(dissenting in part). Most respectfully, I suggest that the court’s holding accords with neither fact nor law. Worse, in disturbing basic legal doctrine, it relies almost entirely on Centeno v City of New York (40 NY2d 932, affg 48 AD2d 812), which, as I propose to show, is an aberrational decision to which we should not defer (see People v Hobson, 39 NY2d 479, 487-488).
In this medical malpractice action for wrongful death, brought on behalf of the widow and children of Harold B. Topel against Dr. Harold Levinson, a psychiatrist, a classical prima facie case was established.1 A highly qualified specialist in this field of medicine, called as an expert witness by the plaintiff, testified, on the basis of record facts, that the treatment rendered by the defendant deviated from accepted medical, practice and that the lack of reasonable care inherent in this departure proximately caused the suicidal death in this case. The single expert produced by the defendant testified to the contrary. The issues sharply drawn, the case was submitted to a jury, which returned a verdict for the plaintiff.
Nevertheless, on the defendant’s motion, the Trial Judge, in a lengthy memorandum which, while it ruled in terms of whether plaintiff had made out a case on the law, in substance presented an argument more akin to one on the facts, dismissed the complaint. On appeal from the ensuing judgment to the Appellate Division, Second Department, all four Justices who sat on the case, after expressly noting their “disagreement with the trial court’s analysis of the testimony of plaintiff’s expert witness”, also took the pains to state that “The weight to be accorded the testimony of an expert witness is a matter within the province of the jury and not the trial court” (76 AD2d 862). However, other than to so comment, that court, obviously constrained by Centeno, to which it cited to explain its ruling, and essaying no independent analysis, echoed the *686Trial Term’s decision that plaintiff had not made out a prima facie case.2 By leave of this court, the case is now here pursuant to CPLR 5602 (subd [a], par 1, cl [i]). On the analysis which follows, I believe a reversal is now required.
It is elementary that, in deciding whether a prima facie case was made out, we must take the proof most favorably to the plaintiff. Embarking on that task at once, I start with the evening of March 14, 1972, when the decedent, a New York City detective, was brought to the hospital’s emergency room by his wife with the history that he twice that day had attempted to take his life by strangulation.3 Examinations that night, first by the doctor at the emergency room and then by the hospital’s attending psychiatrist, brought successive diagnoses of “agitated depression” and “psychotic depression”. The second physician then communicated with the defendant, who accepted responsibility for the care of the decedent as a private patient. It is not denied that he was advised of the patient’s suicidal manifestations.
On March 15, the morning following the patient’s admission, after familiarizing himself with the medical record, Dr. Levinson examined his patient for the first time. His entry on the chart described his patient as “a depressed, paranoid, agitated, suicidal 48-year old policeman with obsessional, delusional delirium”. He prescribed three different drugs and recommended a plan for electroconvulsive therapy. At trial, called to the stand by plaintiff’s counsel as an adverse witness, Dr. Levinson was to concede that, when he made these entries, it was in the belief that the patient’s disease was “life-threatening”. He also told the *687jury that he had instructed the nurses to see to it that the “patient was followed and seen every 15 minutes” to prevent Detective Topel from acting out his suicidal tendencies. On the same day, a nurse’s entry noted that Topel “was trying the windows”.
On his third day at the hospital, March 16, the detective’s symptoms had not abated. Nurse Paternostro’s entry for the 8:00 a.m. to 4:00 p.m. shift includes: “Patient severely depressed, frightened, T’m no use to anyone any more; I’m crazy; I’ve lost my mind.’” The nurse testified that she was so concerned about the severity of the condition that she decided to place the patient under “constant observation”. The proof goes on that this phrase means precisely what it suggests — the patient was to be under the direct observation of qualified hospital personnel 24 hours a day. The intent was to prevent self-inflicted harm. The defendant’s own entry that day was “severely depressed, paranoid, suicidal and suffering from agitated depression”.4
On the next day, March 17, the patient continued under constant observation. That morning, when he was examined, it was discovered that his blood pressure had increased and that he had developed tachycardia, an accelerated heart beat. Defendant thereupon changed the medication in the belief that the tachycardia was a side effect of the drugs initially prescribed. That evening, despite what was, if anything, an aggravation of his patient’s emotional condition, and the previously perceived need for constant vigilance, observations were reduced to 15-minute intervals.5 Dr. Levinson also was later to testify: “The patient was complaining of contact. Théy were getting him more and more delusional. He thought the nurses were driving him crazy and were going to kill him. He was getting more and more agitated”.
