In Littleton v. Good Samaritan Hosp. & Health Ctr. (1988), 39 Ohio St.3d 86, 529 N.E.2d 449, this court determined that under certain circumstances a psychiatrist can be held liable for the violent acts of a voluntarily hospitalized patient following the patient’s release from the hospital. The question left open in Littleton, however, was “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.” Littleton, 39 Ohio St.3d at 92, 529 N.E.2d at 455, fn. 3. This is the issue we must decide today.
*293I
“SPECIAL RELATION” AND THE DUTY TO CONTROL
It is by now an axiom that duty is an essential element of a cause of action for negligence. See Menifee v. Ohio Welding Products, Inc. (1984), 15 Ohio St.3d 75, 77, 15 OBR 179, 180, 472 N.E.2d 707, 710. See, also, Prosser & Keeton on Torts (5 Ed.1984) 164, Section 30. In Ohio, “[t]he existence of a duty depends on the foreseeability of the injury. * * * The test for foreseeability is whether a reasonably prudent person would have anticipated that an injury was likely to result from the performance or nonperformance of an act.” (Citations omitted.) Menifee, supra, 15 Ohio St.3d at 77,15 OBR at 180, 472 N.E.2d at 710. See, also, Hill v. Sonitrol of Southwestern Ohio, Inc. (1988), 36 Ohio St.3d 36, 39, 521 N.E.2d 780, 783. In addition, it is generally recognized that where the defendant “in fact has knowledge, skill, or even intelligence superior to that of the ordinary person, the law will demand of that person conduct consistent with it.” Prosser & Keeton on Torts, supra, at 185, Section 32.
However, foreseeability alone is not always sufficient to establish the existence of a duty. This court has followed the commoñ-law rule, as set forth at 2 Restatement of the Law 2d, Torts (1965) 116-130, Sections 314 to 319, that there is no duty to act affirmatively for another’s aid or protection absent some “special relation” which justifies the imposition of a duty. Littleton, supra, 39 Ohio St.3d at 92, 529 N.E.2d at 455; Hill, supra, 36 Ohio St.3d at 39, 521 N.E.2d at 784; Gelbman v. Second Natl. Bank of Warren (1984), 9 Ohio St.3d 77, 79, 9 OBR 280, 281-282, 458 N.E.2d 1262, 1263.
Restatement Section 314 states the general rule that there is no duty to act affirmatively for another’s aid or protection.2 Section 315 “is a special application *294of the general rule stated in § 314.” 2 Restatement of Torts, supra, at 122, Section 315, Comment a. It provides that there is no duty to control the conduct of a third person to prevent him from causing physical harm to another unless a “special relation” exists between the defendant and the third person or between the defendant and the other. Sections 316 to 319 set forth the relations between the defendant and the third person which require the defendant to control the third person’s conduct. In Littleton, supra, 39 Ohio St.3d at 92-93, 529 N.E.2d at 455, we relied on Section 319 of the Restatement in finding that a special relation exists between a psychiatrist and his patient in the hospital setting. Section 319 states that:
“One who takes charge of a third person whom he knows or should know to be likely to cause bodily harm to others if not controlled is under a duty to exercise reasonable care to control the third person to prevent him from doing such harm.”
The issue, therefore, becomes whether the relationship between a psychotherapist and the outpatient constitutes a “special relation” which imposes a duty upon the psychotherapist to protect others against and/or control the patient’s violent conduct.
In Tarasoff, supra, 17 Cal.3d 425, 131 Cal.Rptr. 14, 551 P.2d 334, the Supreme Court of California found that the psychotherapist-outpatient relationship constitutes such a special relation. In so finding, the court did not engage in a traditional Restatement analysis. Under a traditional Restatement analysis, Section 319 would take center stage. In Tarasoff, the court treated Section 315 et seq. as reflective of an overall principle that affirmative duties to control should be imposed whenever the nature of the relationship warrants social recognition as a special relation. Id., 17 Cal.3d at 435, 131 Cal.Rptr. at 23, 551 P.2d at 343. In this way, the court subjected Section 315 to an expansive reading. Thus, the court noted that “courts have increased the number of instances in which affirmative duties are imposed not by direct rejection of the common law rule [of *295nonliability for nonfeasance], but by expanding the list of special relationships which 'will justify departure from that rule.”3 Id. at fn. 5.
The court then engaged in a two-part analysis. First, the court drew an analogy to cases which have imposed a duty upon physicians to diagnose and warn about their patient’s contagious disease, and concluded that “ ‘by entering into a doctor-patient relationship the therapist becomes sufficiently involved to assume some responsibility for the safety, not only of the patient himself, but also of any third person whom the doctor knows to be threatened by the patient.’ ” Id., 17 Cal.3d at 437, 131 Cal.Rptr. at 24, 551 P.2d at 344, quoting Fleming & Maximov, The Patient or His Victim: The Therapist’s Dilemma (1974), 62 Cal.L.Rev. 1025, 1030.
Second, the court weighed various public policy concerns, concluding that the public interest in safety from violent assaults outweighs the countervailing interests of safeguarding the confidential character of psychotherapeutic communications and the difficulty inherent in forecasting dangerousness. Id., 17 Cal.3d at 437-443,131 Cal.Rptr. at 24-28, 551 P.2d at 344-348.
The court held, therefore, that:
“When a therapist determines, or pursuant to the standards of his profession should determine, that his patient presents a serious danger of violence to another, he incurs an obligation to use reasonable care to protect the intended victim against such danger. The discharge of this duty may require the therapist to take one or more of various steps, depending upon the nature of the case. Thus it may call for him to warn the intended victim or others likely to apprise the victim of the danger, to notify the police, or to take whatever other steps are reasonably necessary under the circumstances.” Id., 17 Cal.3d at 431, 131 Cal.Rptr. at 20, 551 P.2d at 340.
Since Tarasoff, a majority of courts that have considered the issue have concluded that the relationship between the psychotherapist and the outpatient constitutes a special relation which imposes upon the psychotherapist an affirmative duty to protect against or control the patient’s violent propensities. Recognizing that the duty is imposed by virtue of the relationship, these courts acknowledge that the duty can be imposed not only upon psychiatrists, but also on psychologists, social workers, mental health clinics and other mental health professionals who know, or should have known, of their patient’s violent propensi*296ties. The courts do not impose any single formulation as to what steps must be taken to alleviate the danger. Depending upon the facts and the allegations of the case, the particular psychotherapist-defendant may or may not be required to perform any number of acts, including prescribing medication, fashioning a program for treatment, using whatever ability he or she has to control access to weapons or to persuade the patient to voluntarily enter a hospital, issuing warnings or notifying the authorities and, if appropriate, initiating involuntary commitment proceedings.
Most of the courts engage in a Tarasoff-type analysis by which Section 315 is subjected to an expansive reading. Others find a duty to exist under the rule stated in Section 319. Collectively, they recognize that there are various levels of being in “control” pursuant to Section 315, or being in “charge” pursuant to Section 319, with corresponding degrees of responsibility for the patient’s violent actions. Thus, although the psychotherapist may have less ability to control the patient in the outpatient setting than in the hospital setting, this lesser degree of control is not held to justify a blanket negation of the duty to control.
