Opinion

Six days after his birth, plaintiff suffered irreversible brain damage. Through his mother as guardian ad litem, he sued his pediatrician and the hospital in which he was bom. Before trial, plaintiff and the pediatrician agreed to a settlement of $1 million, the limit of the pediatrician’s malpractice insurance policy. At a jury trial, plaintiff was awarded both economic and noneconomic damages. The jury found that the pediatrician was 55 percent at fault, the hospital 40 percent at fault, and the parents 5 percent at fault.

On the hospital’s appeal, a major contention was that under the common law “release rule,” plaintiff’s settlement with the pediatrician also released the nonsettling hospital from liability for plaintiff’s economic damages. The Court of Appeal reluctantly agreed. It observed that although this court “has criticized the common law release rule,” it “has not abandoned it.” Considering itself bound by principles of stare decisis, the Court of Appeal then applied the common law release rule to this case, and it reversed that portion of the trial court’s judgment awarding plaintiff economic damages against the hospital. We granted plaintiff’s petition for review, which asked us, as the Court of Appeal did in its opinion, to repudiate the common law release rule. Today, we do so.

I

Helpful in our review of this case is the Court of Appeal’s lengthy and detailed explanation, not in dispute here, of the relevant facts and the medical conditions—jaundice, hyperbilirubinemia, and kemicterus—that led to plaintiff’s postbirth brain injury. Our brief summary follows.

A. Medical Conditions

The skin and eyes of an infant with jaundice have a yellowish tint, which may be caused by an accumulation in the blood of bilirubin, a waste substance produced by the normal breakdown of red blood cells. All infants have increasing levels of bilirubin for the first three to five days after birth. Unless an exacerbating condition exists, the bilimbin level reduces in about a week as the infant’s liver develops and the bilimbin is expelled. A common way of preventing a rise in the bilimbin level is to give the infant adequate milk, resulting in sufficient stool to expel the bilimbin.

Excessive bilimbin can lead to hyperbilirubinemia, in which bilimbin after migrating to the brain can cause kemictems, leading to severe brain damage. Hyperbilirubinemia is readily treatable by exposure to light (phototherapy), or in more serious cases by a blood-exchange transfusion. The risk of kemictems is higher for some infants than for others. The risk factors include these characteristics: (1) male, (2) East Asian descent, (3) bom at less than 38 weeks’ gestation, (4) exclusively breastfed, (5) braising, (6) jaundice within the first 24 hours, and (7) weight loss.

In April 2001, the Joint Commission (formerly the Joint Commission on Accreditation of Healthcare Organizations) issued Sentinel Event Alert, No. 18 (Alert No. 18) to warn the medical community of the reemergence of kemictems. The alert identified the various risk factors and recommended certain protective measures, such as medical checkups of all newborns within 24 to 48 hours of birth, and educating neonatal caregivers on the danger of and risk factors for kemicterus in newborn infants.

B. Facts Leading to Lawsuit

On Monday, March 24, 2003, Aidan Ming-Ho Leung, of East Asian descent, was bom at Verdugo Hills Hospital in Glendale, Los Angeles County. He was bom at less than 38 weeks’ gestation (37 weeks and two days). On the day of his birth, his mother, Nancy Leung, tried to breastfeed him five or six times, but she could not tell whether he was taking in milk. At least three times she expressed her concern to two of the attending nurses; two entries in Aidan’s hospital medical chart indicated problems with breastfeeding.

The next day, Aidan’s pediatrician, Steven Wayne Nishibayashi, examined Aidan at the hospital. Dr. Nishibayashi told the parents that Aidan was a healthy baby, that two bruises on the side of Aidan’s head were nothing to worry about, that it was safe to take Aidan home, and that a followup appointment should be made for the next week. Later that morning, about 24 hours after his birth, Aidan was discharged from the hospital. The hospital gave Aidan’s parents a manual entitled “Caring For Yourself and Your New Baby,” and the nurses told the parents to consult the manual if there were problems. When the parents arrived home, Aidan’s mother made an appointment for a followup visit with Dr. Nishibayashi for March 31, seven days after Aidan’s birth.

On Thursday, March 27, 2003, Aldan’s parents noticed that his eyes looked yellow and that his lips were chapped. They checked the care manual that the hospital had given them. The manual said that jaundice is common in newborns, that in most cases jaundice can be ignored, and that although jaundice can be dangerous, it rarely is so, depending on various factors such as age, premature birth, and “any other medical conditions.” The manual also stated that any bruises on the head were not dangerous and would heal in a few days, and that any questions about the baby’s jaundice should be directed to the baby’s treating physician.