*688Notes on March 18 continue to track the unallayed course of the illness: “On approach [it] appears that he feels threatened, his redundancy is paramount when he feels anxious, stating ‘no one can help me here, it can’t do me any good’”. The following day, he committed suicide. While left alone in his room, he hung himself with a heavy police belt by affixing it to the doorknob of his closed door and to his neck. He had been allowed to retain and wear it throughout his stay.
On this basis, it was the plaintiff’s contention that failure to keep Detective Topel under constant observation during the acute stage of his illness, when, if anything, the clear and present danger that he would take his life was accelerating rather than receding, was a departure from accepted medical practice, and no less so because its discontinuance catered to the possible comfort or preference of the deranged patient. The proof on this score came from Dr. Manuel Trujillo, a qualified psychiatrist, whose credentials included certification by the American Board of Psychiatry and Neurology, the ongoing directorship of a psychiatric institute and professorial rank at Columbia. In response to a hypothetical question predicated on facts adduced at the trial, including the entire hospital record, Dr. Trujillo testified “that the care and treatment afforded the deceased was not in conformity with the standards of acceptable medical and psychiatric care”, “[t]his patient showed symptoms and signs of having a severe psychotic depression”, “this condition is a condition which has a very high suicidal risk” and “the proper treatment for [it] includes constant observation or other measures that insure that the patients will not carry out their suicidal impulses”. If any further particularization of the departure Dr. Trujillo was referring to was needed, it was supplied in this question and answer: “Q. And, doctor, is the failure to have this patient under constant observation from moment No. 1, when he was admitted *** is that the deviation from the accepted medical psychiatric practice that then existed in 1972? A. That is my opinion, sir”.
Persistent attempts to shake this testimony on cross-examination did not succeed, not even when counsel’s questions altered the hypotheses the witness had *689originally been asked to assume by adding or subtracting components (see Stearns v Field, 90 NY 640; 2 Wigmore, Evidence [Chadbourn rev], §682, subd [b]; §684). For example, though he agreed that in his own practice all patients whom he hospitalized and who had histories of suicidal ideations, gestures or attempts were not put under constant observation, 15-minute checks being appropriate in some cases, he was steadfast that this more limited observation was not to be condoned except “where the suicidal risk was low”. Similarly, he explained that, though he treated some patients who were suicide risks on an outpatient basis, this would only be when he “felt that the risk was not very high or was not imminent”; Dr. Trujillo had previously testified formally to what appears well-nigh conceded, that Detective Topel was “a very high suicidal risk”. And, consistently, when defense counsel specifically elicited his opinion on whether increased agitation on the part of the patient in the face of complete surveillance could support a decision to terminate constant i observation, his answer was that, “given the severity of 1 [Topel’s] condition”, it would run counter to good medical i practice to even consider doing so.
^ The legal principles applicable in this factual contest are neither complex nor confusing. The law generally permits the medical profession to establish its own standard of care. Though, in some circumstances, a physician may be held to an even higher standard, in most cases the one that will govern is accepted medical practice (Toth v Community Hosp. at Glen Cove, 22 NY2d 255, 262-263, supra; see n 2, supra). Also, so long as a physician stays within the bounds of the accepted practice, he or she is immune from liability for pure errors of judgment or for mere lack of success in his or her medical ministrations. But, when there is a departure from that standard, when the physician goes beyond its limits, the medical judgment so exercised carries no immunity against a claim that it constituted negligence (see Davis v Caldwell, 54 NY2d 176). Moreover, when the judgment exercised is outside the permissible range, “however good his intentions may have been”, a physician will not thereby be rendered free from *690liability for this “departure from approved methods in general use” (Pike v Honsinger, 155 NY 201, 210).
