FDA Warning Letters often cite "tests not done" as a significant problem.
If not in the protocol, then there should be a documented Process for re-screening and re-entry (see previous forum responses) and for tests not done. Your new process could be a trial specific process or a process for all trials (SOP, etc). Part of your CAPA could be to make sure that the protocol template and the SOPs are updated to include provision for these problems (re-consent and tests not done). The CAPA should cover whether the trial site(s) may also need to update their documented processes to make sure tests are done (why was this test not done), how to prevent recurrence in other subjects/trials ; and what do to when tests are not done.

These incidents should be include in your escalation process, so that this event is examined in case it is a potential Serious Breach of GCP or the protocol. You need to document this decision. Check whether there are any local requirements to inform the Ethics Committee, oversight committees, management or Regulators.

The Research Quality Association (www.theRQA.com) has a Q & As on screening and consent. The MHRA GCP Guide (2012) has a lots of information on screening; compliance; consent; etc.