Our threshold issue requires us to determine under what circumstances a psychiatrist can be held liable for the violent acts of a *92voluntarily hospitalized patient following the patient’s release from the hospital.
To maintain a wrongful death action on a theory of negligence, a plaintiff must show (1) the existence of a duty owing to plaintiff’s decedent, (2) a breach of that duty, and (3) proximate causation between the breach of duty and the death. Bennison v. Stillpass Transit Co. (1966), 5 Ohio St. 2d 122, 34 O.O. 2d 254, 214 N.E. 2d 213, paragraph one of the syllabus. The same three elements must be shown to establish a negligence action generally, including a survivorship action predicated upon ordinary negligence or medical malpractice. Cf. Hadfield-Penfield Steel Co. v. Sheller (1923), 108 Ohio St. 106, 114, 141 N.E. 89, 91; Baier v. Cleveland Ry. Co. (1937), 132 Ohio St. 388, 391, 8 O.O. 208, 209, 8 N.E. 2d 1, 2; Wills v. Frank Hoover Supply (1986), 26 Ohio St. 3d 186, 188, 26 OBR 160, 161, 497 N.E. 2d 1118, 1120.
We review the issues before us within the framework of the traditional duty-breach-proximate-cause analysis. The complicating factor is that this case has its genesis in the mental health field. Practitioners in the mental health field treat disturbed persons. Using the traditional concepts of tort law, we must decide how these concepts apply to such practitioners for the acts of a disturbed patient.
I
Though neither party disputes the existence of a duty of care owed to Carly by Dr. Murray, for purposes of clarity we deem it important to delineate the basis and scope of this duty and to show it is consistent with established tort principles of Ohio law.
Under Ohio law the existence of a duty depends on the foreseeability of the injury. Menifee v. Ohio Welding Products, Inc. (1984), 15 Ohio St. 3d 75, 77, 15 OBR 179, 180, 472 N.E. 2d 707, 710; Gedeon v. East Ohio Gas Co. (1934), 128 Ohio St. 335, 190 N.E. 924. We find no difficulty with the proposition that a reasonably prudent psychiatrist would have anticipated that an injury to Carly could result from his performance or nonperformance of certain acts, including discharging Theresa from the hospital without adequate precautions.
However, there is no duty under Ohio law to control the conduct of another person so as to prevent him from causing physical harm to another unless a “special relation” exists between the actor and that person which imposes a duty upon the actor to control the person’s conduct. 2 Restatement of the Law 2d, Torts (1965) 122, Section 315 (adopted by this court in Gelbman v. Second Natl. Bank of Warren [1984], 9 Ohio St. 3d 77, 79, 9 OBR 280, 281, 458 N.E. 2d 1262, 1263; Hill v. Sonitrol of Southwestern Ohio, Inc. [1988], 36 Ohio St. 3d 36, 39, 521 N.E. 2d 780, 782). Such a “special relation” exists when one takes charge of a person whom he knows or should know is likely to cause bodily harm to others if not controlled. Restatement, supra, at 129, Section 319; see Restatement, supra, at 123, Section 315, Comment c.
Even though Theresa was a voluntary patient, we find that Dr. Murray had sufficient charge of Theresa in the hospital setting such that a special relation was established.3 This is con*93sistent with our holdings in other contexts, including landholders’ liability, where we stated that “ ‘liability in tort is an incident to occupation or control.’ ” Mitchell v. Cleveland Elec. Illum. Co. (1987), 30 Ohio St. 3d 92, 94, 30 OBR 295, 297, 507 N.E. 2d 352, 354, quoting Cooper v. Roose (1949), 151 Ohio St. 316, 317, 39 O.O. 145, 146, 85 N.E. 2d 545, 546. It is also consistent with Illustration 1 under Section 319 of the Restatement:
“A operates a private hospital for contagious diseases. Through the negligence of the medical staff[,] B, who is suffering from scarlet fever, is permitted to leave the hospital with the assurance that- he is entirely recovered, although his disease is still in an infectious stage. * * * B * * * communicate[s] the scarlet fever * * * to D * * *. A is subject to liability to D * * Restatement, supra, at 130.