To establish what the existing standard is or that there has been a departure from it, because laymen ordinarily are not deemed possessed of sufficient knowledge, training or experience to have attained the competence to testify on this subject, a plaintiff nearly always will be required to produce expert testimony (see McDermott v Manhattan Eye, Ear & Throat Hosp., 15 NY2d 20, 24; Hammer v Rosen, 7 NY2d 376, 380; see, also, 7 Wigmore, Evidence [Chadbourn rev], § 2090; Prosser, Torts [4th ed], p 164; 81 ALR2d 597). Of course, in the end, the weight to be given the testimony of such a witness, as with the testimony of all witnesses, is for the jury or other trier of the facts (Meiselman v Crown Hgts. Hosp., 285 NY 389, 398; see Dougherty v Milliken, 163 NY 527).
On all this, measuring the factual showing that I have recounted by the foregoing legal propositions, it would appear that there is no possible question but that the proof here was ever so much more ample than that minimum quantum which, once adduced by a plaintiff, was supposed to create a factual issue whose resolution belonged to the jury (see 9 Wigmore, Evidence [Chadbourn rev], § 2494). Nonetheless, as far as I can make it out, in essence the majority unaccountably now holds that, no matter how much evidence is produced by the plaintiff, she cannot establish a prima facie case because there simply is no cause of action for the wrong she set out to prove. As already indicated, to justify this drastic departure from fundamental and almost universally respected decisional law, the court places primary reliance on our decision in Centeno v City of New York (40 NY2d 932, supra),6 where, without explication of our own, we rested on a brief memo*691randum of the majority at the Appellate Division (48 AD2d 812).7
Turning to Centeno, further study of that case demonstrates that the Appellate Division’s opinion there was flawed in major respects. Among others, both the single case on which its writing relied, i.e., Taig v State of New York (19 AD2d 182), as well as the three cases which Taig itself cites, i.e., St. George v State of New York (283 App Div 245, affd 308 NY 681), Warner v Packer (139 App Div 207) and Pike v Honsinger (155 NY 201, supra), are each readily distinguishable.
As preface to the analysis that proves this point, in order is a brief summary of the indisputable facts which governed Centeno. The decedent there was a diagnosed paranoid schizophrenic who, after a previous stormy sojourn in New York City’s Metropolitan Hospital, where it was decided that maintenance on thorazine, a powerful anti-psychotic drug, was a sine qua non, he was admitted to Bellevue Hospital as an emergency case on July 24, 1967. For some time prior to this admission, he had ceased taking the thorazine because of troublesome idiosyncratic side effects. The state this had reduced him to by July 24 is described in the record kept in the regular course of the hospital’s business as follows: “Brought by police, handcuffed *** agitated, violent behavior *** appears to be actively hallucinating. Talking of hearing messages from God. Speech relevant for short period, then starts screaming”. On the admission sheet, there was a stamp which said “suicidal patient” and provided boxes to be checked either “yes” or “no”. The “yes” box was marked. It was also noted that he was drug dependent. Nevertheless, despite this foreboding appraisal and a tortuous confirming course, he was moved from admission to discharge with amazing speed. Revealing in this regard is the fact that, after three days of hospitalization, his violent behavior towards hospital personnel found him placed “in seclusion”. Moreover, after four more days, on August 1, the chart showed that *692his psychotic condition had not subsided and that continued hospitalization was still necessary. Nonetheless, within two or three days, this volatile patient was discharged without medication. The hardly unexpected result was that, but a few days later, on August 9, he took his life by leaping through a window.