Generally, the courts focus their attention on balancing the countervailing public interests that were weighed in Tarasoff, including the additional concern that patients be placed in the least restrictive environment and that nonviolent patients not be subjected to hospitalization against their will in an effort to avoid liability. These courts conclude that the interests of society to be protected against the violent acts of mental patients outweigh the concerns of confidentiality, overcommitment, and difficulty of predicting violent acts. Hamman v. Maricopa Cty. (1989), 161 Ariz. 58, 775 P.2d 1122; Schuster v. Altenberg (1988), 144 Wis.2d 223, 424 N.W.2d 159; Evans v. Morehead Clinic (Ky.App.1988), 749 S.W.2d 696; Bardoni v. Kim (1986), 151 Mich.App. 169, 390 N.W.2d 218; Peck v. Counseling Serv. of Addison Cty., Inc. (1985), 146 Vt. 61, 499 A.2d 422; Brady v. Hopper (C.A.10, 1984), 751 F.2d 329; Lundgren v. Fultz (Minn.1984), 354 N.W.2d 25; Lipari v. Sears, Roebuck & Co. (D.C.Neb.1980), 497 F.Supp. 185; Jablonski v. United States (C.A.9, 1983), 712 F.2d 391; McIntosh v. Milano (1979), 168 N.J.Super. 466, 403 A.2d 500. See, also, Annotation, Liability of One Treating Mentally Afflicted Patient for Failure to Warn or Protect Third Persons Threatened by Patient (1978), 83 A.L.R.3d 1201; 2 American Jurisprudence, Proof of Facts 3d (1988) 327; Sear, The Psychotherapist’s Duty to Warn (1990), 40 Fedn. of Ins. & Corporate Counsel Qtrly. 406, 416-418.
In addition, a number of courts have relied on Tarasoff in finding that a psychotherapist can be held hable for the violent acts of a patient following the patient’s release from the hospital. Wofford v. E. State Hosp. (Okla.1990), 795 P.2d 516; Naidu v. Laird (Del.1988), 539 A.2d 1064; Petersen v. State (1983), 100 Wash.2d 421, 671 P.2d 230; Chrite v. United States (E.D.Mich.1983), 564 F.Supp. 341; Bradley Ctr., Inc. v. Wessner (1982), 250 Ga. 199, 296 S.E.2d 693. See, also, *297Annotation, Liability of One Releasing Institutionalized Mental Patient for Harm He Causes (1971), 38 A.L.R.3d 699. See, also, Div. of Corr., Dept. of Health & Social Serv. n Neakok (Alaska 1986), 721 P.2d 1121 (parolee case relying on Tarasoff).
The parties do not dispute that the psychotherapist-outpatient relationship justifies the imposition of a common-law duty upon the psychotherapist to control the violent propensities of the patient. In fact, Dr. Brown readily admits that “[i]n Tarasqff, supra, the Supreme Court of California set then-novel but reasonable parameters on a psychotherapist’s liability for violent acts of outpatients.” However, our research discloses that Tarasoff does not enjoy universal acceptance. Some courts have concluded that the typical psychotherapist-outpatient relationship lacks sufficient elements of control necessary to satisfy Sections 315 and/or 319. These courts reason that the duty to control is corollary to the right, power, or ability to control, and criticize Tarasoff for not specifically addressing the issue of a psychotherapist’s control over the outpatient. In addition, some of these courts find that public policy militates against the imposition of a duty in the outpatient setting. Boynton v. Burglass (Fla.App.1991), 590 So.2d 446, 448-449; Santa Cruz v. Northwest Dade Community Health Ctr., Inc. (Fla.App.1991), 590 So.2d 444, 445; Wagshall n Wagshall (1989), 148 A.D.2d 445, 447, 538 N.Y.S.2d 597, 598-599; King v. Smith (Ala.1989), 539 So.2d 262; Currie v. United States (C.A.4, 1987), 836 F.2d 209, 213; Cooke v. Berlin (1987), 153 Ariz. 220, 224-225, 735 P.2d 830, 834-835; Hasenei v. United States (D.Md.1982), 541 F.Supp. 999,1009. See, also, Stone, The Tarasoff Decisions: Suing Psychotherapists to Safeguard Society (1976), 90 Harv.L.Rev. 358, 365-366; Note, Affirmative Duty After Tarasoff (1983), 11 Hofstra L.Rev. 1013, 1029-1030.
In light of the opposing views on the subject, we deem it necessary to resolve the issue of duty in the outpatient setting by balancing the various factors which are the focus of judicial attention on both sides of the issue. These factors consist of the following: (1) the psychotherapist’s ability to control the outpatient; (2) the public’s interest in safety from violent assault; (3) the difficulty inherent in attempting to forecast whether a patient represents a substantial risk of physical harm to others; (4) the goal of placing the mental patient in the least restrictive environment and safeguarding the patient’s right to be free from unnecessary confinement; and (5) the social importance of maintaining the confidential nature of psychotherapeutic communications. See Littleton, supra, 39 Ohio St.3d at 93-94, 97-98, 529 N.E.2d at 456-57, 459. See, also, Perreira v. State (Colo.1989), 768 P.2d 1198, 1214-1215.
A
In weighing these factors, we must bear in mind that duty is not an immutable concept, nor is it grounded in natural law. As Prosser & Keeton explains, “[t]he *298statement that there is or is not a duty begs the essential question — -whether the plaintiffs interests are entitled to legal protection against the defendant’s conduct. * * * ‘[D]uty’ is not sacrosanct in itself, but is only an expression of the sum total of those considerations of policy which lead the law to say that the plaintiff is entitled to protection.” Prosser & Keeton on Torts, supra, at 357-358, Section 53. Thus, “ ‘duty’ is only a word with which we state our conclusion, and no more.” Id. at 281, Section 43.
Accordingly, there is no more magic inherent in the conclusory term “special relation” than there is in the term “duty.” Both are part and parcel of the same inquiry into whether and how the law should regulate the activities and dealings that people have with each other. As society changes, as our sciences develop and our activities become more interdependent, so our relations to one another change, and the law must adjust accordingly. “Duty” is not a rigid formalistic concept forever embedded in the standards of a simplistic yesteryear. Relations perhaps regarded as tenuous in a bygone era may now be of such importance in our modern complicated society as to require certain assurances that risks associated therewith be contained. These principles do not shed their inherent flexibility when applied in the context of a defendant’s duty to control the violent conduct of a third person. See Lipari, supra, 497 F.Supp. at 192-193; McIntosh, 168 N.J.Super. at 495, 403 A.2d at 515; Tarasoff, 17 Cal.3d at 442, 131 Cal.Rptr. at 27-28, 551 P.2d at 347-348; Prosser & Keeton on Torts, supra, at 356-358, Section 53; Harper & Kime, The Duty to Control the Conduct of Another (1934), 43 Yale L.J. 886, 904-905.
B
There is indeed a current running through the relevant Restatement sections that in order for a special relation to exist between the defendant and the third person, the defendant must have the ability to control the third person’s conduct. Moreover, the cases from which these sections derive indicate that the ability to control is not the fictitious control which provides the basis for vicarious liability. Instead, “control” is “used in a very real sense.” Harper & Kime, The Duty to Control the Conduct of Another, supra, 43 Yale L.J. at 891. Further, it would be tantamount to imposing strict liability to require the defendant to control a third person’s conduct where he lacks the ability to do so.
However, those courts which find the ability to control to be lacking in the outpatient setting tend to take a rather myopic view of the level or degree of control needed to impose the duty. They appear to assume that in order to satisfy Section 315 in general, or Section 319 in particular, there must be actual constraint or confinement, whereby the third person’s physical liberty is taken *299away or restricted.4 In viewing the issue in this way, these courts fail to recognize that the duty to control the conduct of a third person is commensurate with such ability to control as the defendant actually has at the time. See, e.g., 2 Restatement of Torts, supra, at 116, Section 314, Comment a; at 124, Section 316, Comments a and b; at 126, Section 317, Comment c; and at 127, Section 318, Comment a. See, also, Lundgren, 354 N.W.2d at 27-28; McIntosh, supra, 168 N.J.Super. at 483, 403 A.2d at 508-509, fn. 11. In other words, it is within the contemplation of the Restatement that there will be diverse levels of control which give rise to corresponding degrees of responsibility.