Therefore, we find that Dr. Murray had a duly to take reasonable precautions to protect Carly from Theresa’s violent propensities.
II
The standard of care required of a medical doctor is dictated by the custom of the profession:
“In order to establish medical malpractice, it must be shown by a preponderance of evidence that the injury complained of was caused by the doing of some particular thing or things that a physician or surgeon of ordinary skill, care and diligence would not have done under like or similar conditions or circumstances, or by the failure or omission to do some particular thing or things that such a physician or surgeon would have done under like or similar conditions and circumstances * * Bruni v. Tatsumi (1976), 46 Ohio St. 2d 127, 75 O.O. 2d 184, 346 N.E. 2d 673, paragraph one of the syllabus. A psychiatrist, as a medical specialist, is held to the stan-. dard of care “of a reasonable specialist practicing medicine or surgery in that same specialty in the light of present day scientific knowledge in that specialty field * * Bruni, supra, at paragraph two of the syllabus.
Defendants argue that the malpractice standard of ordinary care should not be applied to psychiatric decisions releasing voluntary mental patients from a hospital. They make four points in support of their argument.
First, defendants claim that psychiatrists are unable to predict their patients’ potential for violence with any degree of accuracy.4
Second, because a reasonable psy*94chiatrist of “ordinary skill and training” cannot make an accurate prediction of a patient’s violent behavior, it can never be said that a given prediction exhibited a lack of “ordinary skill.” In other words, there is no standard in the psychiatric profession with which to measure a psychiatrist’s judgment of a patient’s propensity for violence.5
Third, modern psychiatry favors the early release of mental patients.6 If a psychiatrist knows that he will face liability for failing to foresee a patient’s future violent behavior, the predictable result will be a court-mandated end to “oút-patient” treatment, and the massive confinement of all patients who display even a remote possibility of violent behavior.7
Fourth, the General Assembly, in delineating the liability standards pertaining to civil commitment, holds a psychiatrist to only a good faith stan*95dard of care for decisions to commit or discharge a mental patient.8 The standard of care required for voluntary hospitalization should reflect the General Assembly’s wisdom in formulating a standard for involuntary hospitalization.
In place of the conventional malpractice standard, defendants urge this court to adopt a “professional judgment rule.” Under this rule, a psychiatrist will not be liable for releasing a patient who subsequently harms another if, after carefully examining all relevant data, the psychiatrist makes a professional medical judgment that the patient does not pose an immediate danger to others.
The professional judgment standard has received the most attention in the New York courts where it applies to all phases of professional malpractice, including psychiatry. O’Shea v. United States (E.D. N.Y. 1985), 623 F. Supp. 380, 384, citing Bell v. New York City Health & Hospitals Corp. (1982), 90 App. Div. 2d 270, 279, 456 N.Y. Supp. 2d 787, 793. The seminal case in New York held that; “The rule requiring a physician and surgeon to use his best judgment does not hold him liable for a mere error of judgment, provided he does what he thinks is best after careful examination.” Pike v. Honsinger (1898), 155 N.Y. 201, 210, 49 N.E. 760, 762. Since Pike, the New York courts have applied the professional judgment rule to exculpate physicians, and especially psychiatrists, from liability for diagnostic decisions,9 treatment decisions,10 decisions to release hospital patients who later harm themselves or others,11 and decisions denying hospital admission.12
*96The New York courts have not applied the professional judgment rule to immunize psychiatric decisions where there was a failure to evaluate the condition of potentially dangerous patients before discharging them from the hospital, Homere v. New York (1975) , 48 App. Div. 2d 422, 370 N.Y. Supp. 2d 246 (discharge decision, based on a patient evaluation, was made forty-one days before the actual discharge date. In the intervening period the patient’s condition markedly worsened and no reevaluation of his condition was conducted before actual discharge.); Bell v. New York City Health & Hospitals Corp., supra (psychiatrist failed to inquire into a patient’s delusions and auditory hallucinations); Huntley v. New York (1984), 62 N.Y. 2d 134, 476 N.Y. Supp. 2d 99, 464 N.E. 2d 467 (suicide plan of patient not told to psychiatrist in charge of leave privileges); or where there was a failure to keep detailed and proper medical notes and, consequently, it could not be established whether any evaluation of the patient’s suicidal propensities had been made by a qualified psychiatrist. Cohen v. New York (1976) , 51 App. Div. 2d 494, 382 N.Y. Supp. 2d 128, affirmed (1977), 41 N.Y. 2d 1086, 396 N.Y. Supp. 2d 363, 364 N.E. 2d 1134.