At the trial of the suit that ensued, a qualified psychiatrist, on this and other proof of “hostility”, “depression”, “delusional thoughts”, “hallucinatory experiences”, “loose associations” and “assaultive tendencies”, testified that Centeno’s condition made him a constant candidate for “an impulsive act which would result in a suicide, even if five minutes before he had not necessarily given it thought”. The witness, speaking “with a reasonable degree of medical certainty”, explained that, at the time of his discharge, Centeno “needed additional treatment both of a pharmaceutical nature in the form of drugs of various sorts and confinement to an institution where there is a regulation of his activities, a guard on what he might do to himself or others, until such time as he showfed] a degree of real stability” and “that this man was rather superficially treated and released much too soon for his own or anybody else’s good”.
On these same facts, he expressed the opinion that, “in accordance with the accepted medical practice and accepted psychiatric practice in July 1967”, “it is very much my opinion and I can’t state it too strongly that [this] highly agitated, disturbed, delusional, hallucinating individual should not under any circumstances * * * [have been] released from the hospital with or without medication after such a brief stay” and “certainly not discharged * * * from the hospital with no medication and no possibility of suitable follow up” (emphasis supplied).8 It was on such a record plus proof of causation that defendant’s motion to dismiss was granted.
Getting back now to the Appellate Division’s affirming memorandum in Centeno, I begin with a verbatim state*693ment of the first proposition it puts forward: “There is not an iota of evidence to indicate that [the discharging psychiatrist’s] opinion was other than honest. It is therefore completely irrelevant that plaintiff-appellant’s expert testified as is set forth in the dissent. Disagreement between professional experts does not in these circumstances provide the basis for a holding that a prima facie case of malpractice was presented requiring jury consideration.” This, of course, ran completely counter to the developed doctrine which has prevailed in this area of law from the seminal Pike case down (Pike v Honsinger, 155 NY 201, 211, supra [“however good his intentions may have been”]). The Centeno memorandum’s second proposition, quoting haec verbae from Taig, after stating the obvious, i.e., that “decision to release the patient from the hospital * * * [was a] professional medical judgment” presumes to pronounce that “Although another physician might disagree as to the form and period of treatment to be followed, a liability would not arise *** if the professional judgment to discharge him was in fact erroneous”. This sanctification of medical judgment per se, irrespective of whether it is good, bad or indifferent; or whether it accords with or is opposed to good medical practice; or whether it results from want of requisite knowledge or skill; or whether it flows from a failure to make a diligent medical evaluation, would set medical malpractice law, if not all professional malpractice law, on its head (see Pigno v Bunim, 43 AD2d 718, affd 35 NY2d 841). Besides, in Centeno, as in Topel, there simply was no attempt merely to pit one medical judgment against another. By no means was the expert testimony in either case confined to averments that the treatment rendered, though founded on judgment, was in error. Rather, it was that the premature discharge of Centeno, and the failure to constantly observe Topel, if indeed not patently unreasonable, deviated from accepted psychiatric medical practice.
Hieing then to the three authorities on which Centeno leans (via their incorporation in its quotation from Taig), we look in vain for confirmation of the pronouncements the Appellate Division published. In the first one, St. George v State of New York (283 App Div 245, affd 308 NY 681, *694supra), a claim was brought against the State for alleged negligence in releasing a psychotic Matteawan State Hospital patient who subsequently killed four people. The Court of Claims Judge who heard the case, which, like all trials in that court, proceeded without a jury, held the hospital liable, essentially on factual findings of inadequate investigation of the patient’s medical condition. On appeal, for its part, the Appellate Division, on review of the record, arrived at an opposite factual determination, as it is empowered to do (CPLR 5501, subd [c]; Cohen v Hallmark Cards, 45 NY2d 493, 498). In affirming the Appellate Division’s order, without opinion, we did no more than choose between the opposing factual determinations (NY Const, art VI, § 3, subd a; CPLR 5501, subd [b]). Thus any argument that St. George undermines the existence of an issue of fact when a medical judgment’s soundness is challenged by competent proof that it is dehors acceptable practice is bootless.