The argument could be made that the two Illustrations to Section 319 describe situations where the potentially harmful individual is confined and then negligently released. However, the rule stated in Section 319 is noticeably broader. Just as there are diverse levels of control, there are different levels of taking “charge.” In contrast to the “takes charge” language of Section 319, Section 320 sets forth the duty of one who “takes the custody” of another under certain conditions to control the conduct of third persons as to prevent them from harming the other. The comments to Section 320 indicate that the term “custody,” as used in Section 320, is more suggestive of restrictions on liberty. Thus, “the illustrations appended to [Section 319], which are drawn in the context of a private hospital or sanitarium for the insane, are obviously not by way of limitation.” McIntosh, 168 N.J.Super. at 483, 403 A.2d at 509, fn. 11.
Although the outpatient setting affords the psychotherapist a lesser degree of control over the patient than does the hospital setting, it nevertheless embodies sufficient elements of control to warrant a corresponding duty to control. As we noted supra, there are a number of anticipatory measures that could be taken in the outpatient setting to prevent the patient’s violent propensities from coming to fruition. To find that such measures lack the quality of control would require us to feign ignorance of the facts and testimony in the case sub judice.
*300Dr. Brown testified that neuroleptic medication controls symptoms of schizophrenia in approximately seventy percent of schizophrenics. Lambert acknowledged that the symptoms of schizophrenia can be controlled with antipsychotic medication such as Navane, and that one of the possible risks associated with taking a medication-controlled schizophrenic off his medication is that he can become dangerous to himself or others. Until Dr. Brown weaned Matt off his medication, Matt was a medication-controlled and treatment-compliant patient.Drs. Goff and Tanay were both of the opinion that Matt would have remained compliant with his treatment and medication had Dr. Brown not weaned him off the Navane, and that had Matt remained on medication, he would not have had the overt psychotic symptoms that led him to kill his parents and injure his sister.
In addition, Drs. Goff and Tanay testified that at various points in time after October 11, 1990, the last time Dr. Brown saw Matt, a number of other steps could have been taken by Dr. Brown, FFCC and its employees to prevent Matt’s dangerous proclivities from manifesting. According to their testimony, Dr. Brown should have closely monitored Matt’s condition for at least six months after withdrawing his medication. If this had been done, Dr. Brown could have reinstated Matt’s medication upon the reappearance of symptoms. Since this was not accomplished, Matt’s condition eventually deteriorated to the point of noncompliance while under the care of FFCC. At this point, Drs. Goff and Tanay opined, FFCC should have taken aggressive action to persuade Matt to continue treatment and have his medication reinstated by Dr. Brown. Such action should have included, among other things, strong family involvement, making Matt’s participation in vocational therapy contingent upon continued treatment, and telling Matt that he faced involuntary hospitalization unless he resumed taking his medication. If such measures had proved to be ineffective, and Matt nevertheless continued to deteriorate as he did, he would need to be involuntarily hospitalized, which, Drs. Goff and Tanay opined, should have taken place in May or June 1991.
Thus, we conclude that the psychotherapist-outpatient relationship embodies sufficient elements of control to warrant a corresponding duty to control.5
*301c
Society has a strong interest in protecting itself from those mentally ill patients who pose a substantial risk of harm. See R.C. 5122.01(B). See, also, Tarasoff, 17 Cal.3d at 440, 442, 131 Cal.Rptr. at 26, 27, 551 P.2d at 346, 347-348. To this end, society looks to the mental health profession to play a significant role in identifying and containing such risks. See, generally, R.C. Chapter 5122. The mental health community, therefore, has a broadly based responsibility to protect the community against danger associated with mental illness. See, e.g., Lipari, 497 F.Supp. at 190; McIntosh, 168 N.J.Super. at 489, 403 A.2d at 512. This responsibility is analogous to the obligation a physician has to warn others of his patient’s infectious or contagious disease.
In Jones v. Stanko (1928), 118 Ohio St. 147, 160 N.E. 456, at paragraphs one and two of the syllabus, the court held that a physician not only has the statutory duty to report his patient’s contagious disease to the appropriate governmental agency, but also the duty “to exercise ordinary care in giving notice of the existence of such contagious disease to other persons who are known by the physician to be in dangerous proximity to such patient.” Jones is important for four reasons. First, it demonstrates that Ohio common law recognizes that a physician can have a duty to others with whom he has no professional relationship. Second, it accepts that a duty can arise by virtue of the public interest in containing certain risks. Third, it places a duty upon the physician to act affirmatively to protect others from a danger not only of which he is aware, but also of which he should be aware. Fourth, the duty owed by the physician to diagnose and treat his patient’s condition for the benefit of others is the same duty already owing to the patient.
As explained in McIntosh, 168 N.J.Super. at 490, 403 A.2d at 512, a patient’s dangerous propensities “may affect [others] in much the same sense as a disease may be communicable. The obligation imposed by this court, therefore, is similar to that already borne by the medical profession in another context.”
D
In Littleton, supra, 39 Ohio St.3d at 97, 529 N.E.2d at 459, we recognized the difficulty inherent in forecasting whether a particular patient may pose a danger to others. Such difficulty, however, does not justify a blanket denial of recovery. The concept of due care adequately accounts for the difficulty of rendering a definitive diagnosis of a patient’s propensity for violence. The psychotherapist is not expected to render a perfect prediction of future violence. All that is *302required is that he or she arrive at an informed assessment of the patient’s propensity for violence. Littleton, 39 Ohio St.3d at 98-99, 529 N.E.2d at 460; Perreira, 768 P.2d at 1216-1217; Schuster, 144 Wis.2d at 246, 424 N.W.2d at 168-169; Lipari, 497 F.Supp. at 192; McIntosh, 168 N.J.Super. at 483, 403 A.2d at 508; Tarasoff, 17 Cal.3d at 438, 131 Cal.Rptr. at 25, 551 P.2d at 345. Moreover, to hold that assessments of dangerousness are so plagued by uncertainty as to be without value would raise serious questions as to the entire present basis for civil commitment.6 See R.C. 5122.01(B)(2). See, also, Perreira, 768 P.2d at 1217; Schuster, 144 Wis.2d at 247-248, 424 N.W.2d at 169; McIntosh, 168 NJ.Super. at 494-495, 403 A.2d at 514.
E
One goal of modern psychiatry is to place patients in the least restrictive environment. See, e.g., Perreira, 768 P.2d at 1219. The patient has a right to be free of unnecessary confinement. Littleton, 39 Ohio St.3d at 97, 529 N.E.2d at 459. Mental hospitals are not dumping grounds for all persons whose behavior might prove to be inconvenient or offensive to society. In O’Connor v. Donaldson (1975), 422 U.S. 563, 575, 95 S.Ct. 2486, 2494, 45 L.Ed.2d 396, 407, the United States Supreme Court put it this way:
“May the State fence in the harmless mentally ill solely to save its citizens from exposure to those whose ways are different? One might as well ask if the State, to avoid public unease, could incarcerate all who are physically unattractive or socially eccentric. Mere public intolerance or animosity cannot constitutionally justify the deprivation of a person’s physical liberty.”
These are important interests, not to be lightly infringed upon. Accordingly, there is some trepidation concerning the imposition of a duty because of the fear that therapists will attempt to protect themselves from liability by involuntarily *303hospitalizing nonviolent mental patients. See Littleton, 39 Ohio St.3d at 97-98, 529 N.E.2d at 459, fn. 18.