At least one court outside New York has applied a limited standard of review to a hospitalization case. The court in Currie v. United States (M.D. N.C. 1986), 644 F. Supp. 1074, affirmed on different grounds (C.A. 4, 1987), 836 F. 2d 209, held that a “psychotherapist judgment rule” should be used to review a therapist’s decision not to seek involuntary hospitalization for an outpatient who subsequently became violent. The court stated:
“Under such a ‘psychotherapist judgment rule,’ the court would not allow liability to be imposed on therapists for simple errors in judgment. Instead, the court would examine the ‘good faith, independence and thoroughness’ of a psychotherapist’s decision not to commit a patient. * * * Factors in reviewing such good faith include the competence and training of the reviewing psychotherapists, whether the relevant documents and evidence were adequately, promptly and independently reviewed, whether the advice or opinion of another therapist was obtained, whether the evaluation was made in light of the proper legal standards for commitment, and whether other evidence of good faith exists.” (Citation omitted). Id. at 1083.
The court derived its psychotherapist judgment rule from the business judgment rule used to review decisions of corporate directors:
“In the business judgment rule, courts defer to the decisions of disinterested directors absent bad faith or self-interest. Many of the considerations cited as justifications for the business judgment rule are applicable to the present case. For example, as with business decisions, the court is not particularly qualified to review commitment decisions involving mental health and dangerousness. In addition, these types of commitment procedures require quick action, and ‘after-the-fact litigation is a most imperfect device to evaluate’ those decisions, as in the corporate setting. * * * Finally, policy considerations favor giving psychotherapists, as well as corporate directors, significant discretion to use their best judgment, recognizing that ‘[a] rule which penalizes the choice of seemingly riskier alternatives *97* * * may not be in the interest’ of the parties or society.” (Citations omitted). Id.
Plaintiff contends there is no basis for deviating from the malpractice standard of care set forth in Bruni. In response to defendants’ arguments, plaintiff asserts (1) although there are limits to the degree of accuracy, many practitioners believe that prediction of violent behavior is possible13; (2) recent scholarship suggests that predictive accuracy may be better than previously thought14; (3) there are objective professional standards for evaluating violent potential15; and (4) holding psychiatrists only to a professional judgment rule in effect requires a psychiatric patient to assume the risk of improper treatment.
We find the arguments in support of adopting the professional judgment rule persuasive. Though a psychiatrist’s ability to predict violent behavior is probably better than a layperson’s, and there does appear to be some consensus within the mental health community on the factors relevant to a diagnosis of violent propensities,16 diagnosing both the existence of violent propensities and their severity is still a highly subjective undertaking. Psychiatric evaluations of any given fact pattern are bound to vary widely.17 And once a determination is made that a patient possesses a propensity for violent behavior, deciding upon a course of treatment poses difficult questions. The patient’s right to good medical care, including freedom from unnecessary confinement18 and unwarranted breaches of confidentiali*98ty,19 must be balanced against the need to protect potential victims.20 Courts, with the benefit of hindsight, should not be allowed to second-guess a psychiatrist’s professional judgment.
On the other hand,- a psychiatric *99patient is not required to assume the risk of improper treatment. Where there are professional standards of care a psychiatrist is required to conform to the standards at all times or suffer liability. Where there are no professional standards, a psychiatrist must exercise good faith judgment based on a thorough evaluation of all relevant factors. Professional standards will be used to determine which factors are relevant and whether an evaluation was thorough.