That same day, Aidan’s mother telephoned the office of pediatrician Nishibayashi and told the responding nurse about Aidan’s yellowish tint. The nurse told her not to worry but said she would check with the doctor. When the nurse returned to the telephone, she asked whether Aidan was “feeding, peeing, and pooping.” Aidan’s mother responded, “Yes.” After saying that Aidan seemed fine, the nurse suggested putting Aidan in the sunlight. When Aidan’s mother mentioned his chapped lips, the nurse told her to apply lotion. When the mother asked whether she should bring Aidan in that day or wait for the scheduled appointment with Dr. Nishibayashi four days later, the nurse said to wait until that appointment.

The next day (Friday) and the day thereafter (Saturday), Aidan’s mother continued trying to breastfeed him and, as suggested by Dr. Nishibayashi’s office, put him in the sunlight, but the jaundice remained. By Saturday evening, Aidan appeared lethargic. Early Sunday, Aidan was very sleepy and would not wake up to be fed. His mother telephoned Dr. Nishibayashi’s office and left a message with his answering service. An on-call physician returned the call and, after listening to a description of Aidan’s symptoms, said to immediately take Aidan to the emergency room at Huntington Memorial Hospital in Pasadena. There Aidan was given a blood-exchange transfusion to reduce the level of bilirubin, but it was too late. Aidan had already developed kemicterus, resulting in severe brain damage.

C. The Lawsuit, the Trial, and the Court of Appeal’s Decision

Through his mother, Nancy, as guardian ad litem, Aidan brought a negligence action against his pediatrician and the hospital in which he was bom.

Before trial, plaintiff settled with defendant pediatrician for $1 million, the limit of the pediatrician’s malpractice insurance policy. Defendant pediatrician agreed to participate as a defendant at trial, and plaintiff agreed to release him from all claims. The pediatrician petitioned the trial court for a determination that the written settlement agreement met the statutory requirement of having been made in “good faith,” seeking to limit his liability to the amount of the settlement. (Code Civ. Proc., § 877 [judicial determination of settlement in good faith discharges the settling party “from all liability for any contribution to any other parties”].)

The trial court denied that motion, as it found the settlement to be “grossly disproportionate to the amount a reasonable person would estimate” the pediatrician’s share of liability would be. (See Tech-Bilt, Inc. v. Woodward-Clyde & Associates (1985) 38 Cal.3d 488, 499 [ 213 Cal.Rptr. 256 , 698 P.2d 159 ] (Tech-Bilt) [a settlement is in good faith when the trial court has determined it to be “within the reasonable range of the settling tortfeasor’s proportional share of comparative liability for the plaintiff’s injuries”].) Plaintiff and defendant pediatrician nevertheless decided to proceed with the settlement.

At trial, a jury found both defendant pediatrician and defendant hospital negligent. The jury awarded plaintiff $250,000 in noneconomic damages; $78,375.55 for past medical costs; $82,782,000 (with a present value of $14 million) for future medical costs; and $13.3 million (with a present value of $1,154,000) for loss of future earnings. The jury apportioned negligence as follows: 55 percent as to the pediatrician, 40 percent as to the hospital, and 2.5 percent as to each of Aidan’s parents. The judgment stated that, subject to a setoff of $1 million, representing the amount of settlement with the pediatrician, the hospital was jointly and severally liable for 95 percent of all economic damages awarded to plaintiff. Defendant hospital appealed, and plaintiff filed a cross-appeal.

The Court of Appeal agreed with defendant hospital that under the common law release rule, plaintiff’s settlement with, and release of liability claims against, defendant pediatrician also released nonsettling defendant hospital from liability for plaintiff’s economic damages. The court did so reluctantly, observing that although our court has criticized the common law release rule, it has not abandoned it. We granted plaintiff’s petition for review.