This fear, however, has no reliable statistical support. See Perreira, 768 P.2d at 1219; McIntosh, 168 N.J.Super. at 496, 403 A.2d at 515. Instead, the statistical evidence that is available indicates that “Tamsoff has not discouraged therapists from treating dangerous patients, nor has it led to an increased use of involuntary commitment of patients perceived as dangerous.” Givelber, Bowers & Blitch, Tamsoff Myth and Reality: An Empirical Study of Private Law in Action, 1984 Wis.L.Rev. 443, 486.
Moreover, as the court explained in Lipari, 497 F.Supp. at 192-193:
“This argument misinterprets the nature of the duty imposed upon the therapist. The recognition of this duty does not make the psychotherapist liable for any harm caused by his patient, but rather makes him liable only when his negligent treatment of the patient caused the injury in question. * * *
“Thus, * * * a psychotherapist is not subject to liability for placing his patient in a less restrictive environment, so long as he uses due care in assessing the risks of such a placement. This duty is no greater than the duty already owing to the patient.” See, also, Perreira, supra, 768 P.2d at 1219.
F
As to the importance of safeguarding the confidentiality of psychotherapeutic communications, this interest comes into play primarily when the psychotherapist is required to warn a potential victim of a patient’s propensity for violence. However, as we noted in Littleton, 39 Ohio St.3d at 98, 529 N.E.2d at 459, fn. 19, “an exception [to confidentiality] 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 [Breach of Medical Confidence in Ohio (1986), 19 Akron L.Rev. 373].
“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.” See, also, Paquin, Confidentiality and Privilege: The Status of Social Workers in Ohio (1992), 19 Ohio N.U.L.Rev. 199, 241-243; Givelber, Bowers & Blitch, supra, 1984 Wis.L.Rev. at 468-472.
*304G
R.C. 5122.34 provides that:
“Persons, including, but not limited to, boards of alcohol, drug addiction, and mental health services and community mental health agencies, 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, determination of appropriate placement, 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. No person shall be liable for any harm that results to any other person as a result of failing to disclose any confidential information about a mental health client, or failing to otherwise attempt to protect such other person from harm by such client. This section applies to expert witnesses who testify at hearings under this chapter.” (Emphasis added.)
The emphasized language represents the relevant changes made by the General Assembly under Am.Sub.S.B. No. 156. 142 Ohio Laws, Part I, 838. Am.Sub. S.B. No. 156 was passed on March 17,1988, and was generally effective on March 28, 1988, the date it was approved by the Governor. Id. at 852-853. However, R.C. 5122.34 as amended by that bill became effective on July 1,1989. Id. at 848. Thus, at the time we decided Littleton, supra, amended R.C. 5122.34 had already passed, but was not yet effective. See id., 39 Ohio St.3d at 95, 529 N.E.2d at 457, fn. 8.
FFCC argues that, pursuant to the first sentence of R.C. 5122.34, its decision with respect to whether Matt should have been hospitalized against his will is immunized to the extent it was rendered in good faith. In this regard, FFCC contends that the first sentence of R.C. 5122.34 operates not only to provide limited immunity for decisions to hospitalize a mental patient, but also for decisions not to hospitalize.
Although FFCC makes no argument with respect to the second sentence of R.C. 5122.34 added by Am.Sub.S.B. No. 156, two legal writers have suggested that this additional language purports to preclude Tarasoff-type liability. Eagle & Kirkman, Baldwin’s Ohio Mental Health Law (2 Ed.1990) 127-129, Section T 3.04; Hulteng, The Duty to Warn or Hospitalize: The New Scope of Tarasoff Liability in Michigan (1989), 67 U. of Det.L.Rev. 1, 11. We consider these issues together because they suffer from the same general flaw of attempting to extend R.C. 5122.34 beyond the boundaries set by the General Assembly.
R.C. 5122.34 does not apply to immunize mental health professionals from liability in all contexts. It operates only in the area of civil commitment. It applies to mental health professionals “who procedurally or physically assist in *305the hospitalization or discharge, determination of appropriate placement, or in judicial proceedings of a person under this chapter.” The decision not to initiate commitment procedures is a decision not to become involved in proceedings under R.C. Chapter 5122, thus taking the psychotherapist outside the protection of R.C. 5122.34. Accordingly, when FFCC itself determined that Matt was not eommittable, it decided not to “assist” in proceedings “under this chapter.”
Similarly, in granting unqualified immunity to mental health professionals for failing to warn or otherwise protect others from harm caused by their client, the second sentence of R.C. 5122.34 presupposes that affirmative action was taken under R.C. Chapter 5122. In other words, the General Assembly has made a policy decision that those who assist in hospitalizing, placing or discharging a person under R.C. Chapter 5122 are not hable for failure to warn or protect. Thus, a mental health professional who has not assisted in having a mental patient committed, placed, or discharged under R.C. Chapter 5122 is not entitled to the immunity bestowed by R.C. 5122.34.
Accordingly, FFCC is not entitled to qualified immunity under the first sentence of R.C. 5122.34 although, as set forth infra, it is entitled to what amounts to the same thing under the “psychotherapist judgment rule.” More important, the second sentence of R.C. 5122.34 does not preclude the finding that a special relation exists between the psychotherapist and the outpatient which imposes a common-law duty on the therapist to take affirmative steps to control the patient’s violent conduct.
H
We conclude that the relationship between the psychotherapist and the patient in the outpatient setting constitutes a special relation justifying the imposition of a duty upon the psychotherapist to protect against and/or control the patient’s violent propensities. The outpatient setting embodies sufficient elements of control to warrant the imposition of such a duty, and such a duty would serve the public’s interest in protection from the violently inclined mental patient in a manner that is consistent with Ohio law. On the other hand, the imposition of such a duty would not impose undue burdens upon the therapist or result in overcommitment, nor would it significantly affect the confidential character of psychotherapeutic communications.
II
THE “PSYCHOTHERAPIST JUDGMENT” RULE
In Littleton, 39 Ohio St.3d 86, 529 N.E.2d 449, we balanced the various policy considerations in determining the appropriate standard of care to be applied to *306psychiatric decisions releasing voluntarily hospitalized mental patients from the hospital. The court concluded that the malpractice standard of ordinary care should not be applied, but that a “professional judgment standard” should be applied similar to the “business judgment rule.” Accordingly, the holding of Littleton describes an application of the professional judgment standard to three different situations in the voluntary hospitalization setting.
We explained in part as follows:
“ ‘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., 39 Ohio St.3d at 96, 529 N.E.2d at 458, quoting Currie v. United States (M.D.N.C.1986), 644 F.Supp. 1074, 1083, affirmed on alternative grounds, Currie, supra, 836 F.2d 209.
Similarly, in Tarasoff, supra, 17 Cal.3d at 438, 131 Cal.Rptr. at 25, 551 P.2d at 345, the court explained that:
“Within the broad range of reasonable practice and treatment in which professional opinion and judgment may differ, the therapist is free to exercise his or her own best judgment without liability; proof, aided by hindsight, that he or she judged wrongly is insufficient to establish negligence.”
Plaintiffs-appellants argue that the good faith judgment rule applies only to decisions on “whether hospitalization of the patient is required,” and only in “those cases in which a standard of care cannot be determined by the expert witnesses.” We disagree.
The nature of the duty imposed requires the therapist to determine the interrelated questions of whether a patient poses a risk of harm to others and how to prevent such harm from coming to fruition. It does not necessarily require in all cases that the therapist hospitalize or medicate the patient, warn others, or take any single specific step to avert the danger. What is required, if anything is required, depends upon the facts and allegations of each case. Thus, when courts indiscriminately refer to the “duty to warn” or the “duty to hospitalize,” they are merely describing how particular therapists, under particular circumstances, failed to fulfill their overall duty. We reject the notion that divergent standards of care should be applied to what is essentially a single duty.