Therefore, we hold that a psychiatrist will not' be held liable for the violent acts of a voluntarily hospitalized mental patient subsequent to the patient’s discharge if (1) the patient did not manifest violent propensities while being hospitalized and there was no reason to suspect the patient would become violent after discharge, or (2) a thorough evaluation of the patient’s propensity for violence was conducted, taking into account all relevant factors, and a good faith decision was made by the psychiatrist that the patient had no violent propensity, or (3) the patient was diagnosed as having violent' propensities and, after a thorough evaluation of the severity of the propensities and a balancing of the patient’s interests and the interests of potential victims, a treatment plan was formulated in good faith which included discharge of the patient.
Ill
As reflected in his discharge summaries, Dr. Murray knew that Theresa had a potential to harm Carly. He was told of Theresa’s fears of harming Carly as well as her statement of June 1 that she planned to kill Carly. He, Dr. Wales and the other members of the team knew at the time of Theresa’s release from the hospital that she had a potential for harming Carly. However, Dr. Murray concluded, as did the others, that Theresa did not have a “plan” to harm Carly or possess a severe and imminent potential for violence. He felt the plan to place Carly with Greg’s relatives was adequate to protect Carly. Absent bad faith or lack of a thorough evaluation of Theresa’s condition, Dr. Murray will not be held liable for this judgment, even though it led to the death of Carly.
Dr. Litvak testified that Dr. Murray did not conduct a thorough evaluation of Theresa’s homicidal potential. Basing his opinion on his review of the hospital records and the depositions of Dr. Murray and Dr. Wales, Dr. Litvak concluded that Dr. Murray should have talked with Theresa and asked her what her thoughts were, trying to ascertain whether those thoughts were rational or not, what her reasons were for wanting to hurt the baby, what means she had thought of for hurting the baby, if she had a specific date, time and place that she might be doing it, and whether or not she had actually attempted to do anything like that in the past; that Dr. Murray should have ascertained in general how impulsive a person Theresa might be, and then whether her illness might make her more impulsive, unpredictable and unreliable; that Dr. Murray should have determined whether Theresa had previously acted aggressively toward anyone else and under what cir*100cumstances; and that Dr. Murray should have determined whether family members had seen Theresa acting in such a way as to indicate she would hurt the baby or heard her talking about hurting the baby. Dr. Litvak found no indication that Dr. Murray went into that amount of detail in the records Dr. Litvak reviewed.
It is clear from Dr. Murray’s own testimony that he did not talk with Theresa about her statement of June 1. However, it is not- clear from the record whether another member of the treatment team talked .with her about it. Perry believed she d(id, but could not remember. More important, it is not clear whether Dr. Murray or another team member should have talked with Theresa about the specific threat in light of her recanting three hours later. Dr. Greenfield, expert witness for Dr. Murray and the hospital, stated he would have wanted to personally talk with Theresa about the note but that failure to do so was not malpractice.-
It is not clear from the record whether Dr. Murray evaluated Theresa’s potential for violence. He was not in daily contact with her while she was at the hospital, nor did he conduct individual therapy sessions with her. However, he did talk with her on occasion and he received information on Theresa from members of the team who were in daily contact with her and conducting therapy. Jn addition, he received information from Dr. Wales and members of Theresa’s family.
The plaintiff faults Dr. Murray for not being more involved in personally treating Theresa. However, “team treatment” appears to be a legitimate mode of therapy within the psychiatric community. See Gowan v. United States (D. Ore. 1985),. 601 F. Supp. 1297, 1301-1302. We find that a genuine issue exists as to whether Dr. Murray acted as would a reasonable psychiatrist in evaluating Theresa’s potential for violence.
The jury was instructed to consider whether Theresa’s release from the hospital was premature, whether the treatment plan for the release was inadequate, and whether there was a failure to warn necessary persons of the known or reasonably foreseeable dangerous consequences of the release. Unfortunately, the basis of the jury’s verdict against Dr. Murray was never tested. The jury could have found Dr. Murray liable for negligence in evaluating Theresa’s condition or for one of the other grounds asserted by the plaintiff. However, under the professional judgment rule, the other grounds, including premature discharge and failure to warn, can be a basis for liability only in the absence of good faith or a failure to exercise professional judgment. Therefore, in light of this new standard, a new trial should be conducted on Dr. Murray’s liability.