II

Under the traditional common law rule, a plaintiff’s settlement with, and release from liability of, one joint tortfeasor also releases from liability all other joint tortfeasors. That common law rule originated in England at a time when, under English law, a plaintiff could sue in a single action only those tortfeasors who had acted in concert against the plaintiff. In this context, the rule developed that if a joint tortfeasor paid compensation to the plaintiff, and received in exchange a release from liability, the remaining joint tortfeasors were also released. (Rest.2d Torts (appen.) § 885, reporter’s notes, p. 162.) The common law rule’s rationale is that there can be only one compensation for a single injury and because each joint tortfeasor is liable for all of the damage, any joint tortfeasor’s payment of compensation in any amount satisfies the plaintiff’s entire claim. (5 Witkin, Summary of Cal. Law (10th ed. 2005) Torts, § 70, pp. 142-143.)

The rule, however, can lead to harsh results. An example: A plaintiff might have settled with a joint tortfeasor for a sum far less than the plaintiff’s damages because of the tortfeasor’s inadequate financial resources. In that situation, the common law rule precluded the plaintiff from recovering damages from the remaining joint tortfeasors, thus denying the plaintiff full compensation for the plaintiff’s injuries. (Mesler v. Bragg Management Co. (1985) 39 Cal.3d 290, 298 [ 216 Cal.Rptr. 443 , 702 P.2d 601 ].) In an effort to avoid such unjust and inequitable results, California courts held that a plaintiff who settled with one of multiple tortfeasors could, by replacing the word “release” in the settlement agreement with the phrase “covenant not to sue,” and by stating that the agreement applied only to the parties to it, preserve the right to obtain additional compensation from the nonsettling joint tortfeasors. (Kincheloe v. Retail Credit Co., Inc. (1935) 4 Cal.2d 21, 23 [ 46 P.2d 971 ]; Lewis v. Johnson (1939) 12 Cal.2d 558, 562 [ 86 P.2d 99 ]; Holtz v. United Plumbing & Heating Co. (1957) 49 Cal.2d 501, 504 [ 319 P.2d 617 ]; 5 Witkin, Summary of Cal. Law, supra, Torts, § 70, p. 143.) As this court later recognized, “the distinction between a release and a covenant not to sue is entirely artificial.” (Pellett v. Sonotone Corp. (1945) 26 Cal.2d 705, 711 [ 160 P.2d 783 ].) We explained: “As between the parties to the agreement, the final result is the same in both cases, namely, that there is no further recovery from the defendant who makes the settlement, and the difference in the effect as to third parties is based mainly, if not entirely, on the fact that in one case there is an immediate release, whereas in the other there is merely an agreement not to prosecute a suit.” (Ibid.)

It was against that backdrop of criticism of the traditional common law release rule that the California Legislature in 1957 enacted Code of Civil Procedure section 877. (Stats. 1957, ch. 1700, § 1, p. 3077; see 5 Witkin, Summary of Cal. Law, supra, Torts, § 70, p. 142.) The statute modified the common law release rule by providing that a “good faith” settlement and release of one joint tortfeasor, rather than completely releasing other joint tortfeasors, merely reduces, by the settlement amount, the damages that the plaintiff may recover from the nonsettling joint tortfeasors, and that such a good faith settlement and release discharges the settling tortfeasor from all liability to others. (Code Civ. Proc., § 877, subds. (a), (b).) But because the statute governs only good faith settlements, and the trial court here determined that the settlement was not made in good faith (see ante, at p. 299), the statute does not apply to this case.

We reject defendant hospital’s contention that in enacting Code of Civil Procedure section 877 in 1957, the Legislature signaled an intent to preclude future judicial development of the law pertaining to settlements involving joint tortfeasors, thus preventing us from abrogating the common law release rule. As we observed in American Motorcycle Assn. v. Superior Court (1978) 20 Cal.3d 578 [ 146 Cal.Rptr. 182 , 578 P.2d 899 ] (American Motorcycle), nothing in the statute’s legislative history suggests an intent to foreclose the courts from rendering future decisions that would further the statute’s main purpose of ameliorating the harshness and the inequity of the common law rule at issue. (Id. at p. 601.) So also here, our decision today abrogating the common law release rule furthers the statute’s legislative purpose.

Here, adherence to the common law release rule would, as a result of plaintiff’s settlement with defendant pediatrician for $1 million (the limit of the pediatrician’s medical malpractice insurance policy), relieve nonsettling defendant hospital from any liability for plaintiff’s economic damages, even though the jury apportioned to the hospital 40 percent of the fault for plaintiff’s severe postbirth brain damage (assigning 55 percent of the fault to defendant pediatrician) and calculated plaintiff’s total economic damages at roughly $15 million. Under the common law release rule, plaintiff, injured for life through no fault of his own, would be compensated for only a tiny fraction of his total economic damages, a harsh result.