*307The psychotherapist judgment rule is a compromise between “declining to recognize a duty at all, or recognizing a duty and liability under a traditional negligence theory.” Currie, 644 F.Supp. at 1083. The overriding concern is to protect the public from assault by the violent mentally ill. Fundamentally, the duty is imposed because the therapist is the best, if not the only, line of defense society has against the danger posed by the violent mental patient. Because of their special training, skill and contact with the patient, psychotherapists are especially equipped to thwart the danger.
However, the public interest would not be served by a standard of care which intrudes upon the integrity of the psychotherapeutic relationship. To maintain its efficacy, the psychotherapeutic relationship must instill confidence in the patient and afford a certain degree of autonomy for the therapist in rendering decisions. Although the concerns over predictability, overcommitment, and confidentiality are overstated, they appear nevertheless to cause some trepidation on the part of the patient and therapist alike which can affect the viability of the relationship. See, generally, Givelber, Bowers & Blitch, supra, 1984 Wis.L.Rev. 443; Stone, The Tarasoff Decisions: Suing Psychotherapists to Safeguard Society, supra, 90 Harv.L.Rev. 358. Society’s interest in being free from harm caused by mentally ill persons is, in the final analysis, dependent upon “[t]he patient’s right to good medical care.” Littleton, 39 Ohio St.3d at 97, 529 N.E.2d at 459. Without good medical care, society would stand unprotected from mental patients with violent propensities.
Thus, the professional judgment rule adopted in Littleton seeks to strike an appropriate balance by not allowing the psychotherapist to act in careless disregard of the harm presented to others by violently inclined patients, yet preserving the confidence, autonomy, and flexibility necessary to the psychotherapeutic relationship. There is nothing in the analysis itself that would suggest a different result in the outpatient setting.
Accordingly, we hold that when a psychotherapist knows or should know that his or her outpatient represents a substantial risk of harm to others, the therapist is under a duty to exercise his or her best professional judgment to prevent such harm from occurring.7
*308Ill
THE “SPECIFIC THREATS TO SPECIFIC VICTIMS” RULE
Dr. Brown urges the court to adopt the “specific threats to specific victims” rule articulated in Brady, 751 F.2d at 331, and Thompson v. Alameda Cty. (1980), 27 Cal.3d 741, 167 Cal.Rptr. 70, 614 P.2d 728. Under this rule, “absent allegations in a complaint that a psychiatrist is aware of his patient’s specific threats to specific victims, there exists no legal duty or obligation on the part of the psychiatrist for harm done by the patient.” Brady, 751 F.2d at 331.
The rule, at least as articulated in Brady, encompasses three elements as a precondition to liability: (A) that the therapist is actually aware that the patient represents a threat of harm to others, (B) that the threat of harm be specific, and (C) that the target of such threats be precisely and specifically identified. Although these elements tend to overlap and meld together in application, for purposes of analysis we will consider them separately.
A
Awareness
In Ohio, a cause of action for professional negligence is not dependent upon actual awareness of the potential for harm. Instead, the duty to protect others is imposed when the medical professional knows or should know that the patient is likely to cause harm to others. See Littleton, 39 Ohio St.3d at 92, 529 N.E.2d at 455; Jones, supra, 118 Ohio St. at 153, 160 N.E. at 458. Moreover, as the court in Bardoni, 151 Mich.App. at 180, 390 N.W.2d at 223-224, aptly explained:
“[I]f a duty to take reasonable steps to protect a third person is limited only to those victims which are actually known to the psychiatrist, * * * an extremely negligent psychiatrist may not ascertain that the patient is even dangerous or that the patient is dangerous to anyone in particular. For example: a psychiatric hospital’s records of a patient indicate that the patient has violently focused his aggression on a named or particular person, * * * but the treating psychiatrist fails to read the records and is not apprised of this fact. Thus the target of the violence is identified or readily identifiable, but the defendant psychiatrist is not even aware that there exists a particular target of his patient’s aggression. The relevant determination then becomes whether the psychiatrist should have ascer*309tained, by acting in accordance with the standards of his profession, e.g., by reading the hospital records, that there existed a target of his patient’s aggression and the identity of the target. Thus, whether the treating psychiatrist actually knows of a target and whether that target is actually identified or readily identifiable by the psychiatrist is not always the appropriate focus in determining the extent of a psychiatrist’s duty to persons endangered by his patient.” (Citations omitted.)
B
Specific Threats
R.C. 5122.01 provides as follows:
“(B) ‘Mentally ill person subject to hospitalization by court order’ means a mentally ill person who, because of his illness:
tt # H* ‡
“(2) Represents a substantial risk of physical harm to others as manifested by evidence of recent homicidal or other violent behavior, evidence of recent threats that place another in reasonable fear of violent behavior and serious physical harm, or other evidence of present dangerousness.”
Thus, in Ohio evidence of specific threats is not necessarily required in order to hospitalize a mentally ill patient against his will. A mentally ill patient is also subject to involuntary commitment when his dangerous proclivities are found to be imminent by virtue of “evidence of * * * other violent behavior * * * or other evidence of present dangerousness.” It would indeed be curious to apply a stricter standard as a precondition to the psychotherapist’s common-law duty to control, especially since that duty, in its most confining aspect, would require the therapist to institute those very commitment proceedings. Moreover, there are obviously circumstances where, despite the lack of a specific threat, it can nevertheless be ascertained that the patient presents a danger to others and who the likely victims will be. See Hammam, supra, 161 Ariz. at 64, 775 P.2d at 1128; Jablonski, supra, 712 F.2d at 398. “Society must not become the victim of a dangerous patient’s ambiguity.” Schuster, supra, 144 Wis.2d at 256, 424 N.W.2d at 172-173.
C
Specifically Targeted Victims
The advent of Thompson and Brady has compelled courts to confront the question of whether and when the “readily identifiable victim” rule is applicable and, if applicable, what degree of specificity is needed. The parties in the case sub judice tend to categorize the various cases which have addressed this issue *310according to whether they adopt the readily identifiable victim rule or reject it in favor of a more encompassing standard of foreseeability. Having drawn this line of demarcation, the parties would have us choose between what they perceive to be opposing options. However, we find the query as to which is the better option to be a misdirected question in this case. The views seen by the parties as being divergent are for the most part separate standards applied in different contexts.
The “readily identifiable victim” rule has its genesis in Thompson, where it was applied only to allegations concerning the failure to warn, as the allegations concerning the defendants’ failure to otherwise control the patient’s violent conduct were disposed of on other grounds. The court in Thompson summarized its holding by explaining that “[wjithin this context and for policy reasons the duty to warn depends upon and arises from the existence of a prior threat to a specific identifiable victim.” Id., 27 Cal.3d at 758, 167 Cal.Rptr. at 80, 614 P.2d at 738. The court reasoned that requiring warnings to the public at large would be “unwieldy and of little practical value,” producing “a cacophony of warnings that by reason of their sheer volume would add little to the effective protection of the public.” Id., 27 Cal.3d at 754-755, 167 Cal.Rptr. at 77, 614 P.2d at 735.
After Thompson, cases that have limited the duty of care to identifiable victims have focused on the failure to warn. Within this context, they have also restricted the application of the rule, allowing a psychotherapist to be held liable to others in close relationship or close proximity to the victim, or finding that certain circumstances fall between the extremes of requiring warnings to be issued to the public at large and limiting it to a specifically named victim. On the other hand, cases that have rejected the rale have focused on a myriad of other alternative precautions alleged to have been reasonably necessary to control the patient. These cases vary in the standards of foreseeability that they impose, ranging from limiting the scope of the duty to those within the “zone of danger” to expanding it to cover the general public or all the world.