IV
The final issue we address concerns the trial court's denial of the plaintiff’s motion to add Theresa as a new party plaintiff with a cause of ao tion sounding in medical malpractice. The plaintiff’s motion was made in March 1985, well after the statute of limitations had run on Theresa’s claim. To come within the statute of limitations, the amended complaint would have to relate back to the date of the original complaint.
Civ. R. 15(C) governs the relation back of amended pleadings:
“* * * An amendment changing the party against whom a claim is asserted relates back if the foregoing provision is satisfied and, within the period provided by law for commenc*101ing the action against him, the party to be brought in by amendment (1) has received such notice of the institution of the action that he will not be prejudiced in maintaining his defense on the merits, and (2) knew or should have known that, but for a mistake concerning the identity of the proper party, the action would have been brought against him.”
The Ohio Rules of Civil Procedure, including Civ. R. 15(C), were patterned after the Federal Rules of Civil Procedure. The 1966 Advisory Committee Notes to federal Rule 15(c) state: “The relation back of amendments changing plaintiffs is not expressly treated in * * * Rule 15(c) since the problem is generally easier. * * * [T]he chief consideration of policy is that of the statute of limitations, and the attitude taken in * * * Rule 15(c) toward change of defendants extends by analogy to amendments changing plaintiffs.”
The primary purpose of Civ. R. 15(C) is to preserve actions which, through mistaken identity or misnomer, have been filed against the wrong person. No mistake or misnomer is alleged here, and Theresa waited well over three years to bring an action.
Federal courts have allowed relation back when the new plaintiff is the real party in interest21 or an original plaintiff brings a new cause of action in a different capacity,22 but generally not when a new plaintiff brings a new cause of action. See Marlowe v. Fisher Body (C.A. 6, 1973), 489 F. 2d 1057, 1064; People of the Living God v. Star Towing Co. (E.D. La. 1968), 289 F. Supp. 635, 641. Therefore, we affirm the dismissal of the plaintiffs motion to add Theresa as a party plaintiff.
Based on our resolution of these issues, it is unnecessary to address other issues raised by the parties. We reverse the judgment of the court of appeals and remand this cause for a new trial on all issues, including, damages, based on the professional judgment standard.
Judgment reversed and cause remanded.
Moyer, C.J., Sweeney, Locher, Holmes and Wright, JJ., concur. Douglas, J., concurs in part and dissents in part.We are not deciding whether a psychiatrist’s duty to protect a person from the violent propensities of the psychiatrist’s patient extends to the outpatient setting. See, generally, Tarasoff v. Regents of the University of California (1976), 17 Cal. 3d 425, 131 Cal. Rptr. 14, 551 P. 2d 334.