The second cited case, Warner v Packer (139 App Div 207, supra), provided no greater succor for Centeno. There the plaintiff brought suit for damages against medical examiners who, she alleged, negligently certified her as insane, as a consequence of which she for a time had been wrongfully detained in a mental institution. True, the Appellate Division held that no jury question had been presented. But it was because the plaintiff did not introduce any expert proof of the existence of a standard with which the conduct of the defendants could be compared (Warner v Packer, supra, at pp 214-215). Warner, then, is antithetical to the suggestion that the expert testimony in Topel, or that which was introduced in Centeno, had not provided the necessary link for a prima facie case.
The last of the three Taig-Centeno citations is to Pike v Honsinger. The wellspring of modern malpractice law, Pike gave a central place to proof of “departure from approved methods” in its determination that a case of medical malpractice was made out and, as discussed earlier, was destructive of Centeno’s “honest opinion” argument. One cannot help wonder then whether its citation was pure ritual. Oh, yes, it does make reference to immunity “for a pure error of judgment”, but it does so in two pages which *695make it crystal clear that, like an “honest opinion” or “good intentions”, this is only the beginning and not the end of the inquiry on which liability must come to rest.
None of the Taig cases lending support to Centeno’s conclusion, I also focus, for the purpose of completeness on Centeno’s remaining argument, a continuation of its quotation from Taig. It consists in the main of a number of nondecisive platitudes: “the prediction of a future course of a mental illness is a judgment of high responsibility”, “in some instances it involves a measure of calculated risk”, “if liability were imposed each time the prediction of future course of mental illness went wrong, few releases would ever be made” (emphases added). To put these comments in perspective, it must be remembered that Taig was a Court of Claims nonjury case and that the Appellate Division, in affirming, was dealing with an appeal on which it was reviewing facts.
Because there was no divergence between the Court of Claims and the Appellate Division, the case never came to our court and so we had no occasion to countercomment. But, had we occasion to do so, many queries could readily have come to mind: If a mental illness’ future course is a “high responsibility”, why not all the more reason to hold the one who carries it to strict conformity with the line drawn by accepted medical practice, representing as it does a consensus of his fellows’ practical experience as an appropriate minimum norm? If “a measure of calculated risk” may be involved only “in some instances” does not a factual determination as to whether in any particular “instance” the exercise of judgment falls outside the bounds of accepted use become all the more appropriate? Is emphasis on the element of predictability as to the future course of mental illness misplaced when all a psychiatrist is asked to do is to take accepted precautionary measures to guard against given risks (e.g., Topel and Centeno, where the risks of suicide not only existed but were labeled for all who had “high responsibilities” to see)? How can one hypothesize the consequences of imposing liability “each time” a judgment as to the future course of a mental illness “goes wrong”, when there is no such thing as an “each time” imposition? And, if behind all this lurk other ques*696tions — medical, social or economic — why are these not more appropriately considered by the Legislature, with facilities and penchants for factual investigation and experimentation not available to the judiciary (see Simcuski v Saeli, 44 NY2d 442, 456 [concurring opn])?
Inspection of the visible foundation stones of Centeno having thus disclosed their lack of substance, it may not be amiss to note that its precedential value is further eroded by a glaring omission. It completely ignores a long line of cases which recognize a duty to protect a patient with known self-destructive tendencies from himself (see, e.g., Hirsh v State of New York, 8 NY2d 125; Daley v State of New York, 273 App Div 552, affd 298 NY 880; Martindale v State of New York, 269 NY 554; Comiskey v State of New York, 71 AD2d 699; Gries v Long Is. Home, 274 App Div 938; Callahan v State of New York, 179 Misc 781, affd 266 App Div 1054; Robertson v Towns Hosp., 178 App Div 285). This adds but another reason why I cannot see how the court in the present case, while acknowledging that the “line between medical judgment and deviation from good medical practice is not easy to draw” [for the physician or for the Judge?] (majority opn, at p 684), finds it possible, in a medical situation in which five minutes of supervisory inattention may make the difference between life and death, to undertake, as a matter of law, to superimpose its own medically and psychiatrically unqualified opinion that something less than complete supervision accorded with accepted medical practice.