The rationale for the common law release rule was “that there could be only one compensation for a joint wrong and since each joint tortfeasor was responsible for the whole damage, payment by any one of them satisfied plaintiff’s claim against all.” (Tech-Bilt, supra, 38 Cal.3d at p. 493 ; see Lamoreux v. San Diego etc. Ry. Co. (1957) 48 Cal.2d 617, 624 [ 311 P.2d 1 ]; 5 Witkin, Summary of Cal. Law, supra, Torts, § 70, p. 142.) That rationale assumes that the amount paid in settlement to a plaintiff in return for releasing one joint tortfeasor from liability always provides full compensation for all of the plaintiff’s injuries, and that therefore anything recovered by the plaintiff beyond that amount necessarily constitutes a double or excess recovery. The assumption, however, is unjustified. For a variety of reasons— such as the settling defendant’s limited resources or relatively minor role in causing the plaintiff’s injury—a plaintiff may be willing to release one tortfeasor for an amount far less than the total necessary to fully compensate the plaintiff for all injuries incurred. As Dean Prosser observed in his criticism of the common law release rule: “There is a genuine distinction between a satisfaction and a release.” (Prosser & Keeton on Torts (5th ed. 1984) § 49, p. 332.)

In light of the unjust and inequitable results the common law release rule can bring about, as shown in this case, we hold that the rule is no longer to be followed in California.

We now consider our holding’s effect on the apportionment of liability among joint tortfeasors when, as here, one tortfeasor’s settlement, resulting in a release of liability, was determined by the trial court not to have been made in good faith, thus rendering inapplicable the apportionment scheme under Code of Civil Procedure section 877.

III

In deciding how to apportion liability in a negligence action when a plaintiff’s settlement with one of several defendants has been determined by a trial court not to have been made in good faith, we begin with two legal concepts that are central to California negligence law. The first concept is comparative fault, the second is joint and several liability. Under comparative fault, “liability for damage will be borne by those whose negligence caused it in direct proportion to their respective fault.” (Li v. Yellow Cab Co. (1975) 13 Cal.3d 804, 813 [ 119 Cal.Rptr. 858 , 532 P.2d 1226 ].) Under joint and several liability, “each tortfeasor whose negligence is a proximate cause of an indivisible injury remains individually liable for all compensable damages attributable to that injury.” (American Motorcycle, supra, 20 Cal.3d at p. 582 .)

As has been recognized, “[n]o perfect method exists for apportioning liability among a plaintiff, a settling tortfeasor, and a nonsettling tortfeasor.” (Rest.3d Torts, Apportionment of Liability, § 16, com. c, p. 133.) Three alternative approaches have developed. (Rest.2d Torts, § 886A, com. m, pp. 343-344.) We describe each.

Under the first approach, the money paid by the settling tortfeasor is credited against any damages assessed against the nonsettling tortfeasors, who are allowed to seek contribution from the settling tortfeasor for damages they have paid in excess of their equitable shares of liability. (Rest.2d Torts, § 886A, com. m, p. 343.) We will call this the setoff-with-contribution approach.

Under the second approach, as under the first, nonsettling tortfeasors are entitled to a credit in the amount paid by the settling tortfeasor. But, unlike under the first alternative, nonsettling tortfeasors may not obtain any contribution from the settling tortfeasor. (Rest.2d Torts, § 886A, com. m, p. 343.) We will call this the setoff-without-contribution approach.

The third approach differs from the first and second by subtracting from the damages assessed against nonsettling tortfeasors the settling tortfeasor’s proportionate share of liability, rather than the amount paid in settlement. (Rest.2d Torts, § 886A, com. m, p. 344.) We will call this the proportionate-share approach.

Which of these three approaches should we apply when, as here, one tortfeasor settles but another does not, and the settlement is judicially determined not to meet Code of Civil Procedure section 877’s good faith requirement? That is the issue we explore and resolve below.