In support of this point, defendants cite numerous commentators. See Stead-man, Predicting Violence Leading to Homicide (1986), 62 Bull. N.Y. Acad. Med. 570, 576 (“* * * our understanding of violence and homicide is so poor that we are able accurately to predict [its occurrence] only in rare cases”); Appelbaum, Tarasoff and the Clinician: Problems in Fulfilling the Duty to Protect (1985), 142 Am. J. Psychiatry 425, 426 (“Determining future dangerousness is a task at which no profession has yet been demonstrated to be successful and for which psychiatrists probably have no special skills.”); Wettstein, The Prediction of Violent Behavior and the Duty to Protect Third Parties (1984), 2 Behavioral Sci. & Law 291, 312 (concerning the clinical prediction of violent behavior, “predictive reliability and accuracy remain poor, and significant numbers of false positive predictions continue to be made, given the relatively low prevalence of serious violent conduct in psychiatric patients.”); Cocozza & Steadman, The Failure of Psychiatric Predictions of Dangerousness: Clear and Convincing Evidence (1976), 29 Rutgers L. Rev. 1084,1098-1099 *94(“The findings of this study taken together with the other works reviewed in this paper would appear to represent clear and convincing evidence of the inability of psychiatrists or anyone else to predict dangerousness accurately.”); Ennis & Litwack, Psychiatry and the Presumption of Expertise: Flipping Coins in the Courtroom (1974), 62 Cal. L. Rev. 693, 695-696 (“Based upon our reading of the professional literature, we have concluded that * * * there is no evidence warranting the assumption that psychiatrists can accurately determine who is ‘dangerous. * * *.’ ”); Diamond, The Psychiatric Prediction of Dangerousness (1974), 123 U. Pa. L. Rev. 439, 440 (psychiatrists cannot predict dangerousness with “reasonable accuracy”); and People v. Burnick (1975), 14 Cal. 3d 306, 326, 121 Cal. Rptr. 448, 501, 535 P. 2d 352, 365, quoting Murel v. Baltimore City Criminal Court (1972), 407 U.S. 355, 365, fn. 2 (Douglas, J., dissenting from dismissal of certiorari) (“Predictions of dangerous behavior, no matter who makes them, are incredibly inaccurate, and there is a growing consensus that psychiatrists are not uniquely qualified to predict dangerous behavior and are, in fact, less accurate in their predictions than other professions.”).
Mills, The So-Called Duty to Warn: The Psychotherapeutic Duty to Protect Third Parties From Patients’ Violent Acts (1984), 2 Behavioral Sci. & Law 237, 248.
Defendants cite two cases in support of this proposition. Johnson v. United States (M.D. Fla. 1976), 409 F. Supp. 1283, 1293, reversed on other grounds (C.A. 5, 1978), 576 F. 2d 606 (“[I]t is sound psychiatric practice to emphasize outpatient therapy rather than lengthy custody and supervision. Modern psychiatry has recognized the importance of making every reasonable effort to return a patient to an active and productive life. Thus, the patient is encouraged to develop his self-confidence by adjusting to the demands of everyday existence. In this view, mental hospitals are not seen as dumping grounds for all persons whose behavior society might find inconvenient or offensive; institutionalization is the exception, not the rule) and is called for only when a paramount therapeutic interest or the protection of society leaves no choice. * ** * [Mjodern psychiatric practice does not require a patient to be isolated from normal human activities until every possible danger has passed.”); and Schrempf v. New York (1985), 66 N.Y. 2d 289, 295, 496 N.Y. Supp. 2d 973, 977, 487 N.E. 2d 883, 887.
See Taig v. New York (1963), 19 App. Div. 2d 182, 183, 241 N.Y. Supp. 2d 495, 496-497 (“The prediction of the future course of a mental illness is a professional judgment of high responsibility and in some instances it involves a measure of calculated risk. If a liability were imposed on the physician or the State each time the prediction of future course of mental disease was wrong, few releases would ever be made and the hope of recovery and rehabilitation of a vast number of patients would be impeded and frustrated.”); see, also, Cameron v. New York (1971), 37 App. Div. 2d 46, 322 N.Y. Supp. 2d 562, affirmed (1972), 30 N.Y. 2d 596, 331 N.Y. Supp. 2d 30, 282 N.E. 2d 118.
R.C. 5122.34 states in part:
“Persons acting in good faith, either upon actual knowledge or information thought by them to be reliable, who procedurally or physically assist in the hospitalization or discharge or in judicial proceedings of a person under this chapter, do not come within any criminal provisions, and are free from any liability to the person hospitalized or to any other person.” (Emphasis added.)
Significantly, former R.C. 5122.34 (136 Ohio Laws, Part II, 2062) exculpated “persons acting reasonably and -in good faith” as contrasted with “persons acting in good faith, ” the language in the current statute. (Emphasis added.)
See Oelsner v. New York (1985), 66 N.Y. 2d 636, 495 N.Y. Supp. 2d 359, 485 N.E. 2d 1024; St. George v. New York (1954), 283 App. Div. 245, 248, 127 N.Y. Supp. 2d 147, 150, affirmed (1954), 308 N.Y. 681, 124 N.E. 2d 320.