Finally, because Centeno rears its head, it is with utmost respect that I exercise the privilege of suggesting that, sensitive though a court must be to considerations of stare decisis, this is an appropriate case in which to apply the wisdom of the question we put and the answer we supplied on another occasion: “Which is the stare decisis: The odd cases or the line of development never fully criticized or rejected? * * * Of course, it would be foolhardy not to recognize that there is potential for jurisprudential scandal in a court which decides one way one day and another way the next; but it is just as scandalous to treat every errant footprint barely hardened overnight as an inescapable *697mold for future travel” (People v Hobson, 39 NY2d 479, 487-488, supra).
For all these reasons, the order appealed from should be modified by reversing the judgment in favor of the defendant Levinson and remitting the pase to the Appellate Division for consideration of the facts.
Chief Judge Cooke and Judges Jasen, Gabrielli, Jones and Meyer concur in memorandum; Judge Fuchsberg dissents in part and votes to modify in an opinion; Judge Wachtler taking no part.
Order affirmed.
. For reasons which include those stated by the majority in its memorandum decision, I make no issue of the dismissal of the case against the hospital. Therefore, all my references to the defendant are to Dr. Levinson alone.
. It also cited our decision in Toth v Community Hosp. at Glen Cove (22 NY2d 255), but Toth, which, unlike Centeno, did not involve a suicidal psychiatric patient, merely states the general propositions of medical malpractice law on which all parties here agree. Moreover, in Toth, the jury, unlike in the case before us now, found as a fact that the physician acted in accordance with accepted medical practice. So our opinion there dealt, in the main, with the discrete question of the hospital’s liability and with whether, as to a second physician, the case should have been submitted to the jury on the alternative theory that a physician may be held to even a higher standard than accepted practice if he possesses “superior skill, knowledge or experience”.
. As other trial evidence indicated, in the period preceding his hospitalization, Detective Topel was on sick leave for extreme agitation, insomnia, weight loss and obsession, with fears that he would be discharged from the police force and bring disgrace upon his family. On March 11, three days prior to admission, he had told his wife, “I am going to suffocate myself” and “It is easy to go out of a window * * * it would be over quick”.
. At the trial, Levinson claimed that the constant observation was instituted as a result of his telephone order to the nurse. However, the nurse testified that it was she who made the decision to place Topel on constant observation and that she did not remember any such telephone order. Levinson’s claim was also contradicted by his prior testimony that he never ordered “constant observation”. The medical chart included no such telephone order.
. Neither at that time nor at any subsequent time before Topel died did Levinson ever see a report of a follow-up electrocardiogram he had ordered.
. The court also cites to Cohen v State of New York (41 NY2d 1086, affg 51 AD2d 494) and Johnson v Yeshiva Univ. (42 NY2d 818), but these cases have little, if any, relevancy here. In Cohen, the only issue presented for our review was the correctness of the Appellate Division’s reduction of the damage award. In Johnson, we simply held that no question of fact was raised when the only showing was that the use of a particular medical test was not a departure from accepted medical practice; in Topel, as we know, Dr. Trujillo expressly disputed Dr. Levinson’s conclusion that the treatment conformed to accepted practice.
. In Centeno, my dissenting memorandum emphasized “that the fact that a departure from accepted medical practice occurs in a psychiatric rather than in a nonpsychiatric setting is only an element to be weighed along with all the other circumstances in a particular case and is not the premise for an application of different legal principles” (40 NY2d 932, 933).
. In sharp contrast, when Centeno, only months earlier, had been hospitalized at Metropolitan, though, after 10 days, the thorazine treatment instituted there appeared to have reduced his symptoms, the doctors would not release him to convalescent care without first keeping him under observation for an additional two weeks to make certain that his condition was under control.