The second approach—setoff without contribution by the settling tortfeasor to the nonsettling tortfeasor—is easy to dispose of, as it is not an option here. Although the Legislature has statutorily adopted this apportionment method, it has expressly limited its application to settlements made in good faith. (Code Civ. Proc., §§ 877, 877.6, subd. (c).) The Legislature did not define what “good faith” means in this context, but this court has held that the trial court should “inquire, among other things, whether the amount of the settlement is within the reasonable range of the settling tortfeasor’s proportional share of comparative liability for the plaintiff’s injuries.” ( Tech-Bilt, supra, 38 Cal.3d 488, 499 .) We further explained that the factors to be considered in determining whether a settlement was made in good faith include the settling defendant’s financial condition and insurance policy limits, a rough approximation of the plaintiff’s total recovery and the settling defendant’s proportionate share of fault, and any evidence of collusion or fraud, as well as “a recognition that a settlor should pay less in settlement than he would if he were found liable after a trial.” (Ibid.)

Here, as we noted earlier, the trial court ruled that plaintiff’s settlement with defendant pediatrician was not made in good faith. The pediatrician challenged that ruling by petitioning the Court of Appeal for a writ of mandate, but that court summarily denied the petition, and the issue is not before us here. For purposes of our discussion, we accept that the settlement was not in good faith. As we noted earlier, the Legislature has expressly limited the second apportionment method, setoff without contribution, to settlements made in good faith. (Code Civ. Proc., § 877.) Applying the setoff-without-contribution approach to settlements not made in good faith would effectively nullify that statutory provision. Consequently, that apportionment method is not available here.

Of the two remaining alternatives—setoff with contribution and proportionate share—nonsettling defendant hospital argues that we should adopt the latter. Plaintiff, however, prefers the setoff-with-contribution approach. In deciding which of these two alternatives to adopt, we evaluate their practical implications for tort plaintiffs and for settling and nonsettling joint tortfeasor defendants, their consistency with established legal principles, and their likely effect on the public policy to promote settlement and on the interest of judicial economy. (See McDermott, Inc. v. AmClyde (1994) 511 U.S. 202, 211 [ 128 L.Ed.2d 148 , 114 S.Ct. 1461 ] (McDermott) [identifying, as criteria for selecting an approach, consistency with principles of comparative fault, promotion of settlement, and judicial economy].)

We begin by considering the practical effect of the two apportionment approaches in terms of the amounts ultimately paid by each joint tortfeasor and the total amount recovered by the plaintiff. We also consider the extent to which each approach is consistent with the two basic legal concepts we mentioned earlier: comparative fault, and joint and several liability.

Under the setoff-with-contribution approach, the settlement has little or no practical effect on the defendants’ ultimate liabilities or the plaintiff’s ultimate recovery, nor is the resulting apportionment inconsistent with either the comparative fault principle or the rule of joint and several liability. Without the settlement, each tortfeasor is, under the principle of joint and several liability, fully liable for all of the plaintiff’s damages less only the amount of fault attributed to the plaintiff. If another joint tortfeasor is for any reason unable to satisfy its share of liability, the remaining tortfeasors are still liable for all the economic damages awarded to the plaintiff, even though the amount exceeds their proportionate shares of liability.

With the settlement, under the setoff-wzY/z-contribution approach, this liability exposure of joint tortfeasors is not affected. In a suit against the nonsettling defendants, the plaintiff may recover damages less the settlement amount and the amount attributable to the plaintiff’s own fault. Consistent with the comparative fault principle, the nonsettling tortfeasors may then seek contribution from the settling tortfeasor for any amount they must pay to the plaintiff in excess of their proportionate shares of liability. If in such a contribution action the settling tortfeasor is for any reason unable to fully discharge its share of liability, each nonsettling tortfeasor, consistent with the rule of joint and several liability, remains liable for the difference.

The net result, under the setoff-wzY/z-contribution approach, is that the plaintiff recovers the total economic damages amount (less an amount attributable to the plaintiff’s own negligence), with the settlement amount providing part of that recovery and the judgment against the nonsettling tortfeasors providing the rest. An action for contribution remains available to ensure that the amounts ultimately paid by each tortfeasor are, so far as possible, consistent with that party’s proportionate share of fault for the plaintiff’s damages. For the parties, the only practical difference that the settlement makes is that the settling tortfeasor’s proportionate share of liability is paid in two parts rather than one, with the first part being paid to the plaintiff in the form of the settlement amount, and the second part to the nonsettling tortfeasors as contribution.

We next consider the proportionate-share approach. As explained below, it does have a significant practical impact on the liability positions of the parties. Moreover, it is not consistent with California’s law of joint and several liability.