At first blush, the cases may lead one to believe that a choice must be made between adopting or rejecting the readily identifiable victim rule across the board. This is because many of the cases involve allegations concerning the psychotherapist’s failure to take a number of steps to control the patient or protect the potential victim in addition to the giving of warnings. Upon closer scrutiny, however, it becomes apparent that the “readily identifiable victim” rule is applied only in the context of failure to warn. On the other hand, when the propriety of other steps alleged to have been required are examined, the focus shifts away from the special policy considerations and practical difficulties which have been said to justify limiting the scope of the duty in failure-to-warn cases. Thus, the “readily identifiable victim” rule is born, lives, and grows in failure-to-warn cases. Wofford, 795 P.2d at 520; Hamman, 161 Ariz. at 64, 775 P.2d at *3111128; Perreira, 768 P.2d at 1210; Bardoni, 151 Mich.App. at 180, 390 N.W.2d at 223-224; Neakok, 721 P.2d at 1128-1129, 1131-1132; Petersen, 100 Wash.2d at 428, 671 P.2d at 237; Hedlund v. Orange Cty. Superior Court (1983), 34 Cal.3d 695, 705, 194 Cal.Rptr. 805, 810, 669 P.2d 41, 46-47; Jablonski 712 F.2d at 398; Lipari, 497 F.Supp. at 194-195; Bradley Ctr., Inc., 250 Ga. at 200-202, 296 S.E.2d at 695-696. See, also, Reisner v. Regents of the Univ. of California (1995), 31 Cal.App.4th 1195, 1199-1200, 37 Cal.Rptr.2d 518, 520-521.
The appropriate question, therefore, is not which approach should be chosen, but whether it is wise to extend the readily identifiable victim rule beyond the failure-to-warn case. In considering this issue, we find ourselves in agreement with the following well-reasoned analysis of Currie, supra, 644 F.Supp. at 1079-1080:
“One cannot read Thompson without being struck by the emphasis the court placed on the practical difficulties which the lack of a ‘readily identifiable’ victim placed on a duty to warn. Without knowing specifically who the victim would be, the therapist could not know who to warn. The court discussed the possibility of a general warning, but noted that the effectiveness of such a warning would be greatly reduced because people would pay less attention to them. 27 Cal.3d at 753-760, 167 Cal.Rptr. at 77-80, 614 P.2d at 735-738. In sum, the court found that, unlike in Tarasoff, Thompson’s reduced benefits of general warnings would not outweigh the problems such as breach of patient confidentiality, now worsened by broader dissemination of his threats.
“This conclusion cannot be reached in a ‘duty to commit’ balancing test. Unlike a duty to warn case, in which the therapist needs to know the identity of the victim in order to adequately act, the therapist in a duty to commit case need only know that the patient is dangerous generally in order to adequately commit him. As a practical matter, the victim’s identity is irrelevant to whether the doctor can adequately act — by committing the patient, the therapist is able to protect all possible victims.
“The court does not believe that it is wise to limit any duty to commit according to the victim. Arguably, the patient who will loll wildly (rather than specifically identifiable victims) is the one most in need of confinement. In negligent release cases, a defendant’s duty generally has not been limited to readily identifiable victims, and the court believes a similar rule is appropriate here. Citizens outside of the ‘readily identifiable’ sphere but still within the ‘foreseeable zone of danger’ are potential victims a therapist should consider if he has a duty to them and a means of adequately protecting them.” (Citations and footnote omitted; emphasis sic.) See, also, Schuster, 144 Wis.2d at 256-259, 424 N.W.2d at 172-174.
*312We need not determine at this time whether and to what extent the readily identifiable victim rule should attach in a failure-to-warn case. The case sub judice does not involve any allegation that Dr. Brown or FFCC was negligent in failing to warn Matt’s family. The policy considerations and practical difficulties leading to a limitation on the duty in failure-to-warn cases are not present here. Accordingly, we reject such a limitation in this case. “ ‘It is not necessary that the defendant should have anticipated the particular injury. It is sufficient that his act is likely to result in an injury to someone.’ ” Di Gildo v. Caponi (1969), 18 Ohio St.2d 125, 130, 47 O.O.2d 282, 285, 247 N.E.2d 732, 736, quoting Neff Lumber Co. v. First Natl. Bank (1930), 122 Ohio St. 302, 308-310, 171 N.E. 327, 330.
IV
REMOTENESS
Dr. Brown also argues that he owed no duty to plaintiffs-appellants as a matter of law because the shootings took place over nine months after the last time he treated Matt. Thus, he contends, it cannot be found that the necessary element of control was present during those nine months, especially since FFCC took over Matt’s care.
This argument commingles two distinct issues, that of duty and proximate cause. The control portion of the argument goes to the issue of duty. In this regard, the argument may be reduced to a proposition that there can be no duty unless the patient was under the care of the therapist at the time the harm was inflicted. Dr. Brown cites no authority for this position. Moreover, such a proposition runs counter to negligence principles. Viewing the facts most favorably to plaintiffs-appellants, Civ.R. 56(C), the gravamen of Dr. Brown’s alleged negligence in this case is the very act of withdrawing medication and relinquishing care of Matt. It is clearly unsound to absolve a negligent defendant because of the very act which made his conduct negligent.
The issue of remoteness in time — a question of proximate cause — cannot be resolved as a matter of law. As Prosser & Keeton explains:
“Remoteness in time or space may give rise to the likelihood that other intervening causes have taken over the responsibility. But when causation is found, and other factors are eliminated, it is not easy to discover any merit whatever in the contention that such physical remoteness should of itself bar recovery. The defendant who sets a bomb which explodes ten years later, or mails a box of poisoned chocolátes from California to Delaware, has caused the result, and should obviously bear the consequences.” (Footnotes omitted.) Prosser & Keeton on Torts, supra, at 283, Section 43.
*313Thus, no fixed rule can be established as to how quickly the harm must occur in order to hold the defendant liable. Some courts have found periods ranging between three and a half months to two years and five months to be too remote, while other courts have found periods ranging from five and one-half months to three years not to be too remote. Reisner, 31 Cal.App.4th at 1201, 37 Cal. Rptr.2d at 521 (three years not too remote); Wofford, 795 P.2d at 521 (two years and five months too remote. Also, notes that a lapse of time of three and one-half months has been found too remote); Wagshall, 148 A.D.2d at 446, 538 N.Y.S.2d at 598 (seven to eight months too remote); Naidu, 539 A.2d at 1075 (five and one-half months not too remote); Neakok, 721 P.2d at 1123, 1135-1136 (six months not too remote); Lundgren, 354 N.W.2d at 28-29 (one and one-half years not too remote).
Physical or temporal remoteness, therefore, may be an important consideration in whether negligent conduct is a substantial factor in producing harm; but the mere lapse of time, in the absence of intervening causes, is not of itself sufficient to prevent the defendant’s negligence from being the legal cause, regardless of how much time has passed. Naidu, supra, 539 A.2d at 1075, Restatement of Torts, supra, at 434, Section 433, Comment f; 2 American Jurisprudence, Proof of Facts 3d, supra, at 344, Section 4.
In light of the testimony of Drs. Goff and Tanay that part of Dr. Brown’s negligence lay in the relinquishment of Matt’s care to FFCC and that Dr. Brown’s negligence did contribute to the events of July 25, 1991, it cannot be found as a matter of law that FFCC’s actions constitute an intervening cause. Accordingly, there remains a genuine issue of material fact as to whether Dr. Brown’s alleged negligence in treating Matt was a substantial factor in producing the injuries and deaths occurring on July 25,1991.