See Schrempf v. New York, supra, at 296, 496 N.Y. Supp. 2d at 978, 487 N.E. 2d at 888; Topel v. Long Island Jewish Medical Center (1981), 55 N.Y. 2d 682, 446 N.Y. Supp. 2d 932, 431 N.E. 2d 293 (no liability for decision to put suicidal patient on fifteen-minute observation instead of continuous, even though patient subsequently committed suicide, because expert testimony did not negate the factors which led to the decision.); McDonnell v. Cty. of Nassau (May 15, 1986), N.Y. Sup. Ct., Nassau Cty. No. 11152/73, unreported.
See St. George, supra, at 248, 127 N.Y. Supp. at 150-151 (“Future human behavior is unpredictable, and it would place an unreasonable and unfair burden upon the State if it were to be held responsible in damages for everything that a person does after he has been discharged or released from one of its State institutions, even though the release was through an error of judgment, unless there is something more present than is contained in this record.”); Taig, supra; Cameron, supra; Centeno v. New York (1976), 40 N.Y. 2d 932, 389 N.Y. Supp. 2d 837, 358 N.E. 2d 520; Fiederlein v. New York Health & Hospitals Corp. (1981), 80 App. Div. 2d 821, 437 N.Y. Supp. 2d 321, affirmed (1982), 56 N.Y. 2d 573, 450 N.Y. Supp. 2d 181, 435 N.E. 2d 398.
See O’Shea v. United States (E.D. N.Y. 1985), 623 F. Supp. 380 (no liability even though patient subsequently killed *96someone, because psychiatrist made a careful examination of patient’s condition at the time of the requested admission).
See Givelber, Bowers & Blitch, Tarasoff, Myth and Reality: An Empirical Study of Private Law in Action (1984), Wis. L. Rev. 443, 462-464.
See Monahan, The Prediction of Violent Behavior: Toward a Second Generation of Theory and Policy (1984), 141 Am. J. Psychiatry 10, 11 (“No one thinks that the prediction of violence is on the verge of attaining a validity comparable to that of the prediction of the weather. But fewer people are now convinced of the inevitability of ACLU-type evaluations of the field, that psychiatrists and psychologists are inaccurate ‘about 95 percent of the time.’ There may indeed be a ceiling on the level of accuracy that can ever be expected of the clinical prediction of violent behavior. That ceiling, however, may be closer to 50% than to 5% among some groups of clinical interests.”).
See Appelbaum, supra (142 Am. J. Psychiatry 425), at 426.
See Kroll & Mackenzie, When Psychiatrists Are Liable: Risk Management and Violent Patients (1983), 34 Hospital & Community Psychiatry 29.
In response to the question whether two doctors will come to the same diagnosis, Dr. Ronald Fox, expert witness for Dr. Wales, responded: “The evidence is that two people making the same diagnosis as a result of about, oh, I would say 200 or 300 studies that have been done show very clearly that the agreement on diagnosis runs from about the 50 to 60 percent range. In fact, if you take two experienced people, they could be psychiatrists or psychologists, it makes no difference, tell them a person is depressed, and you only want them to make a judgment as to whether it is mild, moderate, or severe, the agreement in that case, how severe is the depression, shows agreement at less than 60 percent. It runs 55 percent, which means you have a very large error in those kinds of judgments.”
One commentator suggests that concern about liability for the violent acts of their patients has led some psychiatrists to involuntarily hospitalize patients who may or may not be mentally ill solely for the purpose of preventing them from engaging in violent behavior. Appelbaum, The New Preventive Detention: Psychiatry’s Problematic Responsibility for the Control of *98Violence (1988), 145 Am. J. Psychiatry 779; see, also, Givelber, supra, at 478.
Under Ohio law, involuntary commitment can be undertaken for the benefit of the patient, society, or both. R.C. 5122.01(B). A mentally ill person can be involuntarily hospitalized for the protection of society if that person represents “a substantial risk of physical harm to others,” R.C. 5122.01(B)(2), and psychiatrists can cause a person they believe represents such a risk to be temporarily taken into custody' pursuant to emergency hospitalization procedures, R.C. 5122.10.
