Tuberculosis is caused by the tubercle bacillus. In order to get a confirmed diagnosis of tuberculosis, the bacillus must be found in the tissue. Other signs/symptoms, like apical opacities on the chest x-ray and cough, confirm the suspicion of TB, and may serve as "presumptive" critieria if the tubercle bacillus cannot be cultured from tissue, but if you ask any physician or read any general medical text, you will be told that the tubercle bacillus is the cause of TB. So one might ask: "Oh, doctors, why doesn't everybody exposed to the tubercle bacillus get TB?" Reasons like insufficient exposure and a strong immune system are cited in response to that question. Therefore, exposure to the bacillus is necessary, although not sufficient, for the development of tuberculosis. An example of a necessary but not sufficient causative agent.
Given that a cause can be necessary although not sufficient, let's look for cases where other combinations of necessary or sufficient agents are deemed causative by medical science. Please consider the case of smoking and lung cancer. An oft-cited statistic is that smoking causes about 90% of lung cancers. This figure includes former smokers, even those who quit decades before diagnosis. We know that smoking is not sufficient to cause lung cancer because 80% (or so) of smokers do NOT get lung cancer. (I am not a paid spokesperson for the tobacco industry; please bear with me here.) We also know that smoking is not necessary to develop lung cancer (of any histological type) because 10-13% of people diagnosed with lung cancer are never-smokers (disregarding the remote smoking question for the time being.) So here is a case of medical science concluding causality from a condition that is neither necessary nor sufficient to develop the disease.
I have asked physician colleagues what the criteria for causality are. Why things like gender, age, racial heritage, and so forth are treated as nuisances (confounders) versus causes when they are plainly linked to the disease. I've never gotten a satisfactory answer, or at least one that was logically sound to me. If only men are susceptible to prostate cancer, then having a prostate gland is a necessary, but not sufficient, condition for prostate cancer. If medical diagnosis is driven by logic, and if other medical causes are necessary but not sufficient, then what is the logical basis for excluding male gender from the list of causes of prostate cancer, particularly when we know that male hormone deprivation shrinks prostate tumors? Take away the "cause" and the problem becomes solvable. Maybe it's just too obvious that to have prostate cancer, you must have a prostate gland. Or perhaps demonizing maleness is anathema to medical hegemony, which is based in male power and domination. Perhaps we cannot consider maleness a problem in the way we regard a bacillus. We never have done so, although the medical machine does have a long tradition of treating most female ailments with removal of the uterus. (For more information on this statement, search the etymology of the word "hysteria.") I say this knowing that about 50% of young physicians are women - nothing kept them from joining up with the winning side. That wouldn't change their gender, now would it?