V
APPLICATION OF STANDARD OF CARE TO FACTS
Viewing the evidence most favorably to plaintiffs-appellants, we find that reasonable minds could conclude that Dr. Brown failed to exercise his best professional judgment in diagnosing and controlling Matt’s propensity for violence. When Dr. Brown first saw Matt on July 19, 1990, he noted that Matt was “recently discharged from a mental health unit of some sort in Philadelphia, Pennsylvania, on Elavil and Navane,” and sought to obtain the C.A.T.C.H. records which he received sometime prior to Matt’s next scheduled appointment on August 16, 1990. According to Drs. Goff and Tanay, those records revealed that Matt was a medication-controlled schizophrenic who, without medication, would more likely than not have a relapse of his psychosis, placing him at substantial risk for conflict with his parents and potential violence.
*314However, instead of reading the C.A.T.C.H. records or contacting Dr. Ladenheim, Dr. Brown noted on October 11,1990, the date of his last appointment with Matt, that he had “no further insight as to just what happened to get him [Matt] into the mental hospital in Philadelphia or why they may have prescribed neuroleptic [medication] for him.” Thus, he diagnosed Matt in part as a malingerer intent on gaining Social Security disability benefits; and weaned him off his medication.
In Littleton, supra, 39 Ohio St.3d at 96, 529 N.E.2d at 458, we relied in part upon Bell v. New York City Health & Hospitals Corp. (1982), 90 A.D.2d 270, 456 N.Y.S.2d 787, in constructing the professional judgment standard to be applied in the hospital setting. In Bell, the court explained that:
“A decision that is without proper medical foundation, that is, one which is not the product of a careful examination, is not to be legally insulated as a professional medical judgment. Stated otherwise, ‘[p]hysicians are not liable for mistakes in professional judgment, provided that they do what they think best after careful examination. * * * However, liability can ensue if their judgment is not based upon intelligence and thus there is a failure to exercise any professional judgment.’ ” (Citations omitted and emphasis sic.) Id., 90 A.D.2d at 280-281, 456 N.Y.S.2d at 794.
Applying the standard to the facts of the case, the court in Bell concluded that:
“Dr. Hermann [did not] make any attempt to ascertain from the Veteran’s Administration Hospital, the patient’s medical history and treatment, even though he noted that Bell ‘refuse[d] to discuss his past psychiatric history.’ Suicidal ideations or tendencies may well have gone undetected as a result of inadequate examination.” 90 A.D.2d at 284, 456 N.Y.S.2d at 796.
Similarly, the failure to review prior medical records has played a prominent part in a number of cases holding that the psychotherapist breached his duty to others to diagnose and control the patient’s violent propensities. Hamman, 161 Ariz. at 59, 775 P.2d at 1123; Naidu, 539 A.2d at 1069, 1073; Peck, 146 Vt. at 66, 499 A.2d at 426; Jablonski, 712 F.2d at 393, 394, 398. See, also, Bardoni, 151 Mich.App. at 180, 390 N.W.2d at 224.
Accordingly, we find that reasonable jurors could conclude that Dr. Brown’s failure to review the C.A.T.C.H. records or contact Dr. Ladenheim to obtain a thorough and accurate history on Matt amounted to something less than the exercise of professional judgment. Thus, summary judgment was entered improperly in favor of Dr. Brown, and the decision of the court of appeals is affirmed as to this issue.
We find further that summary judgment in favor of FFCC was inappropriate for several reasons. First, the testimony of Drs. Goff and Tanay, if believed, *315reveals that FFCC and its employees conducted a meaningless evaluation into whether Matt was subject to involuntary hospitalization in May or July 1991. Dr. Goff testified that Matt was committable by June 14, 1991, and “that a vocational therapist has no business making decisions about what is appropriate and what’s not appropriate in terms of involuntary hospitalization.” Dr. Tanay testified that Matt was likely committable in May 1991, and that Lambert and Reid “had no business doing what they were doing. And [FFCC] had no business assigning them that role. * * * This was a responsibility absolutely beyond their competence.” Instead, he opined, at least one psychiatrist or psychologist well-trained ■ in the area of dealing with psychotic individuals is required for a meaningful evaluation of whether a patient is subject to involuntary hospitalization.
This also is the type of situation which Littleton contemplates as indicating the absence of professional medical judgment. In Littleton, we placed reliance in part on Cohen v. New York (1976), 51 A.D.2d 494, 382 N.Y.S.2d 128, affirmed (1977), 41 N.Y.2d 1086, 396 N.Y.S.2d 363, 364 N.E.2d 1134, in constructing the psychotherapist judgment rule. Id., 39 Ohio St.3d at 96, 529 N.E.2d at 458. In Cohen, the court found the absence of professional medical judgment where the patient, a schizophrenic who committed suicide on the date he was let out of the psychiatric ward, had not been evaluated by a qualified psychiatrist. The court explained as follows:
“The determinative factual issue is whether or not a qualified psychiatrist was actively supervising the care of the decedent. * * *
« * * *
“While this record establishes that patients were discussed at team meetings, it is not established that any evaluation of this decedent’s suicide propensities was made by a qualified psychiatrist during his stay at the hospital. * * *
“Wfiiat happened in this case was the result of ill-defined policy, in the present circumstances, as to the power of floor nurses to impose restraints on a patient’s freedom and requiring full evaluation by qualified psychiatrists at periodic intervals. The fault herein is not with Doctors Bjork, Rosenberg or Sverd as individuals, but rather with the lack of policies requiring more direct management of a patient’s treatment by a qualified psychiatrist.” Id., 51 A.D.2d at 496-497, 382 N.Y.S.2d at 130.
The second reason why summary judgment in favor of FFCC was inappropriate is that there is sufficient testimony which, if believed, establishes that FFCC was ill-equipped to deal with psychotic individuals such as Matt. Dr. Tanay testified that FFCC “was not equipped to deal with psychotics. It had neither the resources, and I would even believe not even the interest, in providing care for psychotics. * * * [Instead] psychotics were used, exploited to get funds, but *316they were provided virtually no meaningful care.” Explaining further, Dr. Tanay stated:
“[The] combination of a psychiatrist unable to make a diagnosis of serious mental illness, and a facility staffed by non-medical personnel, was a disaster waiting to happen. * * * [W]e have here a setup, an organizational structure that cannot help but result in tragic consequences.”
Third, although disputed, social worker Barbara Sharp testified that FFCC had an unwritten policy that it would not initiate involuntary hospitalization proceedings, but would become involved only after such were initiated by the patient’s family. The court of appeals found that “there is no evidence that this unwritten policy guided the counseling center in this case.” We find, however, that there is sufficient evidence which allows for an inference that FFCC acted in accordance with such a policy.
FFCC received repeated notices of Matt’s deteriorating condition. It was informed of Matt’s violent outbursts and of his deposit on the purchase of a gun. Mrs. Morgan specifically stated her concern that Matt may become violent, and repeatedly begged FFCC to do something before violence erupted. Instead of taking her seriously, Lambert believed that she was exaggerating and emergency assessments were performed by what Drs. Goff and Tanay described as untrained personnel. Under these circumstances, a reasonable juror could infer that FFCC’s unwritten policy guided their actions.
In light of the foregoing, we find that a genuine issue of material fact remains as to whether FFCC exercised professional judgment in evaluating and treating Matt. Accordingly, we hold that summary judgment was inappropriate in favor of FFCC and its employees, and reverse the court of appeals as to this issue.
VI
ORDER OF “TRIFURCATION”
By separate orders, the trial court “trifurcated” the issues of liability, compensatory damages, and punitive damages. Plaintiffs-appellants argue that “[t]he language of Civ.R. 42(B) clearly places the burden on the moving party to demonstrate that it will suffer prejudice if a request for a separate trial is not granted.”