Under proportionate-share apportionment, the plaintiff’s total recovery for economic damages is limited to the settlement amount plus the proportionate shares of the nonsettling tortfeasors. If the settlement payment turns out to be less than the settling tortfeasor’s proportionate share, as determined by the trial court or the jury, the plaintiff may not recover the difference from any of the tortfeasors and thus is precluded from obtaining full compensation. Therefore, the settling tortfeasor’s liability is not its proportionate share, but only the amount paid in settlement. Because the nonsettling tortfeasors’ liability to the plaintiff for economic damages cannot exceed their proportionate shares, their liability is not joint and several, but several only.

We conclude that the practical implications of the two available apportionment approaches and their consistency or inconsistency with basic tort principles support our adoption of the setoff-with-contribution alternative over the proportionate-share alternative. As explained earlier, setoff-withcontribution apportionment does not change the respective positions of the parties and is fully consistent with both the comparative fault principle and the rule of joint and several liability. In contrast, proportionate-share apportionment alters the parties’ positions and would require us to recognize a new exception to our established law of joint and several liability.

We next consider the relative merits of the setoff-with-contribution approach and the proportionate-share approach in regard to promoting the public policy of settling a case before trial. (See Cassel v. Superior Court (2011) 51 Cal.4th 113, 132 [ 119 Cal.Rptr.3d 437 , 244 P.3d 1080 ] [public policy of encouraging resolution of disputes short of litigation]; T. M. Cobb Co. v. Superior Court (1984) 36 Cal.3d 273, 281 [ 204 Cal.Rptr. 143 , 682 P.2d 338 ] [public policy of encouraging settlements].) That public policy’s laudable goals—decreasing the substantial litigation cost to the parties and lessening the burden that litigation imposes on scarce judicial resources—are not necessarily satisfied in all pretrial settlements. Although good faith settlements are to be encouraged, the same cannot be said of settlements not made in good faith. Indeed, by providing in Code of Civil Procedure section 877 that only good faith settlements will preclude a nonsettling tortfeasor’s contribution action against the settling tortfeasor, the Legislature indicated that settlements not made in good faith should be discouraged. (See Tech-Bilt, supra, 38 Cal.3d at p. 495 .) As explained below, the public policy of encouraging good faith pretrial settlements is furthered by the setoff-withcontribution approach, but not by the proportionate-share approach.'

As we have pointed out {ante, at pp. 305-306), the setoff-with-contribution approach does not change the respective liabilities of the joint tortfeasors. Thus, it provides no incentive for them to enter into a settlement that is not in good faith. In contrast, the proportionate-share approach would encourage settlements not made in good faith: by limiting the liability of the settling tortfeasor (who would be liable only for the settlement amount irrespective of the settling tortfeasor’s proportionate share of liability), and by limiting the liability of the nonsettling tortfeasor (who, under principles of joint and several liability, would no longer be liable for the settling tortfeasor’s proportionate share). Consequently, with respect to a settlement not made in good faith, the setoff-with-contribution approach is superior to the proportionate-share approach in furthering the statutory scheme’s goal of encouraging settlements made in good faith.

With respect to the public interest in promoting judicial economy, we conclude that neither approach is better than the other. Under both approaches, all issues can be resolved in a single action. Under the settlement-with-contribution approach, nonsettling tortfeasors may seek contribution from the settling tortfeasor, but their claims may be combined with the plaintiff’s action against the nonsettling defendants, so that the respective shares of liability can be determined, and all issues resolved, in the same trial. Under the proportionate-share approach, the liability of all tortfeasors is likewise determined in one action, at which the court or jury will establish the total damages and the proportionate fault share of the plaintiff and each of the tortfeasors.

On balance, we conclude that here the settlement-with-contribution approach is preferable to the proportionate-share approach. Unlike proportionate share, settlement with contribution does not change the liability position of the parties from what it would be without a settlement; nor does it require modification of, or a new exception to, our established rule of joint and several liability. Also, the settlement-with-contribution method is, as explained earlier, superior to the proportionate-share approach in serving the policy of promoting settlements made in good faith.

