Great topic. Speaking generally (and not meaning to comment on any particular patient or situation), let me make several points that I think are important with regard to antidepressant medications (many readers know much of this, but others may not). Note that none of these comments is intended to address antidepressant prescribing for children or young adolescents, which is a separate (but related) topic.

1. There are many different pharmaceutical "families" of antidepressants. Several of the most prescribed have about the same [U]statistical[U] effectiveness given a few thousand patients, but have widely differing effects in individuals (largely for pharmacogenetic reasons which we are beginning to understand fairly well in this group of drugs).

2. Antidepressant medications are best indicated for clinically severe depression, particularly some specific depressive diagnoses. They are not intended, and should rarely be used, for temporary and situational stresses (such as divorces or "life stresses"), normal grieving, milder depressions, and the like. Severe depression is a serious medical illness, with a substantial mortalilty and morbidity rate. It should be taken very seriously. There is a very useful local and national advocacy group for people/patients in this latter category, the Depressive and Bipolar Support Alliance (GOOGLE them).

3. Given 1000 or so patients with medication-eligible severe depression (such as a DSM diagnosis of "major depressive disorder") which has not been treated before, a fairly large number (up to 70% in some studies, fewer in others) will respond well to the first carefully-chosen antidepressant drug that is prescribed. That is, they will have a complete or substantial reduction in depressive symptoms within one to two months. As you pointed out, the studies that suggest higher response rates make the points that (a) the clinical evaluation is adequate; (b) the clinician is psychiatrically trained; (c) the medication is logically chosen; (d) the patient is followed reasonably closely for both treatment response and side effects, and often offered specific non-drug therapy as well; (e) the patient adheres to the prescribing schedule, etc. When any of these is/are left out, the results are not as good.

4. If the first medication chosen is not sufficiently effective, changing antidepressants and/or adding synergistic or augmenting medications (usually done later in relatively unsuccessful treatment) is likely to improve response.

5. As with most medications, some of the most common reasons for lack of effectiveness are (a) not taking it as prescribed (sometimes because of side effects, but often not), (b) prescribing of too low a dose (most common when done by non-psychiatrists), (c) stopping the medication prematurely (these drugs often take weeks to work, and once they work usually should be taken for a long time and only carefully discontinued in order to watch for a return of symptoms), and (d) choosing the wrong antidepressant (we used to think they were pretty generic in their effect, and that is indeed the case in many patients, but good studies during the past decade or so have revealed some very useful information about individual patient absorption, excretion, metabolism, and neurotransmitter characteristics that allows carefully-tailored prescribing. The science and genetic testing are fairly new, but not terribly expensive. They are available, though not yet the standard of care for most patients [in my opinion]. For a variety of reasons, many psychiatrists and most nonpsychiatric clinicians do not yet use these tools for most patients.)

6. From the above paragraph, it is obvious that clinicians (medical or nonmedical) should not advise patients/clients to stop their antidepressants without an appropriate evaluation (which may well include talking with the prescribing physician). In particular, nonmedical therapists and counselors should not advise patients/clients to stop any medication without consulting a qualified physician.

7. There are a couple of non-medication psychotherapy approaches that have proved very effective in some patients with severe depression: Cognitive behavioral therapy (CBT) and interpersonal therapy (IP). Some studies indicate effectiveness approaching that of properly-prescribed antidepressant medication (though, overall, using both good psychotherapy and the right antidepressant is often best). It is important to note that although the principles of these therapies seem simple, both are specialized approaches that require a therapist with substantial training and experience. Using a few cognitive-behavioral procedures in ordinary counseling, for example, is generally insufficient for substantial depression.

8. There is at least one other biological approach that is often very effective and should be considered in very serious depression: electroconvulsive therapy (ECT -- please don't call it "shock treatment," since there is very little "shock" to the brain and the brain response -- which occurs without a physical "seizure" in modern ECT -- is the relevant part of the treatment). In recent years, some other kinds of stimulus-based treatment have been explored (e.g., vagal nerve stimulation, transcranial sub-convulsive stimulation), though none has a better combination of clinical research, efficacy, and safety than ECT. Antidepressant medication and specialized psychotherapy can be an important part of post-ECT follow-up.

9. One consideration when deciding which of the above approaches to use is rapidity and predictability of effect. That is, when patients are suicidal, greatly disabled, and/or in great emotional pain, one wants relief to come as soon as feasible. (In situations of substantial risk, such as suicide risk, close monitoring of the patient and treatment response is important during the initial period regardless of treatment choice.) Almost all antidepressant drugs take weeks to reach their full effect (assuming they are going to be helpful in the first place). CBT and IP may act more quickly in some cases (and include the advantage of frequent clinician visits). ECT works faster than the other treatments and is often an excellent choice when risk is high, though it is more clinically intensive (as are many other treatments for acute and serious medical conditions).

10. Antidepressants are over-prescribed by some clinicians, but it is important to realize that serious depression is a very painful, often disabling, often dangerous condition. The increase in recognition of severe, potentially medication-responsive depression probably outweighs the general issue of over-prescription by some doctors. Nevertheless, it is important that prescribing be done by experienced docs after an adequate evaluation. (I prefer psychiatrists, but realize that many primary care physicians and psychiatric PAs & CNPs do a good job.) If in doubt, ask for a psychiatric consultation or second opinion.

11. The media flap a couple of years ago about antidepressant prescribing being associated with slight increases in suicide risk was incredibly misunderstood and overblown. It is true that for some very depressed patients, suicide risk increases during the early phase of recovery (largely because the patient becomes more active); however, both clinical experience and clinical research have firmly established that antidepressants did not deserve the "black box" warning required by the FDA a few years ago (and there is good evidence that suicide rates in adolescents increased as a result of those warnings, because docs prescribed fewer antidepressants for people who needed them).

12. To address your question of ethics and legality: In general, in my view, it is unethical for a clinician not to know what the standard of his or her profession requires with regard to serious depression, to try hard to adhere to that standard of care, and to offer objective information and/or diagnostic & treatment choices relevant to a given patient or client (which may require referral to someone else). Is it "criminal"? Probably not except in extreme cases (have I mentioned that I'm not a lawyer?), but it can certainly create civil legal problems (such as allegations of malpractice or complaints to a licensing agency).

13. Finally, don't pay any attention to public media advertisements for antidepressants (or any other prescription medication, in my opinion).

Probably 'way more than you or others wanted to read, I know.