The entries by the trial court indicate that it was concerned with convenience and judicial economy. Civ.R. 42(B) allows the trial court to order separate trials “in furtherance of convenience or to avoid prejudice, or when, separate trials will be conducive to expedition and economy.” Thus, we agree with the following conclusion made by the court of appeals:
*317“Although we question whether such an order serves sound judicial administration, we cannot conclude that such an order is an abuse of discretion as that term is commonly defined.” See State v. Jenkins (1984), 15 Ohio St.3d 164, 222, 15 OBR 311, 361, 473 N.E.2d 264, 313; Blakemore v. Blakemore (1983), 5 Ohio St.3d 217, 219, 5 OBR 481, 482, 450 N.E.2d 1140, 1142.
In light of all of the foregoing, we affirm the decision of the court of appeals in part, and reverse it in part, as indicated herein, and remand the cause to the trial court for further proceedings not inconsistent with this opinion.
Judgment affirmed in part, reversed in part and cause remanded.
Douglas, F.E. Sweeney and Pfeifer, JJ., concur. Moyer, C.J., Cook and Stratton, JJ., dissent.. The rule has its genesis in the early common-law distinction between “misfeasance” and “nonfeasance.” While there is a duty to refrain from active misconduct working positive injury on others, there is no duty to take affirmative action to aid or protect another from harm. Regardless of the gravity of danger threatening the potential victim, or the relative ease with which the danger can be averted, the traditional rule simply refuses to impose a moral obligation upon a stranger. Thus, e.g., the expert swimmer, with a boat and rope at his disposal, who sees another drowning before his eyes, is under no obligation to help him. Instead he may sit on the dock and watch the man drown. Similarly, an ordinary bystander is under no duty to rescue a child drowning in what he knows to be shallow water, or to prevent a neighbor’s child from hammering on a dangerous explosive. Such decisions have been condemned by legal writers as “shocking in the extreme,” “revolting to any moral sense,” and representative of “an attitude of rugged, perhaps heartless individualism.” Nevertheless, because of the difficulties of setting moral standards and devising workable rules to cover all situations, the rule remains the law. See 2 Restatement of Torts, supra, at 116-117, Section 314, Comment c; Prosser & Keeton on Torts, supra, at 375-376, Section 56; 3 Harper, James & Gray, The Law of Torts (2 Ed.1986) 718-719, Section 18.6; Harper & Kime, The Duty to Control the Conduct of Another, supra, 43 Yale L. J. 886, 887.
*294The American Law Institute explains, however, that “[i]t appears inevitable that, sooner or later, such extreme cases of morally outrageous and indefensible conduct will arise that there will be further inroads upon the older rule.” 2 Restatement of Torts, supra, at 117, Section 314, Comment a At least two courts have already imposed a duty to summon aid for the benefit of another. In Soldano v. O’Daniels (1983), 141 Cal.App.3d 443, 449, 190 Cal.Rptr. 310, 314, the court felt it was “time to re-examine the common law rule of nonliability for nonfeasance in the special circumstances of the instant case.” In concluding, the court stated that “[t]he imposition of liability on the defendant in this case * * * is but a slight departure from the ‘morally questionable’ rule of nonliability for inaction absent a special relationship.” Id., 141 Cal.App.3d at 455, 190 Cal.Rptr. at 318. In Griffith v. Southland Corp. (1992), 94 Md.App. 242, 257, 617 A.2d 598, 606, the court explained that “there is no precedent which permits a bystander to refuse to call 911 when not exposed to imminent danger. Even if there were such an uncivilized and shocking principle, blind allegiance would invite disdain and disrespect for the courts.”
. This analysis has prompted some discussion by legal scholars over whether and to what extent Tarasoff affects traditional notions of duty. See, e.g., Lake, Revisiting Tarasoff (1994), 58 Albany L.Rev. 97; Adler, Relying Upon the Reasonableness of Strangers: Some Observations about the Current State of Common Law Affirmative Duties to Aid or Protect Others, 1991 Wis.L.Rev. 867, 886-898; Murphy, Evolution of the Duty of Care: Some Thoughts (1981), 30 DePaul L.Rev. 147, 169-176.
. For example, in Hasenei, supra, 541 F.Supp. at 1009-1010, the court supported its conclusion by observing that “lack of control by the therapist and maximum freedom for the patient is oft times the end sought by both the psychiatric profession and the law.” Id., citing O’Connor v. Donaldson (1975), 422 U.S. 563, 95 S.Ct. 2486, 45 L.Ed.2d 396. In O’Connor, the high court held that a state cannot constitutionally confine, without more, a nondangerous mentally ill individual who can live safely in freedom. O’Connor is cleai’ly inapplicable to the dangerous individual. The Supreme Court spoke in terms of “locking a person up,” “simple custodial confinement,” and “fencfing] in the harmless mentally ill.” Id., 422 U.S. at 575, 95 S.Ct. at 2493-2494, 45 L.Ed.2d at 407. By relying on O’Connor, the court in Hasenei apparently felt that the level of control needed to satisfy Section 315 is that of actual constraint or confinement. More directly, the court in Santa Cruz, 590 So.2d at 445, reasoned that “[t]his case does not meet section 319 standards. [The patient] was not in Northwest Dade’s custody.” The remaining cases do not directly address the issue other than to conclude, without explanation, that the ability to control is lacking in the outpatient setting.
. This does not mean, however, that the psychotherapist has a duty to control the violent propensities of the outpatient merely because he has the ability to control. If the ability to control automatically gave rise to a corresponding duty to control, the general rule of nonliability for nonfeasance would be undermined. The ability to control is a necessary but not a sufficient basis for finding a special relation. As explained in 2 Restatement of Torts, supra, at 123, Section 315, Comment b:
“In the absence of either one of the kinds of special relations described in this Section, the actor is not subject to liability if he fails, either intentionally or through inadvertence, to exercise his ability so to control the actions of third persons as to protect another from even the most serious harm. This is true although the actor realizes that he has the ability to control the conduct of a *301third person, and could do so with only the most trivial of efforts and without any inconvenience to himself.”
. In response to the question whether he disagreed that “ ‘[d]etermining 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,’ ” Dr. Tanay responded: “I think psychiatrists are capable of determining that someone has a high potential for being dangerous. If not, then I think we have wasted billions of dollars in employing psychiatrists in evaluating people before we commit them and before we hospitalize them.”
It is also worthy of comment that the mental health profession claims expertise in predicting the stability and responsibility of prospective employees. See, generally, Rouse & Butcher, Annotated Bibliography on the Use of the MMPUMMPI-2 in Personnel and Educational Selection (1995); Applying Psychology in Business: The Handbook for Managers and Human Resource Professionals (1990) 835, Chapter 83. It would be somewhat inconsistent to claim the ability to predict likely human behavior in the employment setting, based solely on a single test and/or interview, yet claim that a psychotherapist is without ability to predict dangerous propensities in the outpatient setting.
. In applying this standard, courts should not lose sight of the nature of the duty imposed. The psychotherapist is not necessarily free of liability merely because he or she considered one of several alternatives. What the psychotherapist judgment rule requires is that the therapist consider all alternatives. Otherwise, an evaluation can hardly be considered thorough. For example, where a therapist knows or should know that his or her patient poses a substantial risk of harm to others and decides not to hospitalize the patient, the therapist can still be found liable for failing to explore other options such as medicating the patient or warning potential victims. See Littleton, 39 Ohio St.3d at 102, 529 N.E.2d at 462-463 (Douglas, J., concurring in part and dissenting in part). Similarly, a decision is not necessarily based on professional judgment where *308the psychotherapist is aware or should be aware that the course of action decided upon would be ineffective, e.g., if medication is prescribed but the patient is not compliant, other alternatives may have to be considered. Also, it may become necessary to reevaluate a chosen course of treatment where signs of decompensation become evident. On the other hand, if the therapist, in good faith, chooses an alternative after thoroughly considering all the options, he or she will not be held liable merely because, in hindsight, the choice proved to be wrong.