The United States Supreme Court’s decision in McDermott, supra, 511 U.S. 202 , deserves mention. There, the high court selected the proportionate-share method of apportioning liability, a view later adopted in the Restatement Third of Torts, section 16. McDermott involved a plaintiff’s pretrial settlement with some of the joint tortfeasors. The settling tortfeasors agreed to pay $1 million, an amount that, as later determined at trial, far exceeded their proportionate liability. Because the jury found that the settling tortfeasors were only 30 percent at fault, their proportionate share of liability came to only $630,000 out of the $2.1 million assessed in total damages. A significant difference exists, however, between McDermott and the case now before us. Here, unlike in McDermott , the pretrial settlement was determined not to have been made in good faith. (See ante, at p. 299.) Thus, in McDermott no need existed to decide which method of apportioning liability to select in a case involving a settlement not made in good faith, and hence McDermott did not address that issue.

Also, as we have explained (ante, at p. 304), the good faith settlement provisions of California law preclude us from considering the setoff-without contdbution approach, thus limiting us to choosing between the setoff-wzt/zcontribution approach and the proportionate-share approach. The high court in McDermott, supra, 511 U.S. 202 , was not under any similar restraint. This difference in the choices available to the high court (and the authors of the Restatement) and to this court fundamentally changes the analysis.

Between the two alternatives available to us for consideration, we adopt the setoff-with-contribution approach. As explained earlier, that approach is more consistent with our comparative fault principle and our rule of joint and several liability. Accordingly, we hold that when a settlement with a tortfeasor has judicially been determined not to have been made in good faith (see Code Civ. Proc., §§ 877, 877.6, subd. (c)), nonsettling joint tortfeasors remain jointly and severally liable, the amount paid in settlement is credited against any damages awarded against the nonsettling tortfeasors, and the nonsettling tortfeasors are entitled to contribution from the settling tortfeasor for amounts paid in excess of their equitable shares of liability.

IV

Defendant hospital challenges the Court of Appeal’s determination that the evidence was sufficient to prove that negligence by the hospital was a legal cause of plaintiff’s injuries. We agree with the Court of Appeal, as explained below.

The Court of Appeal correctly noted that in evaluating a claim of insufficiency of evidence, a reviewing court must resolve all conflicts in the evidence in favor of the prevailing party and must draw all reasonable inferences in support of the trial court’s judgment. (See, e.g., Scott v. Pacific Gas & Electric Co. (1995) 11 Cal.4th 454, 465 [ 46 Cal.Rptr.2d 427 , 904 P.2d 834 ].) Thereafter, the Court of Appeal stated that in a medical negligence case, causation is adequately established if the evidence is sufficient to permit the jury to infer that without the negligence there is a reasonable medical probability the plaintiff would have obtained a better result. It then quoted this language from Raven H. v. Gamette (2007) 157 Cal.App.4th 1017, 1029-1030 [ 68 Cal.Rptr.3d 897 ]: “ ‘ “If, as a matter of ordinary experience, a particular act or omission might be expected to produce a particular result, and if that result has in fact followed, the conclusion may be justified that the causal relation exists. In drawing that conclusion, the triers of fact are permitted to draw upon ordinary human experience as to the probabilities of the case.” ’ ”

Disagreeing with that observation, defendant hospital contends that in a medical negligence action, common human experience is insufficient to support a trier of fact’s causation determination. Rather, defendant hospital argues, an actual causal link between the negligence and the injury must be shown; the evidence of causation must show a reasonable medical probability that the plaintiff’s injury was more likely than not the result of the negligence, and only expert testimony can establish causation. The hospital asserts that because plaintiff’s expert witness, Dr. Vinod Bhutani, testified at trial that he could not say what plaintiff Aidan’s bilirubin level was on the third day after his birth, plaintiff did not satisfy his burden of proof that negligence by the hospital was a legal cause of plaintiff’s injury.

Even assuming that the appropriate standard applicable to plaintiff’s claims against the hospital in this case is one of reasonable medical probability and that expert testimony is required (see generally Flowers v. Torrance Memorial Hospital Medical Center (1994) 8 Cal.4th 992, 996-999 [ 35 Cal.Rptr.2d 685 , 884 P.2d 142 ] [discussing standard of care in “ordinary” and “professional” negligence actions]), the evidence presented at trial was sufficient to support the jury’s finding that the hospital’s negligence was a legal cause of plaintiff’s injury.