Since, in Ohio, a psychiatrist’s decision to involuntarily commit someone is judged by a good faith standard, R.C. 5122.34, to apply a lesser standard to other viable methods of discharging the duty to protect a potential victim from a patient’s violence would, in effect, encourage commitment over those other methods.
In Ohio, a physician can be held liable for unauthorized disclosures of medical information. See Hammonds v. Aetna Cas. & Sur. Co. (N.D. Ohio 1965), 243 F. Supp. 793; Nationwide Mut. Ins. Co. v. Jackson (1967), 10 Ohio App. 2d 137, 39 O.O. 2d 242, 226 N.E. 2d 760; Prince v. St. Francis-St. George Hospitals, Inc. (1985), 20 Ohio App. 3d 4, 20 OBR 4, 484 N.E. 2d 265; Levias v. United Airlines (1985), 27 Ohio App. 3d 222, 27 OBR 262, 500 N.E. 2d 370; see, also, R.C. 4731.22(B)(4) (“Willfully betraying a professional confidence” is a ground for physician discipline.); see, generally, Johnston, Breach of Medical Confidence in Ohio (1986), 19 Akron L. Rev. 373. However, an exception exists for disclosures necessary to protect individual or public welfare. See Jones v. Stanko (1928), 118 Ohio St. 147, 160 N.E. 456, paragraph two of the syllabus; see, generally, Johnston, supra.
The American Medical Association has long allowed breaches of confidence when “it becomes necessary in order to protect the welfare of the individual or of the community.” Principles of Medical Ethics of the American Medical Association (1957), Section 9. Apparently, even before the seminal Tarasoff decision, therapists occasionally gave warnings to potential victims and 'their families. Wise, Where the Public Peril Begins: A Survey of Psychotherapists to Determine the Effects of Tarasoff (1978), 31 Stan. L. Rev. 165, 183-184, 190.
Warning a potential victim of a patient’s propensity for violence has been suggested as one way to discharge the duty to protect others from a patient’s violence without infringing on a patient’s liberty interest. See Fleming & Maximov, The Patient or His Victim: The Therapist’s Dilemma (1974), 62 Cal. L. Rev. 1025. However, warning a potential victim has also been strongly criticized as generally counterproductive to effective patient treatment. See Stone, The Tarasoff Decisions: Suing Psychotherapists to Safeguard Society (1976), 90 Harv. L. Rev. 358 (a duty to warn potential victims is incompatible with an effective therapeutic relationship and will deter both patients and therapists from undertaking treatment); Mills, supra, at 250 and 257 (“Confidentiality must give way where the psychotherapist believes that the patient is dangerous, but if an alternate, clinical means can be found for dealing with the patient’s dangerousness, such means are highly preferable.”); Givelber, supra, at 469-472 (survey indicates many therapists feel warning is in conflict with clinical judgment); Weinstock, Confidentiality and the New Duty to Protect: The Therapist’s Dilemma (1988), 39 Hospital & Community Psychiatry 607, 609 (“Reporting and warning can lead to outcomes harmful to both patient and potential victim.”); Schmid, Appelbaum, Roth & Lidz, Confidentiality in Psychiatry: A Study of the Patient’s View (1983), 34 Hospital & Community Psychiatry 353, 355 (concluding that “* * * factors that diminish confidentiality in psychiatric care *99may impair the willingness of patients to seek and to remain in treatment”).
But, see, Wulsin, Bursztajn & Gutheil, Unexpected Clinical Features of the Tarasoff Decision: The Therapeutic Alliance and the “Duty to Warn” (1983), 140 Am. J. Psychiatry 601 (noting some therapeutic advantages of approaching the duty to warn in clinical terms).
Metropolitan Paving Co. v. Internatl. Union of Operating Engineers (C.A. 10, 1971), 439 F. 2d 300, certiorari denied (1971), 404 U.S. 829.
Williams v. United States (C.A. 5, 1968), 405 F. 2d 234.