Arthur Shorr, plaintiff’s expert witness on hospital administration, testified that defendant hospital was negligent in failing to implement any of the recommendations of the Joint Commission’s Alert No. 18. Dr. Bhutani, another expert witness for plaintiff, testified that defendant hospital was negligent when its staff failed to tell plaintiff’s parents that a followup appointment with his pediatrician within two or three days was needed, failed to assess plaintiff’s risk of developing hyperbilirubinemia, failed to discuss that risk with his parents, failed to appropriately educate plaintiff’s parents about jaundice, and failed to instruct plaintiff’s mother on breastfeeding so that she knew whether plaintiff was getting an adequate amount of milk, which is a primary way of preventing the buildup in the blood of the bilirubin that caused plaintiff’s injury. Also, plaintiff presented evidence that defendant hospital provided his parents with misleading information in the manual given to them upon plaintiff’s discharge from the hospital some 24 hours after his birth. We agree with the Court of Appeal that this evidence sufficiently supports the jury’s determination that defendant hospital’s negligence was a legal cause of plaintiff’s injury.

Defendant hospital contends that because a hospital does not practice medicine, as a matter of public policy its conduct should not be considered a legal cause of plaintiff’s injuries. Defendant hospital asserts it should not be required to provide medical advice “beyond directing the patient to call the doctor with concerns.” We disagree.

As the Court of Appeal observed in Mejia v. Community Hospital of San Bernardino (2002) 99 Cal.App.4th 1448, 1453 [ 122 Cal.Rptr.2d 233 ], quoting Bing v. Thunig (1957) 2 N.Y.2d 656 [ 143 N.E.2d 3, 8 , 163 N.Y.S.2d 3 ]; “ ‘Present-day hospitals, as their manner of operation plainly demonstrates, do far more than furnish facilities for treatment. They regularly employ on a salary basis a large staff of physicians, nurses and internes [sic], as well as administrative and manual workers, and they charge patients for medical care and treatment, collecting for such services, if necessary, by legal action. Certainly, the person who avails himself of “hospital facilities” expects that the hospital will attempt to cure him, not that its nurses or other employees will act on their own responsibility.’ ” Although hospitals do not practice medicine in the same sense as physicians, they do provide facilities and services in connection with the practice of medicine, and if they are negligent in doing so they can be held liable. Here, defendant hospital implicitly recognized that point when it requested, and the trial court gave, this jury instruction; “A hospital must provide procedures, policies, facilities, supplies, and qualified personnel reasonably necessary for the treatment of its patients.”

For the reasons given above, the evidence presented at trial was sufficient to support the jury’s finding that defendant hospital’s acts or omissions were a legal cause of plaintiff’s brain injury.

V

The Court of Appeal reversed the judgment against defendant hospital insofar as the hospital was held liable for plaintiff’s economic damages. Consequently, the Court of Appeal did not address the hospital’s contentions that the trial court erred in excluding evidence of future insurance coverage, in calculating interest on future periodic payments, and in requiring the hospital to provide security for future periodic payments. Nor, for the same reason, did the Court of Appeal address on its merits plaintiff’s cross-appeal contending the trial court erred in allowing defendant hospital to acquire an annuity payable to itself as security for the future periodic payments it had been ordered to pay under the judgment.

Having concluded that the common law release rule is no longer to be followed in California, and that therefore defendant hospital remains jointly and severally liable for plaintiff’s economic damages, we must reverse the Court of Appeal’s judgment. On remand, the Court of Appeal will address the unresolved contentions, as noted in the preceding paragraph, raised by defendant hospital and by plaintiff.

Disposition

The judgment of the Court of Appeal is reversed, and the case is remanded to that court for further proceedings consistent with the views expressed in this opinion.

Cantil-Sakauye, C. J., Baxter, L, Werdegar, J., Chin, L, Corrigan, J., and Liu, J., concurred.

The petition of appellant Verdugo Hills Hospital for a rehearing was denied May 1, 2013.

Overruled are this court’s decisions in Tompkins v. Clay Street R.R. Co. (1884) 66 Cal. 163 [ 4 P. 1165 ], Chetwood v. California Nat. Bank (1896) 113 Cal. 414 [ 45 P. 704 ], Bee v. Cooper (1932) 217 Cal. 96 [ 17 P.2d 740 ], Kincheloe v. Retail Credit Co., Inc., supra, 4 Cal.2d 21, Pellett v. Sonotone Corp., supra, 26 Cal.2d 705 , Lamoreux v. San Diego etc. Ry. Co., supra, 48 Cal.2d 617 , and their progeny, to the extent they are inconsistent with the views expressed here.